Carbon Monoxide Detectors

Very few pilots carry any sort of carbon monoxide (CO) detector when they fly, except perhaps the el-cheapo chemical spot detectors that leave much to be desired. Mike Busch analyzes the CO-related accident history, discusses the medical aspects of CO poisoning, and evaluates the various kinds of CO detectors available for under $100.


NOTE: This article was updated in November 2003 to reflect the unavailability of the previously recommended AIM and Senco detectors (both of which are no longer in production), and the introduction of the CO Experts Model 2002 which has now become our top choice for aircraft use.

The dramatic crash of Piper Dakota N8263Y madeall the evening TV newscasts on Friday, January 17, 1997. The experiencedinstrument-rated pilot and his 71-year-old mother had departed FarmingdaleAirport on New York’s Long Island at 11:15 a.m. on a VFR flight to Saranac Lake,N.Y., about two hours’ flying time to the north. But less than a half-hour intothe flight, something went terribly wrong: the pilot-in-command passed out cold.Thirty-six minutes into the flight, the passenger (who was herself a low-timeprivate pilot) radioed Boston Center and told the controller that the pilot wasunresponsive and vomiting, and they were in trouble. After determining that thepassenger was pilot-rated, the controller spent the next 20 minutes trying totalk her down to a landing at Bridgeport, Conn. An Air National Guard helicopterjoined up with the aircraft and participated in the talk-down attempt, butwithout success.

Route of Dakota N8263YForty-five minutes into the flight, the woman reported that she, too,was getting tired and nauseated, and was unable to awaken the pilot. Shortlythereafter, the airplane turned north and started climbing, and the womanstopped responding to radio calls. The aircraft gradually climbed into the5,000-foot cloud bases and continued climbing to 8,800 feet. The helicopter lostsight of the emergency aircraft but attempted to follow it with the help of ATC.About two hours into the flight, the aircraft descended out of the clouds andthe helicopter established visual contact, reporting that the cabin appeared tobe full of smoke and nobody was visible through the windows. Not longafterwards, the helicopter pilot reported that the Dakota had started descendingrapidly and crashed into the woods near Lake Winnipesaukee, N.H. Both occupantswere fatally injured in the crash.

Toxicological tests in the FAA lab in Oklahoma City revealed that the pilot’sblood had a carboxyhemoglobin (CO) saturation of 43%, and the passenger’smeasured 69%. Those concentrations are sufficient to produce convulsions andcoma. NTSB metallurgists determined that the muffler contained a large crack andan irregular hole, both of which appeared to have been leaking exhaust gas forsome time.

NOTE: The full text of the NTSB factual report on this accident [NTSB reference number IAD97FA043] is available here.

A week after the Dakota crash, the AOPAAir Safety Foundation issued a press release cautioning pilots about thedangers of carbon monoxide poisoning, and recommending that pilots ofsingle-engine aircraft install a CO detector. However, ASF executive directorBruce Landsberg noted that CO accidents are “extremely rare,” adding that “asearch of the Air Safety Foundation accident database revealed only twoaccidents caused by carbon monoxide between 1985 and 1994.”

Extremely rare? Don’t bet your life onit!

Much as I hate to contradict my good friend Bruce, my own quick search of theNTSB accident database suggests that CO-related accidents and incidents occurfar more frequently than the AOPA Air Safety Foundation’s statement would haveyou believe. But don’t take my word for it. Take a look at just some of theaccidents and incidents my brief search turned up:

  • March 1983. The Piper PA-22-150 N1841P departed Oklahoma City, Okla., with en route stops planned at Amarillo, Texas, and Tucumcari, N.M. At Tucumcari, the occupants found themselves “staggering a little” and concluded this was from exposure to high altitude. After leveling at 9,600 feet on the next leg of the flight, the right-front-seat passenger became nauseous, vomited, and fell asleep. The pilot began feeling sleepy and passed out. The aircraft began a circling descent, and efforts by the rear seat passengers to revive those in the front seats were unsuccessful. A 15-year-old passenger in the back seat took control of the aircraft by reaching between the seats, but the aircraft hit a fence during the emergency landing. None of the four occupants were injured. Multiple exhaust cracks and leaks were found in the muffler. The aircraft had recent annual and 100-hour inspections. The NTSB determined the probable cause of the accident to be incapacitation of the pilot-in-command from carbon monoxide poisoning. [FTW83LA156]
  • February 1984. The pilot of Beech Musketeer N6141N with four aboard reported that he was unsure of his position. ATC identified the aircraft and issued radar vectors toward Ocean Isle, N.C. Subsequently, a female passenger radioed that the pilot was unconscious. ATC and another aircraft tried to assist, but the aircraft crashed in a steep nose-down attitude, killing all occupants. Toxicological tests of the four victims revealed caboxyhemoglobin levels of 24%, 22%, 35% and 44%. [ATL84FA090]
  • November 1988. The Cessna 185 N20752 bounced several times while landing at Deadhorse, Alaska. The pilot collapsed shortly after getting out of the airplane. Blood samples taken from the pilot three hours after landing contained 22.1% carboxyhemoglobin. The left engine muffler overboard tube was broken loose from the muffler where the two are welded, allowing exhaust gas to enter the cockpit through the heater system. The overboard tube had a .022-inch-deep groove from rubbing against the engine cowling. The muffler had been installed during the last 100-hour inspection, 68 hours prior to the incident. The NTSB determined probable cause to be physical impairment of the pilot-in-command due to carbon monoxide poisoning. [ANC89IA019]
  • July 1990. While on a local flight, the homebuilt Olsen Pursuit N23GG crashed about three-tenths of a mile short of Runway 4 at Fowler, Colo. No one witnessed the crash, but post-crash investigation indicated that there was no apparent forward movement of the aircraft after its initial impact. The aircraft burned, and both occupants died. Investigators found no evidence of mechanical failure, but toxicology tests of the pilot and passenger were positive for carboxyhemoglobin. The specific source of the carbon monoxide was not determined. [DEN90DTE04]
  • August 1990. About 15 minutes into the local night flight in Cessna 150 N741MF, the aircraft crashed into Lake Michigan about one mile from the shoreline near Holland, Mich. Two days later, the wreckage of the aircraft was found in 50 feet of water, with the passenger’s body still strapped in. The pilot’s body was found 12 days later when it washed onto the shore. Autopsies were negative for drowning, but toxicological tests were positive for carboxyhemoglobin, with the pilot’s blood testing at 21%. [CHI90DEM08]
  • July 1991. The student pilot and a passenger (!) were on a pleasure flight in Champion 7AC N3006E owned by the pilot. The aircraft was seen to turn into a valley in an area of mountainous terrain, where it subsequently collided with the ground near Burns, Ore., killing both occupants. Although the aircraft was three years out of annual, investigators found no evidence of pre-impact mechanical failure. A toxicology exam of the pilot’s blood showed a saturation of 20% carboxyhemoglobin, sufficient to cause headache, confusion, dizziness and visual disturbance. [SEA91FA156]
  • October 1992. The pilot of Cessna 150 N6402S was in radio contact with the control tower at Mt. Gilead, Ohio, and in a descent from 5,000 feet to 2,000 feet in preparation for landing. Seven miles south of the airport, the airplane was observed on radar flying away from the airport. Radar contact was lost, and the aircraft crashed into a wooded area, seriously injuring the pilot. Toxicological tests on the pilot’s blood were positive for carbon monoxide. Examination of the left muffler revealed three cracks and progressive deterioration. The NTSB found probable cause of the accident to be pilot incapacitation due to carbon monoxide poisoning. [NYC93LA031]
  • April 1994. Fifteen minutes after takeoff from Long Beach, Calif., the Cessna 182 N9124G began deviating from headings, altitudes and ATC instructions. ATC said the aircraft’s course became increasingly erratic as the flight continued, and that the pilot seemed disoriented. The aircraft drifted significantly off the assigned airway and headings, and did several 360- and 180-degree turns. The pilot reported blurred vision, headaches, nausea, labored breathing, and difficulty staying awake. The aircraft ultimately crashed in a vineyard near Kerman, Calif., following an uncontrolled altitude deviation, and the owner/pilot was seriously injured. Post-crash inspection revealed numerous small leaks in the exhaust system. The pilot tested positive for carbon monoxide even after 11 hours of oxygen therapy. [LAX94LA184]
  • October 1994. A student pilot returned to Chesterfield, Mo., from a solo cross-country flight in Cessna 150 N7XC, complaining of headache, nausea, and difficulty walking. The pilot was hospitalized, and medical tests revealed elevated CO, which required five and a half hours breathing 100% oxygen to reduce to normal levels. Post-flight inspection revealed a crack in an improperly repaired muffler that had been installed 18 hours earlier. [CHI95IA030]
  • March 1996. The pilot of Piper Cherokee 140 N95394 stated that she and her passenger became incapacitated after takeoff from Pittsburg, Kan. The airplane impacted the terrain, but the occupants were uninjured. Both were hospitalized for observation, and toxicological tests for carbon monoxide were positive. A subsequent examination found holes in the muffler. An annual inspection had been completed four flight hours prior to the accident. [CHI96LA101]
  • August 1996. A Mankovich Revenge racer N7037J was #2 in a four-airplane ferry formation of Formula V Class racing airplanes. The #3 pilot said that the #2 pilot’s flying was erratic during the flight. The witness said when they were within a mile of the landing airport, the pilot “pulled stright up, pulled left to the east at full power, then went into a slight descent.” The witness said that he flew up alongside the pilot’s airplane to try to get his attention. “I couldn’t get his eye. He would not even look at me. I chased him about five miles before I lost sight of him. The last time I saw him, he was below 500 feet.” The airplane crashed near Jeffersonville, Ind., killing the pilot. The results of FAA toxicology tests of the pilot’s blood revealed a 41% saturation of carboxyhemoglobin; loss of consciousness is attained at approximately 30%. Examination of the wreckage revealed that the adhesive resin that bound the rubber stripping forming the firewall lower seal was missing. The NTSB determined probable cause of the accident to be pilot incapacitation due to carbon monoxide poisoning. [CHI96FA322]
  • January 1997. The fatal crash of Piper Dakota N8263Y near Lake Winnipesaukee, N.H. described at the beginning of this article. [IAD97FA043]
  • December 1997. Dr. Bob Frayser, a family physician, was piloting his Piper Comanche 400 N8452P from his hometown of Hoisington, Kan., to Topeka when he fell asleep at the controls. The airplane continued on course under autopilot control for 250 miles until it ran a tank dry and glided to a soft wings-level crash-landing in a hay field near Cairo, Mo. The pilot was only slightly injured, and walked to a nearby farm house for help. Toxicology tests revealed a 26.8% carboxyhemoglobin saturation some two hours later. Post-crash inspection revealed that the right muffler had a crack around one of its seams that would allow exhaust fumes into the cabin heat system. [CHI98LA055]
  • December 1997. A new Cessna 182S was being ferried from the factory in Independence, Kan., to a buyer in Germany when the ferry pilot felt ill and suspected carbon monoxide poisoning. She landed successfully, and examination of the muffler revealed that the muffler had been manufactured with defective welds that allowed CO to enter the cabin through the cabin heat system. Subsequent pressure tests by Cessna of new Cessna 172 and 182 mufflers in inventory revealed that 20% of them had leaky welds. The FAA stepped in and issued an emergency Airworthiness Directive requiring muffler replacement on some 300 new Cessna 172s and 182s. [Priority Letter AD 98-02-05]

Still think in-flight CO poisoning occurs too rarely to worry about? I didn’tthink so!

The fact is that deaths from unintentional carbon monoxide poisoning havedropped sharply in recent years, thanks mainly to lower CO emissions fromautomobiles with catalytic converters (60% of CO deaths are motorvehicle-related) and safer heating and cooking appliances. In contrast,CO-related airplane accidents seem to be on the increase as the fleet growsolder and the maintenance infrastructure deteriorates. The recent mufflerproblem on new Independence-built Cessna singles demonstrates that even newairplanes aren’t immune.

How CO kills

Lycoming IO-540 engine on 1997 Skylane.Carbonmonoxide is an invisible, odorless, colorless gas created when fossil fuels(such as gasoline) burn incompletely. In a piston-powered aircraft, engineexhaust contains high concentrations of CO, particularly at mixture settingsricher than peak EGT. The most common way for this CO to find its way into thecabin is through the cabin heat system. The vast majority of single-engineaircraft obtain their cabin heat by ducting ventilation air over the surface ofthe muffler. Therefore, when cabin heat is being used, any cracks or holes inthe muffler can allow CO-rich exhaust gas to contaminate the cabin air. Otherpossible causes include inadequate sealing of the firewall, wheel wells, orother air leak that allows exhaust to leak into the cabin.

Normally, oxygen inhaled into your lungs combines with the hemoglobin in thered cells of your blood to form “oxyhemoglobin.” The oxygen is then transportedthroughout your body by your arteries and capillaries, where it disassociatesfrom the hemoglobin and oxygenates the cells of your tissues and organs(including your brain). The deoxygenated hemoglobin then returns through yourveins to your lungs, where it is combines with more oxygen and the cyclerepeats.

CO Concentration
(parts per million)
35 No obvious symptoms after 8 hours of exposure.
200 Mild headache after 2 to 3hours.
400 Headache and nausea after 1 to 2 hours.
800 Headache, nausea and dizziness after 45 minutes; collapse after 2 hours.
1000 Unconsciousness after 1 hour.
1600 Unconsciousness after 30minutes.
Table 1 – Effects of various CO
concentrations at sea level.
(At altitude, the effects of CO
poisoning and altitude hypoxia
are cumulative.)
Saturation (%)
0 – 10 None. (Smoking yields 3% to 10% COHb.)
10 – 20 Tension in forehead, dilation of blood vessels.
20 – 30 Headache and pulsating temples.
30 – 40 Severe headache, weariness, dizziness, vision problems, nausea, vomiting, prostration.
40 – 50 Same as above, plus increased breathing and pulse rates, asphyxiation.
50 – 60 Same as above, plus coma, convulsions, Cheyne-Stokes respiration.
60 – 70 Coma, convulsions, weak respiration and pulse. Death is possible.
70 – 80 Slowing and stopping of breathing. Death within hours.
80 – 90 Death in less than 1 hour.
90 – 100 Death within minutes.
Table 2 – Effects of various COHb saturations.

When carbon monoxide is inhaled, the CO combines with your hemoglobin to form”carboxyhemoglobin” (COHb). The COHb bond is over 200 times stronger thanoxygen’s bond with your hemoglobin. Thus, the CO effectively puts yourhemoglobin “out of commission” and deprives your body of the oxygen it needs tosurvive. The strong COHb bond explains why even very tiny concentrations ofcarbon monoxide can poison you slowly over a period of several hours, and why itmay take a long, long time for your body to eliminate CO buildups from yourbloodstream.

How long? According to an authoritative medical text (Rosen’s EmergencyMedicine, 3rd Ed., 1992), COHb has a “half-life” of more than fivehours for a patient breathing fresh air. In other words, if you crash-landin a hay field with COHb saturation of 40%, your COHb level can be expected todrop to about 20% after five or six hours, to 10% after another five or sixhours, and so forth. If you’re taken to the emergency room and they put you onoxygen therapy, the half-life drops to 1.5 to 2.5 hours (depending on whetherthe docs put you on a ventilator or just use a face mask). In extreme cases ofCO poisoning, you may be rushed to a large medical center and put into ahyperbaric chamber with pure oxygen at three times normal atmospheric pressure,which reduces the half-life to under a half-hour.

According to the FAA Civil Aeromedical Institute, cigarette smoking willnormally produce a COHb saturation of 3% to 10%. Smokers are consequently farmore vulnerable to CO poisoning in flight, since they’re already in apartially-poisoned state when they first get into the aircraft. Because ofCOHb’s long half-life, smokers would do well to abstain from smoking for 8 to 12hours prior to flight. (Unfortunately, the more common scenario is that the lastcigarette is stubbed out on the tarmac moments before flight, and the next oneis lighted seconds after the aircraft comes to a stop at the destination.)

As the CO level in your blood increases, the amount of oxygen transported toyour body’s cells decreases. It is this oxygen deprivation that makes CO sodeadly. Sensitive parts of your body like your nervous system, brain, heart andlungs suffer the most from this lack of oxygen. Symptoms of mild CO poisoninginclude headache, fatigue, dizziness, vision problems (particularly doublevision), nausea, and increased pulse and respiration. Unfortunately, thesesymptoms are often attributed to flu, indigestion, or the common cold. At higherlevels of COHb saturation, you may suffer difficulty in breathing, loss ofconsciousness, collapse, convulsions, coma, and even death.

Just how sick you’ll get from CO exposure varies greatly from person toperson, depending on age, overall health, the concentration of CO (measured inparts per million), and the duration of exposure. High concentrations can causeincapacitation within minutes, but low concentrations can still be extremelydangerous if you’re exposed for a period of hours. As CO continues to beinhaled, the percentage of COHb gets higher and higher, and you get sicker andsicker. Your eyes are particularly vulnerable to the effects of CO poisoning,and permanent damage can easily occur.

Whereas hypoxia tends to make you turn blue (the medical term is “cyanotic”),CO poisoning has the opposite effect – it makes you turn red. Carboxyhemoglobinis red in color, just as oxyhemoglobin is. (That’s why a pulse oximeter isunable to detect CO poisoning.) But, since CO does not disassociate readily fromhemoglobin the way O2 does, your venous blood remains red rather than turningthe normal bluish color. This morbid little fact is useful mostly to coronersand morticians, however, because by the time CO poisoning has progressed farenough to turn you noticeably red, you’re at least comatose if not dead.

The accompanying tables give you some idea of how various levels of COconcentration in the air and COHb saturation of the blood affect an averageperson. As you can see, a CO concentration of one tenth of one percent (1,000parts per million) is enough to render you unconscious in an hour. OSHA hasestablished the maximum permissible CO level for continuous 8-hour-per-dayexposure in the workplace at 35 parts per million. Personally, I would not careto fly in an airplane that exceeded that level.

Chemical spot CO detectors

Chemical spot detectorGiventhe insidious nature of carbon monoxide poisoning and the apparent increase inthe CO-related accident rate, it seems astonishing that so few pilots install COdetectors in their airplanes (particularly piston singles, which are by far themost vulnerable). Furthermore, among those pilots who do use CO detectors,almost all seem to be using those adhesive-backed cardboard chemical spotdetectors that are commonly sold for about $4.00 apiece under tradenames like”DeadStop” and “HeadsUp” by pilot shops and mail-order outfits.

While I suppose these chemical spot detectors are better than nothing, theyleave a great deal to be desired. For one thing, they have a very short usefullife, claimed to be 30 to 60 days (and experts tell me that anything more than30 days is wildly optimistic). Unfortunately, most pilots who use thesedetectors are very bad about replacing them once a month religiously. C’mon,fess up, you know I’m right!

Oh, by the way, if you did replace them once a month, they’d cost you$50 a year!

MaintenanceFurthermore, these chemical spots are extremely vulnerable tocontamination from all sorts of aromatic cleaners, solvents, and other chemicalsthat are routinely used in aircraft maintenance. Read the fine print on thesethings, and you’ll learn that the detectors will be inactivated and damaged bythe presence of ammonia, chlorine, iodine, bromine, and nitrous gases. Itdoesn’t take much, either. One brand of spot detector actually warns that theammonia produced by the presence of a cat litter box in the home may render thedetector unusable! What’s worse, there’s not necessarily any warning that thedetector has been contaminated. The bottom line is that you might easily beflying around with an inoperative detector (because it’s too old orcontaminated) and not know it. In some ways, that’s worse than not having adetector at all.

Finally, the chemical spot detectors are incapable of detecting low levels ofCO. If you’re lucky, they’ll just barely start turning color at 100 PPM, but soslowly and subtly that you’ll never notice it. For all practical purposes,you’ll get no warning until concentrations rise to the 200 to 400 PPM range (andthat assumes a fresh, uncontaminated detector). Even at these levels, it cantake so long for the color change to take place that you could easily becomeimpaired before you notice it. As I said, these things are arguably better thannothing, but not by much.

A better spot: the QuantumEye

Quantum EyeAn improved version of the chemical spot detector – the Quantum Eye -is manufactured by the Quantum Group Inc. in San Diego, Calif. This unit sellsfor about $10 and claims to have a useful life of 18 months (although my expertstell me that 12 months is more realistic). It has an expiration date printedright on its face to help ensure that it won’t be used beyond its time. It alsohas a color reference wheel printed on its face, making it easier to noticesubtle color changes.

The Quantum Eye utilizes a “biomimetic” sensor element, which is essentiallyan artificially engineered chemical whose affinity for CO is as similar aspossible to that of hemoglobin. The idea is that CO binds to this sensormaterial at approximately the same rate that it binds to hemoglobin, and thusthe biomimetic detector will change color even in the presence of fairly lowlevels of CO if the exposure time is long enough. That’s a big improvement overthe four-dollar chemical spot detectors that are basically insensitive to lowlevels of CO.

The Quantum Eye is not without its problems, however. Just as with thecheaper chemical spot detectors, the Quantum Eye is quite vulnerable to exposureto a wide range of aromatic chemicals commonly used around airplanes, such ascleaners and solvents containing alcohol, ammonia or chlorine. Such contaminantshave a cumulative effect that progressively degrades detector performance overtime. Unfortunately, there’s really no good way to determine the degree ofcontamination or degradation. The only real solution is to replace the detectorregularly, and to try to avoid exposure to aromatics.

Another problem with this type of detector is that it can’t distinguishbetween a short exposure to a high concentration of CO and a long exposure to alow concentration – both produce the same color change. To put this intoaviation terms, the detector may be able to warn you that you’re in trouble(assuming you keep it in your visual scan and notice the color change), but itcan’t tell you what the concentration of CO is and therefore how much time ofuseful consciousness you have left. This might be okay in the home environment -where you can call 911 and then run out the door – but it leaves a lot to bedesired in an airplane where running outside is not exactly a viable option.

If you’re thinking of buying a chemical spot detector, the Quantum Eye is theonly one worth considering, in my opinion. But as you’ll see, there are farbetter alternatives available.

Electronic CO detectors…a troubledhistory

In the early 1990s, a number of companies started selling low-cost electroniccarbon monoxide detectors for consumer use. These seemed to offered greatpromise, but their history has been something of a roller coaster ride.

Electronic CO detectorsIn 1992, Underwriter’sLaboratory issued its UL2034 Standard for low-cost residential CO detectors.A number of manufacturers, including American Sensor, BRK Brands (First Alert),and Nighthawk Systems, quickly introduced UL-approved CO detectors priced in the$50 range. First Alert(then a division of Pittway Corporation) ran a massive campaign of “scaretactic” TV ads featuring the basso profundo voice of actor William Conrad(Cannon, Jake and the Fatman), and quickly became the leadingsupplier of residential CO detectors. The industry really took off when the Cityof Chicago mandated the installation of CO detectors in residences beginningOctober 1, 1994.

Chicago Fire DepartmentThen a funny thing happened. By December 20, 1994, the Chicago firedepartment had logged some 8,500 calls of CO detector alarms, and found that 86%of them turned out to be false alarms! Then, on December 21, 1994,Chicago experienced a temperature inversion and consequent smog problem, and allhell broke loose: more than 1,800 calls were made to “911” within 24 hours,almost all of which turned out to be false alarms. Other cities experiencedsimilar problems. Los Angeles recorded some 3,300 nuisance alarms in onemonth.

First Alert CO detectorsBy far the worst false-alarm offenders were the market-leading FirstAlert units. These made use of the same “biomimetic” (color-change) sensortechnology used by the Quantum Eye spot detector. The sensor module used byFirst Alert simply passed a light beam through the biomimetic spot, and alarmedif the light was sufficiently attenuated (presumably because the spot had turneddark in color). Not only did this mean that the First Alert detector shared allthe problems of the Quantum Eye (such as limited sensor life andcross-sensitivity to gases and vapors other than CO), but the detector wasplagued by false alarms due to the fact that other things could attenuate thelight beam (smoke, contamination, even insects that crawled inside the sensormodule).

UL logoIn response to the false-alarm crisis, Underwriter’s Laboratory revised its UL2034 Standard in Juneof 1995, but the false alarm problems didn’t get any better. Meantime, in late1995 and early 1996, the gas utility industry and the Consumer Product SafetyCommission (CPSC) started getting concerned about the very opposite problem: COdetectors that would not go off when they should! While First Alert had obtainedan exclusive license on the biomimetic sensor technology for residential COdetectors, virtually all other detectors sold prior to 1996 (including #2 and #3market leaders American Sensor and Nighthawk) made use of ametal-oxide-semiconductor (MOS) sensor, which was the only other low-cost sensortechnology available at the time. CPSC tests revealed that some of the MOS-basedunits would fail to alarm even at life-threatening CO concentrations of 1,000PPM or more! Many of these units were recalled.

In short, your choice in 1996 was between two sensor technologies, one(biomimetic) plagued by false positives and the other (MOS) plagued by falsenegatives.

Since then, the industry has gone through considerable upheaval. PittwayCorporation wound up divesting itself of First Alert, which subsequently wentpublic, then nearly bankrupt, and finally was acquired by Sunbeam in 1998.American Sensor wound up going bankrupt, while the assets of Nighthawk Systemswere acquired by fire extinguisher giant Kidde Safety who subsequently redesigned the Nighthawkproducts to use a more reliable electrochemical sensor technology.

In 1998, Underwriters Laboratory finally revised its UL 2034 specification,but did so in a fashion that made all UL-approved residential CO detectors farless attractive for aircraft use. UL published its revised spec in 1998, butimplementation was delayed until January 1, 2000. For a CO detector to beUL-approved for residential use after that date, UL requires that it must notindicate CO levels less than 30 parts per million (PPM), nor alarm at levelsbelow 100 PPM. This requirement was imposed by UL at the request of gasutilities and firefighters to minimize the number of unnecessary emergency callsfrom homeowners. I’m sure this has made the firefighters and gas company folksvery happy. But it sure didn’t please me.

That’s because I believe that aviation safety is best served by a sensitivelow-level detector, not one that’s intentionally “blinded” to concentrationsbelow 30 PPM. For in-flight use, we’re not simply worried about high CO levelsthat can make you ill – we care even more about low levels of CO that canproduce subtle cognitive impairment. Furthermore, when flying at altitude in anunpressurized aircraft, we’re already somewhat impaired by altitude hypoxia, soit doesn’t take much CO to increase impairment to a dangerous level. That’sbecause the effects of altitude and CO are additive. The bottom line is that ULdidn’t do pilots any favors with its 1998 revision to UL-2034.

CO Experts 2002, Kidde NighthawkMost residential CO detectors sold today are designed to operate from115-volt AC power, which makes them poor candidates for in-flight use. However,I found several battery-operated CO detectors with digital readouts that canbe used for aviation applications: the Kidde Nighthawk 900-0089, the AIM Model935, the Senco Sensors Model One (BS7860 version), and the CO Experts Model2002. All are powered by alkaline batteries, and incorporate an electrochemical sensor, an 85dB horn, and a largeLCD digital display. (First Alert also offers a battery-operated model, but itlacks a digital readout, so I didn’t bother to test it.)

Unfortunately, since I originally wrote this article, the units from AIM andSenco Sensors (both Canadian manufacturers) became unavailable. AIM changedhands and ceased production of their CO detectors, and Senco decided no longerto permit their BS7860 detector to be sold in North America (apparently onadvice of their lawyers). This leaves only the Kidde Nighthawk 900-0089 and theCO Experts Model 2002 in the running for use in the cockpit.

Kidde Nighthawk 900-0089Kidde Nighthawk 900-0089

Kidde Safety of Mebane,N.C. markets the largest-selling line of home CO detectors in the United Statesunder the “Nighthawk” brand name. The company has been in business for 75 yearsand is best known as the world’s largest producer of fire extinguishers. Thecompany offers a variety of different CO detector models, but most are hardwiredor plug-in models that operate off AC power, and therefore are unsuitable foraircraft applications. Two Nighthawk models are battery-powered, however: the900-0089 (with digital readout) and the 900-0090 (without). Since I consider adigital readout to be absolutely essential, only the 900-0089 digital readoutmodel was tested for this review.

The Nighthawk 900-0089 is housed in a very attractive white plastic enclosure(5.5″ diameter x 1.4″ deep) with a large, easy-to-read LCD digital display inthe center. Above the display is a loud 85 db horn that provides audible alerts.To the left of the display is a Test/Reset button used to silence the warninghorn and initiate the unit’s self-test feature, and a green Operate LED whichflashes briefly once a minute to indicate that the unit is operational. To theright of the display is a Peak Level button used to display the unit’speak-level memory, and a red Alarm LED which illuminates continuously to warn ofdangerous CO levels, or flashes in the event of a malfunction of the unit.Pressing both buttons simultaneously resets the peak level memory to zero.

The LCD display includes three digits that shows CO levels in parts permillion from 11 PPM to 999 PPM, or zero if CO concentration is below detectablelevels. At the left of the LCD display is a three-segment “battery gauge” and anicon that warns if the CO sensor fails.

Kidde Nighthawk 900-0089 openThe lower half of the case swings aside to reveal the CO sensor and abattery holder that accommodates three AA-size alkaline batteries. (The unitcomes with three Energizer E91 batteries that are good for about one year ofoperation.) To the left of the batteries is the replaceable plug-in CO sensormodule. This electrochemical sensor contains platinum electrodes and a padsaturated with an acid electrolyte. In the presence of carbon monoxide, thesensor acts as a battery and generates a voltage proportional to the COconcentration. The sensor output is processed by a microprocessor which drivesthe digital display and decides when to activate the aural and visual alarmsbased on the concentration and duration of detected CO.

The Kidde 900-0089 is available through numerous retail building supply andhardware outlets, and sells for around $50 (batteries included). The replaceablesensor module is warranted for two years, and the rest of the unit (exclusive ofthe battery and sensor) for five years.

CO Experts Model 2002 Low-Level COMonitor

CO Experts Model 2002 low-level CO monitorTheCO Experts Model 2002 low-level CO monitor is not nearly as well-knownas the Kidde Nighthawk because it is not advertised or sold in retail stores,and is principally distributed through natural gas utility companies and heatingcontractors. The unit is housed in a utilitarian-looking rectangular whiteplastic case (6″x3.75″x1.75″), but inside this unpretentiousexterior I discovered the most sensitive and feature-rich functionality of anyunder-$500 CO detector on the market.

The frontpanel features an LCD display that shows CO concentrations from 5 to 70 PPM;levels below 5 PPM display as “000” and levels above 70 PPM display as”HI.” There are two LED lights: the green “power” lightblinks briefly about once a minute to indicate that the unit is functioning,while the red “alarm” light flashes in sync with the 85dB alarm hornto warn of carbon monoxide.

One unique aspect of the CO Experts Model 2002 is that it provides three alarm levels, both visually (flashing redlight) and aurally (85dB horn). The low-level CO warning activates immediatelyat concentrations between 10 and 25 PPM, and is indicated by two quick beeps andflashes every 60 seconds. The medium-level warning activates immediately atconcentrations between 25 and 50 PPM, and is indicated by two quick beeps andflashes every 10 seconds. The high-level alarm activates after 60 minutes ofexposure to 50 PPM, or 15 minutes of exposure to 70 PPM, and is indicated byfour quick beeps and flashes every 6 seconds. The flashing red light andquick-beeping horn are particularly well-suited for in-flight use, since itmakes it virtually impossible to confuse a CO alarm with other audible alarmssuch as stall warning or gear warning horns.

The front-mounted “test/reset” button performs two functions. If pressed while theunit is alarming, the button will silence (“hush”) the alarm for aperiod of time that depends on the CO concentration — 12 hours at 10 PPM, onehour at 25 PPM, 6 minutes at 50 PPM, 4 minutes at 70 PPM. At other times, thebutton activates a recall of the memory (peak level, how long ago, duration, andCOHb%) and a self-test (red LED and horn). Pressing and holding the buttonclears the memory.

An extensive “fail-safe” design monitors the CO sensor,electronics and battery. Any detected failure results in a single”chirp” of the audible alarm and a single flash of the red LED once aminute, with the error condition indicated on the LCD display. Possibleindications are “BAT” for low-battery, “ERR” for a failureof the electronics, and “SENSOR END” for a sensor failure (orend-of-life).

Like the Nighthawk, the CO Experts Model 2002 utilizes an electrochemical COsensor, but the non-replaceable sensor in the CO Experts unit is much bigger andcontains a far greater supply of electrolyte, allowing it to last for at leastfive years under normal conditions. The unitis powered by a single 9-volt alkaline battery, which lasts about a year. TheLCD display indicates when the battery needs to be replaced and when the sensorhas reached the end of its useful life.

The CO Experts Model 2002 sells for about $100, about twice the price of the Nighthawk. For the extramoney, you get a much more sensitive and feature-rich unit, with a sensordesigned to last for five or more years instead of two. Taking thecost of sensor replacement into account, the long-term cost of ownershipof the Nighthawk and CO Experts detectors are approximately the same.

Comparing the CO Experts and Nighthawk for aviation use

CO Experts 2002, Kidde NighthawkI’ve done a lot of comparative product reviews over the years forAVweb, Aviation Consumer, and other publications, but this onepresented some special challenges. After all, I couldn’t simply take these unitsflying like I might do with a handheld GPS or ANR headset, because all thatwould happen is that they’d sit there and blink their little green LEDs at meand read zero. I asked a friend whether he’d mind if I drilled a hole in themuffler of his Skylane and used it as a flying test bed for CO detectors, but heallowed as how he’d rather I didn’t.

So I decided to get creative. For the past month, I’ve been running both ofthese electronic CO detectors through a series of home-brew lab and torturetests designed to determine whether or not they’re worthy of service in thecockpit. I’ve repeatedly exposed them to both high concentrations of CO (my 1985Dodge Caravan will easily produce well over 1,000 PPM of CO when the catalyticconverter is cold), and low concentrations of CO (I found that a smolderingpaper plate in the kitchen sink does a really nice job of producing 50 to 100PPM).

FlameResponsetime tests

Like most residential CO detectors, the Nighthawk updates its digital PPMdisplay only once per minute. In contrast, the CO Experts unit has a six-secondupdate rate, ten times faster than the Nighthawk. When exposed to low levels of CO (like thesmoldering-paper-plate-in-the-sink test), the CO Experts detector starts registeringCO almost immediately while the Nighthawk seems to take at least two minutes tostart registering a non-zero reading. When the CO source is removed (by dousingthe smoldering-paper-plate-in-the-sink with water), the CO Experts reading startsdropping toward zero right away, while the Nighthawk display seems to take quitea few minutes to return to zero.

Whilethis difference in response rate is probably no big deal with respect to thein-home application for which both of these CO detectors were designed, it seemsto me that it is important when it comes to aviation use. If your home CO detector alarms, your course of action is clear: getgrandma and the kids out of the house, call 911, open some doors and windows,and get out yourself. On the other hand, if your detector alerts you to aCO hazard while in flight, you really don’t have the option of stepping outside… at least not right away.

So how do you respond as PIC? Certainly youshut off the cabin heat if it’s on. Then what? Well, maybe opening the cabinvents will help clear the CO out of the cabin … but then again, maybe that’llmake things worse, depending on where the cabin vent system gets its air fromand the CO’s source. How about opening the cockpit storm window … will that makethings better or worse? What about lowering the landing gear? Leaning themixture more aggressively?

See what I mean? If you find yourself in the position of trying to figure outhow to minimize the CO concentration in the cabin long enough to get safely onthe ground, would you rather have a CO detector that has a fast-acting sensor orone that has a slower-acting sensor? Yes, me too! Here’s where the CO Experts’six-second response time can be invaluable.

Temperature and vibration tests

How the units respond to CO is important, but there are other considerationsas well. For instance, because UL only tests these home units over a temperaturerange of 40F to 100F (4.4C to 37.8C), and since the cabin temperature of anairplane stored out-of-doors can experience much wider excursions than that, Iput these units in the freezer and in the oven to make sure that heat or coldwouldn’t make them fail. And, because airplanes can be subject to some prettynasty in-flight bumps and jolts, I put both detectors through a series ofinformal vibration and shock tests – in other words, I shook them, dropped them,hit them, and just generally abused them.

Both units passed my hot-and-cold tests without obvious trauma, althoughnaturally I was unable to test the long-term effects of temperature extremes.Exposure to cold is unlikely to be a problem (at least down to -40F or so), butprolonged exposure to dry heat could shorten the useful life of the sensor bydrying out the electrolyte.

Kidde Nighthawk 900-0089 openThe shock tests were another matter altogether: the CO Experts detectorpassed them with flying colors, while the Nighthawk flunked miserably. TheNighthawk has at least three different areas of vulnerability in this regard,some more serious than others:

  • The swing-out doors of the Nighthawk’s plastic case do not latch securely closed, and can easily vibrate open in flight. This isn’t a serious problem, and could be cured easily with a couple of strategically placed strips of duct tape.
  • The Nighthawk has no positive provision for retaining the three AA-size alkaline batteries in its battery holder. In my judgment, the batteries are vulnerable to vibrating out of the holder during turbulent in-flight conditions, especially as the unit gets older and the spring clips in the holder start to lose some of their tension. Again, this problem is not insoluble, and could be remedied with some foam rubber wedged between the batteries and the lower swing-out door before the door is duct-taped shut. (The CO Experts uses one 9-volt alkaline battery that doesn’t seem to have any vibration problem.)
  • The most serious problem I encountered with the Nighthawk 900-0089 was that virtually any manhandling of the unit would often result in the unit acting as if the “Test/Reset” or “Peak Level” buttons had been pressed. This occurred virtually any time I jolted, shook, dropped, squeezed, or even accidentally brushed an arm or leg against the unit. Often, this caused the horn to chirp loudly, the display to go to “888” and the unit to go through its 30-second cycle before returning to normal. I judged this to be a problem that is very annoying at best and potentially hazardous at worst, and could find no obvious work-around. To me, this problem virtually disqualifies the Nighthawk for in-flight use. (The CO Experts withstood my shake-rattle-and-roll tests without so much as a peep.)

Internal construction

In trying to determine why the Nighthawk was so sensitive to vibration whilethe CO Experts was seemingly immune, I decided to disassemble both units and have agood look at their internal construction. (Don’t try this at home, kids … it’llvoid the warranty.) There is a night-and-day difference between the units.

Kidde Nighthawk internal constructionThe Nighthawk contains four separate electronic subassemblies: the maincircuit board mounted to the rear cover, a separate display board mounted to thefront cover, a replaceable plug-in sensor module, and a spring-terminal-typebattery holder integral to the front cover. A ribbon cable connects the mainboard to the display board, while two lengths of hookup wire run between themain board and the battery holder. This makes for lots of potential problems.The wires and ribbon cable are neither supported nor strain-relieved, sovibration could cause them to fatigue and break (most likely at the solderjoints). The plug-in sensor module is not clamped or safetied, so it couldconceivably vibrate loose from its socket (although the sensor is so light thatit’s not terribly likely). And, as mentioned previously, there’s nothing to keepthe batteries from vibrating loose from their holder.

In contrast, the CO Experts unit seems much more bulletproof inside. Everything ismounted to a single circuit board and mounted rigidly in a plastic case that’smuch thicker than the Nighthawk’s. There are no wires or connectors or socketedICs. So it’s not surprising that it holds up well in a high-vibrationenvironment.


For aircraft use, I think it’s essential to use a sensitive detector capableof displaying and alarming at very low concentrations of CO. In aircraft use,after all, we’re not just concerned about protecting the health of theoccupants, but also about preventing cognitive impairment of the pilot. To makematters worse, low levels of CO can be extremely hazardous in aircraft becausethe effects of CO and of altitude (hypoxia) are cumulative. A low level of COthat you might never notice at sea level could easily make you very woozy ifencountered at a cabin altitude of 10,000 feet. For example, the very mild COpoisoning caused by smoking a cigarette just prior to flight can raise one’sphysiological altitude by 5,000 feet or more.

How much CO is too much? It depends a lot on whom you ask. OSHA (the U.S.Occupational Safety and Health Administration) originally established a maximumsafe limit for continuous exposure to CO in the workplace of 35 PPM, then laterraised it to 50 PPM under pressure from industry. On the other hand, the U.S.Environmental Protection Agency (EPA) issues a health hazard alert when theoutdoor concentration of CO rises above 9 PPM for an extended period, or above35 PPM for one hour. The FAA now requires no more than 50 PPM duringcertification testing of new general aviation aircraft (FAR Part 23), but thevast majority of GA aircraft were certified under the older CAR 3 standard whichinvolved no CO testing, and the FAA requires no regular re-testing of aircraftduring maintenance (although I think it’s high time they did).

Here’s my take: Given the concern about cognitive impairment and theaggravating effect of hypoxia, I consider CO concentrations of 10 PPM or more inthe cockpit to be something worth worrying about, and concentrations above 20PPM to be grounds for landing at the next reasonable opportunity to determinethe cause of the CO contamination. A CO concentration of 35 PPM or more in thecockpit of an unpressurized aircraft should be treated as an emergency: thepilot should go on supplemental oxygen immediately (if available), and make aprecautionary landing as soon as possible.

This is why I don’t care for the 30 PPM minimum display level imposed byUL-2034-1998, and why I think a more sensitive detector such as the CO Experts Model 2002 is a much better choice for cockpit use.(Frankly, I’d rather have a more sensitive detector in my home, too, regardlessof what the gas company wonks and firefighters may think.)

CO Experts Model 2002 low-level CO monitorRecommendation

After researching the subject rather extensively, I’ve concluded thatthere’s basically no contest. The only currently available CO detector (under $100) I’d trust in my airplane is the CO Experts Model 2002Low-Level CO Monitor. It employs by far the best technology available,and appears to be built to “aircraft quality” standards. It offers five years ormore ofthe most reliable and sensitive CO protection available for about $20 a year. This is actually cheaper by far than those cardboard spotdetectors that are, in my opinion, leave much to be desired.

At present, the only viable challenger appears to be the Kidde Nighthawk900-0089, which is far better known but in my opinion not nearly as good. I consider this unitadequate for residential use, but I honestly can’t recommend it for aircraft orvehicle use where it’s exposed to vibration, because it’s just not built to takeit. The Nighthawk is also blind to CO concentrations below 30 PPM, while the COExperts starts to display at 5 PPM and provides a low-level alarm at 10 PPM.

While Kidde Nighthawks and First Alerts can be found ubiquitously in almostevery hardware and building supply store (as well as K-Mart, Wal-Mart, etc.),the CO Experts units are not nearly as well known or easy to find. You won’t find these units advertised or sold it in retail stores, which is why nobodyever heard of them until we started talking about them on AVweb. None ofthe big pilot supply houses offer these units, and that’s a pity, because I’mconvinced that this device belongs in every cockpit.

AeromedixYou can purchase the CO Experts Model 2002 through your localheating contractor, or buy it online through

Personally, I have four of these CO Experts units: two for the house, one forthe airplane, and another for the car. Something to consider.

Some Usage Notes

Many readers have asked me for advice concerning where and how to mount a COdetector in the aircraft, and how to respond if the alarm goes off in flight.Neither of those are easy questions, but let me take a crack at answeringthem.

Mounting in an aircraft

First of all, keep in mind that no under-$100 electronic CO detector (includingthe CO Experts model that I recommend) is TSO’d, STC’d, or otherwise approved foraircraft use. (There is at least one FAA-approved panel-mount CO detectoravailable, but it’s a lot more expensive and much less sensitive.) Consequently, it’sasking for trouble with the FAA for such a device to be”permanently installed” in the airplane. This is an issue similar to the one youface with a handheld GPS or other non-TSO’d portable equipment. As an A&Pmechanic, I’d recommend mounting the detector with Velcro or some such, ratherthan bolting it to some structural component. I’ve had excellent luck withhigh-strength hook-and-loop mounting strips such as Radio Shack “SuperlockStrips” (part number 64-2360, price $2.99). Do not let the unit lie aroundloose in the cockpit, since it weighs about a pound and could easily become aprojectile in moderate turbulence or worse.

Where’s a good place to mount the detector? Well, the answer depends on yourparticular cockpit layout. The detector folks suggest that younot mount it directly in front of a heater or ventilationoutlet, and that you not mount it in some blind corner where theair is likely to be completely stagnant. In other words, they’d prefer to see itin an area of moderate airflow. It doesn’t really matter whether you mount it uphigh or down low; although CO is slightly lighter than air, it mixes with airso thoroughly that the concentration is likely to be relatively constantthroughout the cabin. What’s most important is to find a mounting locationthat’s readily visible to the pilot and which doesn’t obstruct the pilot’s viewof something else important. That leaves a lot of latitude. In single-engineCessnas, for example, the center pedestal may offer a good mounting site. Inlow-wing aircraft with a single door on the right, the left sidewall may be aworkable location. The floor between the front seats is another goodpossibility. Even the ceiling is possible, although I’m not sure I’d want totrust hook-and-loop mounting strips for that. You’ll simply have to study your cockpit layout and comeup with an optimum location.

Since electronic CO detectors like the CO Experts and Nighthawk use an internalmicroprocessor, you’d also be well advised to check carefully for interferencewith on-board avionics. I’ve not heard any reports of such problems, but commonsense dictates that you be cautious until you’re sure there’s no problem in youraircraft.

What to do if the alarm goes off in-flight

Please don’t wait until the alarm goes off! Look at the digital display fromtime to time.

In order to avoid false alarms, UL-approved CO detectors like the ones from Kiddeand First Alert do not alarm until they think you’vebeen exposed to enough CO for a long enough time to raise your COHb level tosomewhere between 5% and 10%. If you get to this point while flying an airplane,you’re already in deep kimshee! By including the digital readout in your scanfrom time to time, hopefully you’ll discover a CO problem long before it getsserious enough to create a potential crisis. That’s why I consider it soimportant to choose a digital-readout unit detecting low levels of CO … andwhy I’m not at all happy about the UL requirement that prevents residentialdetectors from even displaying CO concentrations below 30 PPM.

If the display starts reading 10 PPM orgreater in-flight, you’ve got a problem that your mechanic needs to look into.If the concentration reaches 35 PPM in-flight, get down now!

Anytime you have a CO problem, the first thing to do is to shut off the cabinheater, which for single-engine airplanes is the predominant source of COcontamination. Twins typically use combustion heaters, which can create a COproblem in unpressurized twins but are very unlikely to do so in a pressurizedtwin.

The second thing to do is to start breathing supplemental oxygen if you haveit aboard, and turn up the flow to its maximum level. Going on O2 will reduceyour intake of CO, and will increase the rate at which the COHb level of yourblood will dissipate.

The third thing to do is to start making plans to land at the earliestpossible opportunity. That’s particularly important if you have any symptoms ofCO poisoning or hypoxia, such as headache, nausea, or double vision. Don’t justkeep on flying and hope that you’ll feel better. Remember that it can take many,many hours for the effects of CO poisoning to abate, even if you’re breathinguncontaminated air or supplemental oxygen.

A fourth step would be to make sure your engine is leaned aggressively tominimize the CO content of the exhaust (remembering that CO is produced byincomplete combustion of fossil fuels). This trick is especially usefulwhen high levels of CO are detected during ground operations. Many pilots taxiaround with the mixture full-rich, and that’s like driving a car with the chokefull on. (Am I dating myself?) Over-rich mixtures result in CO-rich exhaust.Lean for maximum RPM rise at idle and the CO level will plummet – and yourengine will stay cleaner and won’t foul its sparkplugs.

Beyond that, you’ll have to experiment. Whether opening the fresh air ventswill help or hurt is something that depends on the design of your aircraft’sventilation system and the source of the CO contamination. With afast-responding digital CO detector, you can try various ideas for ventilatingthe cabin and see whether they seem to help or hurt CO-wise.

But the main thing is to get it on the ground ASAP and then sort thingsout.