Features November 15, 1998 Issue

CO Detectors: It’s MTI Safe-T-Alert

Combine any enclosed occupied space with a combustion process and you have the recipe for possible carbon monoxide poisoning. An inboard-powered boat, by definition, meets that criterion, but along with the engine there are other sources of CO: stoves, hot water heaters and gensets.

Yet, according to a recent survey by BOAT/U.S., 41% of boat owners say they donít have a CO detector. And many who do (it seems to us) disconnect them because of frequent false alarms, a fault weíve reported on in the past. Itís been several years since weíve reported on actual CO poisoning cases and looked at alarms. Here, we look at both marine and residential alarms.

Danger of CO
In low concentrations, carbon monoxide can cause headache, nausea and other symptoms that may easily be mistaken for other illnesses. In higher concentrations, it can overwhelm a victim, causing loss of consciousness before awareness of the danger is perceived, and death if no help arrives quickly. CO kills by asphyxiatingóthat is, by preventing oxygen from reaching the bodyís tissues. It does that by interfering with the bodyís oxygen transport system, blood hemoglobin.

Hemoglobin picks up CO much more readily (about 200 times more so) than it picks up oxygen, and it releases CO that much less readily. Hemoglobin that combines with CO is called carboxyhemoglobin, or COHb. While COHb is not harmful or toxic in itself, it wonít carry oxygen. When significant amounts of the bodyís hemoglobin are converted to COHb, you can begin to suffer from oxygen deprivation. As long as youíre exposed to CO, at any concentration, the percentage of your hemoglobin thatís converted to COHb will continue to rise, slowly at low concentrations and more rapidly at higher concentrations.

Because the body is so slow to rid the blood of CO (it takes about five hours for the COHb concentration in the blood to drop by half), treating CO poisoning is much less effective than preventing it. So CO detectors and alarms have a genuine role to play.

CO Sources
Internal combustion engines are major sources of CO. Boat-engine exhausts are not cleansed by catalytic converters as are car engines and can contain around 2% CO. Open-flame devices generally produce less, but can still build to toxic levels in small, enclosed galleys or cabins. If the flame is properly adjusted and hot enough and thereís plenty of air, much of the CO produced by an open flame is immediately reburned to form carbon dioxide (CO2), which is not toxic.

On a boat, any leakage of engine exhaust fumes, from your own boat or a neighborís, into the cabin is a very serious situation. Even while under way, you could easily find your own engine exhaust coming into the cabin through an open entry or curtain, either from a following wind or from the so-called ďstation wagon effect.Ē

The relationship between CO concentration, length of exposure and biochemical effects is complex. A standardized set of curves has been developed, representing a calculated equivalent COHb concentration at varying times and concentrations of CO in the air. These curves, called Beta curves, are derived from a mathematical formula based on the measured activity of hemoglobin in healthy men. Beta is a fairly optimistic estimate of COHb blood level; infants, the elderly and people with cardiac problems, among others, may well be affected more rapidly and more seriously.

These curves reveal some sobering aspects of the CO problem. For one thing, the effects are not linear. Doubling the concentration of CO cuts the time to reach a given COHb level by more than half. For example, exposure to 200 ppm of CO for 90 minutes produces a 25% COHb level, which can cause headaches and nausea. Doubling the CO level to 400 ppm reduces the time to about 35 minutes. Another tipoff to the hazard: These are very tiny concentrations; 400 ppm is less than 1/20 of 1%. A concentration of 1/10 of 1%, or 1,000 ppm, can cause death in a half hour.

CO Detectors
There are three basic types of sensors to detect and measure CO: tin oxide, electrochemical and chemical. Each has advantages and disadvantages.

The tin oxide sensor is used by the two marine detectors we evaluated. A Japanese company, Figaro, is the leading producer of this type of sensor. The sensorís resistance varies according to the CO level, and this resistance is detected by a microprocessor which computes the equivalent COHb level. Tin oxide sensors are stable and reliable but subject to errors induced by contamination. Other hydrocarbon fumes can affect the sensor and cause false alarms.

Electrochemical sensors have a cell which produces a voltage that varies with the concentration of CO. They are very accurate and specific to CO while unaffected by other hydrocarbons, so theyíre not prone to false alarms. Unfortunately, theyíre expensive and have a relatively short lifetime, since the ingredients of the cell are consumed during operation, much as in a flashlight battery. Because of these limitations, this sensor is used mainly in test instruments. The small meter provided to us by Fireboy/Xintex as a reference for our test uses an electrochemical sensor.

The third kind is the chemical sensor, in which a chemical changes color in response to CO. Chemicals have been found which closely mimic the behavior of hemoglobin. By shining a light through the chemical, the color density change can be measured by a photocell and an equivalent COHb level computed by a microprocessor. Because detectors using this kind of sensor mimic the biological effect of CO, they have been dubbed biomimetic. Some home detectors use chemical sensors, including the First Alert, discussed below.

False Alarms
False alarms have been a bugaboo of marine CO detectors. There are a couple of theories as to why.

One holds that boats, unlike homes, are prone to having a variety of other hydrocarbon gases in the atmosphere that contaminate the sensors, make them hypersensitive, and cause the alarms to trip needlessly. Boats, especially new ones, may produce a lot of solvent outgassing from the fiberglass hulls and other plastics, carpets and materials. Also, propane and alcohol, often used as lantern and cooking fuels in cabins, may accumulate and confuse the detector.

Another theory holds that many ďfalseĒ alarms are not false at all, but are produced by actual CO levels detected before there are symptoms. Because thereís no other way to confirm the presence of CO (the old ďodorless, colorless gasĒ trick), itís easy to assume itís not there. According to Keith Weldy, Fireboy/Xintexís president, ďIf carbon monoxide had a color and were visible, we would hear a lot less about false-alarming.Ē

Tom Wisniewski, president of Marine Technologies, agrees, citing a case in which several people were found unconscious from their own engine exhaust blown into their boat after extended time trolling. One of MTIís alarms was fitted in the boat but the wiring had been cut to silence it after several ďfalse alarms,Ē which evidently were not false, after all.

Obviously, a detector that reacts to short-term transients, such as a brief puff from an engine exhaust, would be a nuisance. All the makers incorporate some type of time-weighted averaging to measure the cumulative effect of CO over a period of time rather that an instantaneous concentration. The chemical sensors react at the same rate as your bloodstream, and the electronic sensors use algorithms to compute the equivalent effect. We found that our models did not react to brief high-level exposures.

What We Tested
Only two the five makers of the alarms we tested in our 1993 report still make detectors specifically for boats. Fireboy/Xintex makes an updated version of essentially the same model as before, now called the CMD2-M. MTI has a new model, the Safe-T-Alert 60 542. Quantumís updated model is the COStar 9L-i, but the company markets it only for recreational vehicles, and asked us to mention that they do not sell it for use on boats. The Quantum Eye, a dosimeter-type device that is not an alarm, is also still sold.

Among CO detectors for the home, we found three battery-powered models, all listed by Underwriters Laboratories (UL).

How We Tested
Underwriters Laboratories (UL) has two standards which define the performance requirements for CO detectors, one for residential applications and one for marine. Alarms for marine use are listed under UL Standard 1524, which sets criteria for sensitivity and detection limits, and for environmental factors such as shock, vibration, temperature and humidity resistance. Our objective was not to replicate ULís acceptance tests but to set up what we judge to be a real-world evaluation process that checks the basic function of each alarm.

The primary difference between UL standards for marine and residential units involves sensitivity requirements. Residential models are required to sound an alarm at the equivalent of 10% COHb blood levels, marine models only when the threshold of 20% is reached.

We set up a test chamber consisting of a large glass boxóa fish tank, to be precise. Inside the box were placed the alarms, along with a small fan to circulate the air. The box was covered with a sheet of clear plastic taped in place. A dowel rod pierced the top and was used to press the control buttons on the alarms. A small garden sprayer was used to collect automobile exhaust gas, compress it slightly, and inject it into the chamber in small amounts.

We used a calibrated CO sensor (loaned by Fireboy/Xintex), which gives instant readings in ppm to measure the concentration in the chamber. With occasional small additions of CO from our pressurized sprayer to make up for losses, we were able to maintain a fairly stable level of around 100 ppm. We then watched how the alarms responded and noted the times. In a separate test, we checked alarm loudness with a dB meter. In general, all of the alarms pretty much met their performance claims. Here are the specifics.

Two models sold primarily for marine use meet UL Standard 1524. Both connect to the boatís 12-volt system.

Fireboy/Xintex CMD-2M
This model appears basically the same as the one we reported on in 1993, though the maker says the circuitry has been improved. The horn continues to sound until CO levels drop to 75 ppm, and you canít turn off the alarm manually except by disconnecting the power. Its compact case can install easily in almost any location.

One change from the earlier model is apparent. There used to be a warning signal (an amber flashing light and an intermittent beep), followed by the full alarm if the hazard persisted. We didnít see much sense in that, since your reaction should be the same in either case. Fireboy evidently came to the same conclusion and has gone to a single alarm.

The Fireboy/Xintex CMD2-M is nicely made and simple to install. The circuit board is coated with a varnish, which the company says protects it against the effects of salt air. Price is $88, a third less than five years ago.

Bottom Line: A basic, functional alarm.Drawback: Until the hazard is gone, you canít silence the horn unless you disconnect the power.

MTI Safe-T-Alert 60-542
We described this unitís predecessor in 1993 as the plain vanilla entry because it worked so simply. At the danger level, it sounded an alarm. Press a button to quiet the alarm and it sounds again in 7 minutes if the CO level is still high.

This newer MTI model still has the 100-ppm alarm, consisting of a continuously pulsing horn and a flashing red signal that starts within 90 minutes. Thatís equivalent to a blood COHb level of 10%, which is well below what ULís marine standard requires. Whatís new is an added low-level alarm, a short beep every 5 minutes with a flashing yellow signal, triggered by 60 ppm within 240 minutes. But thatís also equivalent to about 10% COHb, so the two different alarms represent approximately the same hazard level. We donít think thereís enough difference to warrant two very different warnings.

Pressing a reset button stops the alarm, which will resume in a short time if the CO level has not dissipated. Thereís also a memory feature activated by a short tap on the reset button. The number of beeps and flashes is a code which indicates the highest CO level the detector has noted, ranging from 60 to 200 ppm. This feature is not intuitive to use and we donít think itís of much value.

The Safe-T-Alert 60-542 is compact and nicely made and easy to mount. It has come down in price and is a good value, now listing for about $70.

Bottom Line: Works fine as a basic alarm. Not improved materially over the earlier model by the added bells and whistles, but then it didnít have to be.

The following models are sold for residential use and are listed under UL Standard 2034, which sets the 10% COHb beta curve as the detection limit. They may not meet UL requirements for marine use.

Kidde Lifesaver 9CO-1
This residential CO alarm is about the same size and shape as a smoke detector and is powered by a replaceable 9-volt battery, good for about a year. It is unusual because it has a low-level warning at levels as low as 60 ppm, equivalent to about 5% COHb, an effective early warning. The full alarm sounds at 100 ppm and higher, calibrated approximately to 10% COHb.

This model is particularly simple to set up and operate. It mounts on a wall or has a stand to set it on a tabletop. A yellow LED and short beeps indicate a warning. A red LED and long beeps indicate a full alarm. Pushing the reset button temporarily silences the warning for 30 minutes and the full alarm for 4 minutes, but the LEDís stay on. The alarms will recur if the CO level has not been reduced. In our test, the low-level warning went off after about 30 minutes at 80 ppm, which is below the 5% COHb level, a very early warning.

This modelís dual alarms offer a meaningful early warning. Itís available in many hardware stores and home centers and lists for $34.95, but is generally discounted.

Bottom Line: A good alarm with an effective early warning. Drawback: It requires fresh batteries.

Kidde Nighthawk
Another residential model, this unit features a nearly real-time digital readout of CO levels in ppm. Itís powered by 3AAA self-contained batteries and has the familiar smoke-detector shape. It can wall-mount or stand on a table or shelf. We checked the digital readout against our calibrated meter and found it off by less than 10%.

A memory stores the highest level recorded since power-up or since last being manually cleared (done by pressing the reset and peak-level buttons simultaneously). A green LED blinks occasionally if all is right. A red LED blinks and the horn chirps to indicate low batteries or other faults. The detectorís circuitry is programmed to warn at the 10% COHb level required for residential models.

The Nighthawk isnít sold for marine use, which we think is a shame. Itís full of desirable features, but isnít UL listed for marine applications. The digital readout gives you an immediate check on conditions at that momentóeven levels below those required to set off the alarm, which we think is useful. Itís easy to install and maintain. It can be bought for under $60 at retail outlets.

Bottom Line: Full of features that work, especially the useful digital readout. Drawback: You need to keep fresh batteries in it, and its lack of approval for marine use.

First Alert FCD1
The First Alert residential CO detector resembles a smoke detector and mounts the same way. Itís powered by a self-contained battery which is packaged in a replaceable module that also includes a new sensor. When the battery is depleted, the alarm goes into a chirp mode. The battery-sensor module is supposed to last at least two years.

This model uses a chemical biomimetic detector that changes color according to itís accumulated CO. The color change is designed to reflect the COHb level a person would reach from that exposure. A photoelectric device measures the color change and triggers the alarm when the danger level is reached.

Thereís an early warning alarm with intermittent beeps at approximately 5% COHb and a full alarm with continuous horn at 10%.

Early First Alert models were implicated in a notorious event in Chicago some years ago, when thousands that were installed in that city started producing an avalanche of emergency calls. It turned out that the detectors were just too sensitive for ordinary home use. Since then, UL changed its requirements and the sensors have been recalibrated. The First Alert is easy to mount on almost any surface and sells for about $30.

Bottom Line: A good alarm with an effective early warning. Drawbacks: Expensive sensor/battery replacement every two years, and lack of approval for marine use.

Quantum Eye
The Quantum Eye is not an alarm. Itís a small card that has a chemical indicator that gives a visual indication of CO conditions. According to the instructions, yellow means normal conditions; green means cautionóventilate the space; and blue means dangeróget out immediately.

We found the Quantum Eye started to show color changes after exposure to about 40 ppm after about 20 minutesófar below any COHb level on the Beta curves. Checking it every once in a while would tip you off to increasing CO levels.

The card comes with a peel-and-stick adhesive spot to attach it to any surface, and a stick-on magnet. It has a useful life of 18 months under normal conditions and costs under $20.

Bottom Line: A useful adjunct to an alarm, but not for protection while you sleep.

We prefer detectors that warn you of the differing danger levels in appropriate and clearly differentiated ways.

The residential units are simple to use, but we would only use one for back-up. Of these, we like the Kidde Nighthawk, primarily because of its real-time digital read-out.

The marine units must pass tougher environmental requirements and you donít have to replace batteries as you do with the residential models. Of the two marine units tested, weíd opt for the MTI, because you can temporarily quiet the alarm horn while you ventilate the cabin, and because it alarms at lower levels.

Contacts- Fireboy/Xintex, 100 Commerce Ave. SW, Grand Rapids, MI 49501; 616/454-8337. First Alert, BRK Brands, 390 Liberty Street Rd., Aurora, IL 60514; 630/851-7330. Kidde, 1394 S. Third St., Mebane, NC 27302; 919/563-5911. MTI, Marine Technologies, 1000 Brown St., Wauconda, IL 60084; 847/487-4940. Quantum Group, 11211 Sorrento Valley Rd., San Diego, CA; 800/432-5599.

Comments (0)

Be the first to comment on this post using the section below.

New to Practical Sailor?
Register for Free!

Already Registered?
Log In