Many carbon monoxide detectors sold in hardware stores sound their alarms for no reason, or fail to sound an alarm even when high concentrations of the deadly gas are present, according to recent laboratory tests.
Industry officials, prodded by the Consumer Product Safety Commission, are drafting tougher standards that may make obsolete the tens of millions currently in use.
Commission officials, who recommend that every home have a carbon monoxide detector, say that the current generation is better than nothing. But they say that when better ones are widely available–which will not be until fall 1998 at the earliest–the commission will decide whether to advise homeowners to replace the ones they have.
In fact, the detectors, whose sales surged after tennis star Vitas Gerulaitis was killed by carbon monoxide poisoning three years ago, have had a rocky time. Chicago began requiring them in most homes in 1994, and then was swamped with calls that turned out to be mostly false alarms. In 1995, standards for the detectors were improved, ending the rampant false alarms.
Carbon monoxide detectors are roughly similar in size and price to smoke detectors, but false alarms in a carbon monoxide detector are more serious because they cannot be checked with the naked eye.
Unlike smoke, carbon monoxide can overwhelm its victims before they have any idea what is happening. Early symptoms are headaches, nausea, drowsiness and confusion. More intense exposure to the gas can kill.
Because the threat is complex, so is the standard, which was set by Underwriters Laboratories, a private nonprofit organization, in 1992 and amended in June 1995; that standard now requires a unit to sound the alarm within 90 minutes if the concentration of the gas reaches 100 parts per million; 35 minutes at 200 parts per million, and 15 minutes at 400 parts per million.
A private laboratory, commissioned by the Gas Research Institute, which is funded by the natural gas industry, tested four brands of detectors over several months, beginning last year, using 24 units of each brand.
The institute sponsored the research because it wants a detector less prone to false alarms, which raise public concerns about the safety of natural gas.
According to Steve Wiersma, the institute’s program leader in health and safety, during the rash of alarms in Chicago, 70 percent were false. He said he was surprised that the research found that in addition to crying wolf, the detectors often fail to sound when they ought to. (Other fuels can release carbon monoxide, but natural gas is so clean that unlike a faulty oil burner, a faulty natural gas burner generally does not produce enough smoke to alert a victim.)
The institute did not identify the brands it tested but said they are all sold nationally. Testers found that half of one brand’s group of 24 units failed: one by alarming too soon and 11 by alarming too late. Of the three other brands tested, between 21 percent and 42 percent failed the test, most by staying silent when they should have been giving the warning.
The president and chairman of First Alert, B. Joseph Messner, said in a telephone interview that the Gas Research Institute’s tests had used his company’s battery-powered detectors and that First Alert, a large manufacturer of the detectors, was so concerned about the results that it sent another 24 units to the same laboratory for testing.
He said the laboratory had not controlled the humidity during the course of the test, and that the introduction of carbon monoxide into the test chamber had lowered the humidity rapidly, in a way that would not occur in a house. At a realistic humidity level, he said, First Alert’s units worked fine.
First Alert disputes the Gas Research Institute findings, he said, adding, “We are 100 percent confident our products will go into alarm” when they are supposed to.
Messner said the false alarm problem had been greatly reduced by the 1995 revision to the standard and that the next revision, which First Alert supports, might eliminate it.
The Consumer Product Safety commission says it does not know whether the units performed so badly in the tests because of a problem with the detectors or with the test itself. It also cited humidity as a factor.
“Scientists will tell you that it would be foolish to draw conclusions from one set of tests,” Begala said. But, she added, “it is cause for us to do our own testing.”
The commission won’t decide for months whether consumers should pull out all existing units and replace them. But Ronald L. Medford, the assistant executive director for hazard identification and reduction, said, “Certainly that’s an issue before the agency.”
And even before its testing is finished, the commission has asked Underwriters Laboratories to toughen its standards.
At the commission, Elizabeth Leland, project manager for the detectors, said that to show they meet the current standard, manufacturers had been allowed to use a fresh detector for each phase of the test, so the procedure never verified whether a detector, after it had alarmed once, would continue to function acceptably.
And the standard did not specify testing the detector as it aged, which is what was done in the tests sponsored by the Gas Research Institute.
Underwriters Laboratories, which issues standards for thousands of electric devices, has been reluctant to talk about the problems with carbon monoxide detectors.
Asked if it had received recommendations from the commission, Paul Patty, an associate managing engineer who is in charge of the standard, said that “standards are dynamic in nature,” and that “we have continuous communication with the commission, as well as others, related to changes in standards.”
The new Underwriters Laboratories standards–which will require that detectors be tested over a period of time and with repeated soundings–may go into effect by the end of the year, but manufacturers would have many months to comply.
How the manufacturers will meet a tougher standard is not clear, but perhaps it will involve new technology.
Existing detectors are of two kinds. The round, battery-powered models are called “biomimetic,” because they have a gel that absorbs carbon monoxide the way the human body does. The gel changes color as it absorbs the gas; then, a sensor measures the color change and sounds an alarm.
Units that run on house current use metal oxide semiconductor technology. In that system, a strip of metal becomes a better conductor as it is exposed to carbon monoxide; a microprocessor senses the change and sounds an alarm.
A newer, electrochemical technology uses a material that reacts chemically when it is exposed to carbon monoxide. The reaction gives off electricity, setting off an alarm.
Aim Safety Co., of Delta, B.C., and Austin, Texas, has begun marketing one such detector, but so far it is available mainly through gas utilities, not stores.
