As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
Most home carbon monoxide alarms use one of two sensing elements. Electrochemical cells produce a small current proportional to CO concentration; metal-oxide semiconductor sensors change resistance as CO adsorbs on a heated tin dioxide surface. Both are blind to smoke, natural gas and propane, and both are checked by a test button that only proves the buzzer and battery circuit work, not the sensor. The listing for the Kidde COBD states an 85-decibel alarm and a backlit display but does not publish the sensor type or the alarm thresholds. Multi-gas monitors such as the Amazon Smart Air Quality Monitor report a CO reading alongside PM2.5, VOCs, humidity and temperature, but a displayed number is not the same as a certified life-safety alarm.
The two sensing chemistries inside a home CO alarm
Almost every residential carbon monoxide alarm sold today contains one of two sensor types, and the difference matters because it changes how the device behaves in humidity, in cold, and over its service life.
Electrochemical cells. A small cell holds an electrolyte between two electrodes. Carbon monoxide arriving at the sensing electrode is oxidised, releasing electrons. The resulting current is tiny, measured in nanoamps or microamps, but it is proportional to the CO concentration, so the circuit can convert it into a parts-per-million reading. These cells are selective: they respond to CO and largely ignore the other gases in a home. They also drift as the electrolyte ages, which is why the alarm has an end-of-life signal and why the cell is sealed and not user-replaceable.
Metal-oxide semiconductor sensors. A bead of tin dioxide is heated to several hundred degrees Celsius. In clean air it has a high resistance. When CO molecules adsorb on the surface they donate electrons, the resistance falls, and the circuit reads that change. These sensors are cheap and robust but less selective: humidity, temperature and other reducing gases shift the baseline, so the firmware has to compensate. They also consume more power, which is why they appear more often in mains-powered units than in battery-only ones.
Neither type is a smoke sensor. An ionisation or optical smoke chamber responds to particles, not to a colourless gas. This is the reason a combined smoke and CO alarm is two sensors in one housing, not one sensor doing two jobs. The listing for the Kidde COBD does not publish which sensor chemistry it uses, so a buyer cannot tell from the bullets alone whether it is electrochemical or metal-oxide.
Why the alarm waits: exposure arithmetic and the threshold tables
A carbon monoxide alarm does not sound at the first molecule it sees. It integrates concentration over time, because the harm from CO is a dose effect: haemoglobin binds CO about 200 to 250 times more readily than it binds oxygen, and the proportion of haemoglobin occupied by CO (carboxyhaemoglobin) rises with both concentration and exposure duration.
The thresholds most residential alarms follow are set out in standards such as UL 2034 and its equivalents. The pattern is that the alarm must not sound too early at low concentrations, because cooking and poorly maintained appliances can produce small transient readings, and must sound quickly at high concentrations. A typical table looks like this:
| CO concentration | Alarm must not sound before | Alarm must sound before |
|---|---|---|
| 70 ppm | 60 minutes | 240 minutes |
| 150 ppm | 10 minutes | 50 minutes |
| 400 ppm | 4 minutes | 15 minutes |
The exact times vary by standard revision and by manufacturer, and the listing bullets for the products here do not publish their threshold table. What the table explains is the everyday behaviour that puzzles owners: a detector can sit quietly next to a source that is producing a low but real concentration, then sound an hour later. It is not faulty. It is doing the arithmetic.
This is also why a displayed number from a multi-gas monitor such as the Amazon Smart Air Quality Monitor should not be read as an alarm threshold. The listing states that it tracks CO among five factors and gives an air quality score, but it does not publish the CO alarm threshold, the averaging period, or the sensor chemistry. A monitor that reports 9 ppm and an alarm that stays silent at 9 ppm are not contradicting each other; they are answering different questions.
What the test button proves, and what it does not
Pressing the test button on a CO alarm runs a self-check of the battery, the buzzer, and the alarm circuit. It does not expose the sensor to carbon monoxide. On an electrochemical cell, the only way to prove the cell still responds is to apply a known gas concentration, which is a calibration procedure, not a button press.
The distinction matters because a button test that produces a loud sound feels like proof. It is proof of the sounder, not of the sensor. A cell that has dried out or a metal-oxide bead whose heater has failed can still pass a button test. The manufacturer’s end-of-life signal exists because the sensor degrades silently; the button cannot see that degradation.
the Kidde Smart Carbon Monoxide Detector and Indoor Air Quality Monitor adds a voice announcement that names the hazard, and the listing states it will say “Replace Alarm” when the unit reaches its end of life. That is a useful honesty feature: it tells the owner the sensor has aged out rather than letting the unit fail quietly. But the voice alarm is still not a calibration. The listing does not publish a calibration interval or a field calibration method.
For readers who want to know how a real calibration works, the procedure is covered in how to calibrate a gas detector. The short version: you apply a certified gas of known concentration, wait for the reading to stabilise, and adjust the instrument until it agrees. A button test skips every one of those steps.
Placement, cross-sensitivity and the gases a CO alarm ignores
Carbon monoxide mixes freely with air, so a CO alarm does not need to be at floor level or at ceiling level the way a propane or natural gas detector does. It does need to be near the sleeping areas, because the risk is greatest while people are asleep and unable to notice symptoms. The listing for the Kidde COBD states that it can be wall-mounted or set on a table using the included hardware, and that it runs on two AA batteries, which means it can be placed without wiring. That portability is useful for travel, but it also means placement is entirely the owner’s decision, and a detector left in a suitcase detects nothing.
The cross-sensitivity question is the one buyers ask most often. A carbon monoxide alarm is not a natural gas alarm. Natural gas (methane) and propane are detected by combustible gas sensors, which work on a different principle, usually catalytic oxidation or infrared absorption. The two alarms are not interchangeable, and a CO alarm will not warn you of a gas leak. This is explained in gas leak detector versus carbon monoxide detector and in does a carbon monoxide detector detect natural gas.
Electrochemical CO cells can also respond weakly to hydrogen, and metal-oxide sensors respond to many reducing gases. In a home the concentrations are usually too low to matter, but in a workshop or near a battery bank the cross-sensitivity can cause nuisance alarms. The product listings here do not publish cross-sensitivity figures, so a buyer with a specific interference concern should treat that as unknown.
Where a CO detector stops being the right instrument
A residential CO alarm is a life-safety device. It is not a diagnostic instrument, and it is not a compliance tool for a workplace. If you need to know whether a boiler is producing CO at 35 ppm or 70 ppm, or whether a flue is spilling, an alarm that sounds somewhere between those values is not enough. You need a CO meter with a published accuracy, a known calibration interval, and a display that resolves the concentration you are measuring. The difference is set out in how to choose a carbon monoxide meter and in best carbon monoxide meters.
There is also a category boundary inside the smart monitor space. A device such as the Kidde Smart CO and IAQ monitor combines a CO alarm with an indoor air quality sensor and app alerts. That is useful for a homeowner who wants one plug-in device for both jobs, and the listing states it is a 2-in-1 unit with voice and app alerts. But the IAQ side is a trend indicator, not a calibrated measurement, and the listing does not publish the CO alarm threshold or the sensor type. A multi-gas monitor such as the Amazon Smart Air Quality Monitor is even further from a life-safety alarm: it reports CO alongside PM2.5 and VOCs, but the listing does not state that it carries a CO alarm certification, and a notification on a phone is not the same as an 85-decibel sounder in a hallway.
Finally, a CO alarm does not detect smoke, and a smoke alarm does not detect CO. If you are deciding between the two, the answer is that you need both, and the combined units exist because people forget to maintain two separate devices. The maintenance burden is the real reason to combine them, not any sensing advantage.
The one sentence to remember
A carbon monoxide detector is a dose integrator with a buzzer, not a gas analyser: it waits, it accumulates, and it sounds when the arithmetic says the exposure has become dangerous, which is why the button test proves the buzzer and not the sensor.
Frequently Asked Questions
Does a carbon monoxide detector detect natural gas or propane?
No. A CO alarm responds to carbon monoxide. Natural gas (methane) and propane require a combustible gas detector, which uses a different sensor. A CO alarm will not warn you of a gas leak, and a gas leak detector will not warn you of CO. See does a carbon monoxide detector detect natural gas.
Why does my CO alarm wait before sounding?
Because the harm from CO depends on concentration multiplied by time. Residential alarms follow threshold tables that deliberately delay at low concentrations and respond quickly at high ones. A delay at 70 ppm is normal behaviour, not a fault. The listing bullets for the products on this page do not publish their threshold table.
Does the test button prove the sensor works?
No. The test button checks the battery, the sounder and the alarm circuit. It does not expose the sensor to CO. Only a calibration with a known gas concentration proves the sensor still responds. The button is a maintenance check, not a calibration.
Can a smart air quality monitor replace a CO alarm?
Not on the evidence in the listings. A multi-gas monitor such as the Amazon Smart Air Quality Monitor reports CO among several factors, but the listing does not state a CO alarm certification, a threshold table or an 85-decibel sounder. A phone notification is not a life-safety alarm. Use a certified alarm for protection and a monitor for information.
How long does a CO sensor last?
Electrochemical cells degrade over time and most residential alarms carry an end-of-life signal. The listings here do not publish a service life figure. the Kidde Smart CO and IAQ monitor states that it will announce "Replace Alarm" when the unit reaches its end of life, which is the honest way to handle a sensor that cannot be recalibrated by the owner.
Where should a CO alarm be placed?
Near sleeping areas, because the risk is greatest while people are asleep. CO mixes with air, so floor level versus ceiling level matters less than proximity to bedrooms. the Kidde COBD can be wall-mounted or set on a table, so placement is the owner's decision; a detector left in a drawer protects nobody.