| This
study assessed the performance of commercially available carbon monoxide
(CO) alarms, several preproduction prototypes, and several alarms
removed from residences by first responders. Alarm performance was
assessed in terms of the Underwriters Laboratories UL 2034 design
specification for sensitivity to CO, immunity to interfering gases,
accuracy of digital displays, and ability to detect CO in ambient air of
varied relative humidity. Findings parallel the results of earlier
studies, indicating that the poor performance of commercial CO alarms
has not significantly improved. These tests reveal that a significant
portion of some widely installed alarm brands did not alarm in CO
poisoning incidents with potentially lethal consequences. In particular,
they did not adequately protect at low relative humidity. Digital
display readings on some alarms showed little relationship to actual CO
levels. Results indicate that certification to UL 2034 does not assure
the performance or reliability of CO alarms. The study sheds light on
the physical and algorithmic determinants of the poor performance of CO
alarms, and proposes solutions for some of their most serious technical
deficiencies. It is recommended that reliability standards, including
specifications for alarm longevity and time-of-manufacture testing, as
previously recommended by the Consumer Product Safety Commission (CPSC)
and GRI, be incorporated in the UL 2034 standard. Gas utilities,
appliance manufacturers and others are urged to require certification to
the Canadian Standards Association CSA 6.19-01 standard, which includes
improved reliability specifications. |