This article is reprinted from the April 2000 issue of Energy Design Update.

Reprinted with permission from Energy Design Update.  Copyright Aspen Publishers.  For information on subscribing to Energy Design Update call 800-638-8437.

Canada's New CO Alarm Standard Deemed "World's Best"

The newly revised Canadian standard for carbon monoxide (CO) alarms, which takes effect this month, is being hailed as on ot the greatest victories in consumer protection.  For the first time ever, manufacturers who want to certify their CO alarms through the Canadian Standards Association (CSA) will have to have a documented reliability testing program in place.

"I've been praying for years that Canada would get behind a strong reliability standard like this, "says George Kerr, an internationally recognized authority on life safety systems and a voting member on the CSA Technical Committee on Carbon Monoxide Detectors.  "I can brazenly say that Canada now has the best CO detector standard in the world!"

The new Canadian standard (CAN/CGA-6.19-M93 is in many regards identical to the standard the Underwriters Laboratories (UL, - Northbrook, Illinois) put into place in October 1998 (UL 2034-98).  Both require CO detectors to alarm before a time-weighted average of CO exposure can create a carboxyhemoglobin level of 10% in the human blood.  But a new amendment in the Canadian version goes a bold step further in ensuring detector reliability, requiring production line testing not only at the time of manufacturing, but also periodically throughout the product's stated lifetime.

"Under the UL standard, an alarm manufacturer can build 20,000 samples, handpick a few, and submit them to UL for testing," Kerr explains.  "If they pass the test, the manufacturer is them free to build a million more of them without any further testing - ever- so long as the primary components remain unchanged."  Kerr says that this long-standing flaw in the UL standard has given manufacturers license to make outlandish guarantees (e.g., a 10 year warranty), knowing that the sensors inside their alarms are likely to drift or fail outright long before the warranty actually expires.  "The figure that most consumers aren't going to keep the proof of purchase and warranty statement anyhow," Kerr adds.

In 1996, Kerr wrote an impassioned letter to the US Consumer Product Safety Commission (CSPC), stating that the fraud had gone on long enough and pleading with CPSC to help pressure UL into adopting lifetime reliability testing in its standard.  The American Gas Association, Gas Research Institute (GRI), and Illinois Fire Chiefs' Conference were also pushing for changes, which eventually culminated in UL 2034-98 (see EDU December 1998).  The revised standard was supposed to have included reliability testing, says Kerr, but UL backed out when the manufacturers complained that it would cost them too much money.  Kerr along with a growing number of other critics, contend that UL upt manufacturers' interests above consumer safety.

By contrast, the CSA standard requires listed manufacturers to randomly pull a  minimum of 230 CO alarms off the production line during the frrst quarter and test them,  If there's a failure, the size of the test sample must be increased,  For example, the sample size per 100,000 alarms would grow to 11,400 if a failure rate of 0.88% exists.  Moreover, if a manufacturer claims a warranty on an alarm, samples pulled off the production line must continue to undergo quarterly tests for the duration of the warranty period.  "In essence, it makes a manufacturer put his money where his mouth is," Kerr says.

Industry insiders are betting that the dramatic changes north of the border with force UL's hand.  "I think the new Canadian standard will prompt UL to roll out a revision to UL 2034-98 in the months ahead," says Steve Wiersma, who oversees GRI's alarm-test program.  "There's also going to be mounting pressure from Europe, there a new community standard (CENELEC) for CO alarms has just been passed.  Some of the alarms that carry the UL label here wouldn't pass muster  under the CENELEC standard."  Wiersma tells EDU that he hopes the results from his current year-long test program at GRI - due out in May - will also encourage UL to adopt tougher standards.

"You have to remember that UL gets paid about $30,000 per test series and a small per- unit fee on every alarm that caries the UL logo, so thy are very responsive to manufacturers," Wiersma says.  "Some of the recommendations that GRI made three or four years ago were just filed away, because the alarm manufacturers didn't want to move on them."

Last August, GRI presented a list of nine recommendations to the UL Technical Advisory Panel for Standard 2034-98.  Among those was a proposal to adopt the reliability testing standard that CSA has enacted.

UL does seem to be moving in the right direction albeit grudgingly.  Technicians from the standards organization recently went out an bought a few dozen UL-listed CO alarms - representing 30 models in all - and tested them.  Though UL insists that the sampling was too small to be statistically significant, there must be a message in the fact that 25% of the alarms failed at least on UL response test. 

Industry observers note, with some irony, that had UL enacted reliability testing with the 1998 revision, last year's recall of one million faulty Nighthawk and Lifesaver CO detectors could have been prevented.  They say that periodic production line testing - like that required in the new Canadian standard - would have exposed the problem in its earliest stages.  (The failures were attributed to out gassing from the plastic packaging, which damaged the sensors before they even arrived in stores.)

"Thanks to the new Canadian standard, we're going to reduce the number of false alarms and false negatives, and save lives," Kerr concludes.  "But it will also be in the manufacturer's best interest over the long run."

He tells EDU that the CSA Technical Committee is already discussing additional measures that would make the standard even stronger.  These might include:

• A requirement that the detector's expiration date (end of useful life) be clearly printed on both the box and the alarm itself.
• A requirement that all CO alarms include a kill switch, which would put the CO alarm into an error or fault mode at the end of its life, before it begins to give off false alarms.

Taking all this into account, EDU recommends that home builders, HVAC contractors, utilities, and others who are buying CO alarms select models that are certified under CAN/CGA-6.19-M93-99.  In Canada, such alarms carry a distinctive blue flame on the label.  In the US, the corresponding standard is IAS-696, with a blue star logo.

Return to CO Alarm Standards    Return to Main Page