As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
The test button on a carbon monoxide alarm exercises the circuit board, the sounder and the display, and on many models it also checks the battery. It does not expose the electrochemical sensor to carbon monoxide, so it cannot confirm that the sensor still responds to gas. A passing button test means the alarm can make noise; it does not mean the alarm can detect. Real verification needs either a calibrated test gas applied to the sensor or a manufacturer-supplied test mode that simulates a sensor output, and most household alarms offer neither to the owner. For a home, the practical routine is to press the button as the manufacturer instructs, keep the unit clean and within its stated service life, and replace it when the end-of-life signal sounds. A battery-powered Kidde CO detector illustrates the type where the button check is the only user test available.
What the test button is actually wired to
A carbon monoxide alarm contains three things that can fail independently: the sensor, the electronics that interpret the sensor, and the sounder that wakes you. The test button is wired to the second and third. Pressing it injects a simulated signal into the alarm circuit, or simply commands the sounder to run, depending on the model. The sensor is left alone.
This matters because the sensor is the part that degrades. Most household CO alarms use an electrochemical cell: a small pouch of electrolyte behind a diffusion membrane, with electrodes that produce a tiny current when carbon monoxide reaches them. The cell consumes itself slowly over its life, and its output falls. A button test cannot see that decline. The alarm can pass every button press for years and still be deaf to gas by the end.
The listing for the Kidde Carbon Monoxide describes a test/hush button that enables quick checks or silencing alerts. That is the standard arrangement: one button, two jobs, and the check is of the alarm’s own circuitry. The listing does not claim the button verifies the sensor, and no manufacturer of a consumer alarm does.
Why a button test feels more convincing than it is
The alarm is loud, the display lights up, and the unit announces itself. That is a satisfying result, and it is easy to read as proof. But the same outcome would occur if the sensor were removed entirely and the button pressed. The test is a check that the alarm can alarm, not that it can sense.
There is a second confusion, between the button test and the alarm’s own self-diagnostics. Many units run continuous internal checks, and some display a fault code or chirp a specific pattern when the sensor drifts out of range. Those diagnostics are more useful than the button, but they are still not a gas test. They catch a sensor that has failed outright or drifted badly; they do not measure response at a known concentration.
A third confusion is the test button on a combination unit. The Kidde smart CO and indoor air quality monitor announces hazard type by voice and reports to an app. A button press there may cycle the voice and the display, which tells you the speaker and screen work. The gas-sensing element is still untested.
The three tests a CO alarm can receive, ranked by what they prove
Not all tests are equal. The table below sets out what each one establishes, what it leaves unknown, and who can perform it.
| Test | What it exercises | What it does not prove | Who can do it |
|---|---|---|---|
| Button press | Circuit, sounder, display, often battery | Sensor response to gas | Owner |
| Manufacturer self-test or diagnostic mode | Sensor output within an expected band, electronics | Accuracy at a stated concentration | Owner, if the model offers it |
| Calibrated test gas at the sensor | Response to a known CO concentration | Long-term drift after the test | Technician with gas and adapter |
The third row is what a professional does with a carbon monoxide meter, not with a household alarm. A meter is a measuring instrument; an alarm is a threshold device. The two are built differently and tested differently. If you want to know the concentration in a room, you need a meter, which is a different purchase from an alarm. The site’s comparison of a CO meter against a CO alarm sets out that distinction.
For most homes, the honest position is that the button test plus the unit’s own diagnostics plus replacement at end of life is the whole of the user-serviceable testing. Anything more requires gas and a trained hand.
How to run the button test without creating a false record
Press and hold the test button as the manufacturer’s instructions specify, usually until the alarm sounds, then release. Note what happened: did the sounder reach full volume, did every display segment light, did the unit return to normal without a lingering fault. Write the result down with the month, because the value of a test log is in the pattern, not the single event.
Do not press the button and then tell yourself the detector is tested for the year. The button test is a weekly or monthly habit on many manufacturers’ schedules, and it is cheap to repeat. The annual task that actually matters is different: check the manufacture date on the back, confirm the unit is inside its stated service life, vacuum the vents, and confirm the end-of-life signal has not been ignored.
Battery models add one more step. The Kidde Carbon Monoxide runs on two AA batteries and the listing states they are included. A button test on a battery unit may draw on the battery, so a weak battery can produce a weak or failed test. Replace batteries on the schedule the manufacturer gives, not when the chirp starts.
Where the test question stops being the right question
Owners often ask how to test a CO detector because they want assurance the thing will work in an emergency. The better question is whether the detector is the right instrument for the space. A CO alarm is a life-safety device for occupied rooms; it is not a diagnostic tool for a boiler, and it is not a substitute for a combustible gas detector. If the worry is a gas leak rather than carbon monoxide, a CO alarm will not help, as the site explains in whether a CO detector detects natural gas.
Nor is a CO alarm a confined-space monitor. A single gas detector for CO does not confirm that a space is safe to enter, because it measures one hazard and ignores oxygen deficiency and flammable gas. If the work is entering tanks or pits, the instrument is a multi-gas detector with a pre-entry test, not a household alarm with a button.
For a landlord or a technician who needs to document that an alarm responded to gas, the answer is a calibrated CO meter and a cylinder of test gas, applied at the sensor with the manufacturer’s adapter. That is a different class of equipment and a different procedure. The carbon monoxide meters page covers what those instruments publish and what they leave out.
Choosing an alarm whose testing you can actually live with
If you are buying rather than testing, favour units whose diagnostics you can read. A digital display that shows the sensed level, as on the First Alert Plug Carbon and the Kidde 3rd, gives you more than a beep: it shows whether the sensor is tracking zero in clean air and whether a fault is being reported. The listing for the First Alert Plug Carbon states an electrochemical sensor and a peak level feature that displays the highest concentration measured. The listing for the Kidde 3rd states a digital LED display showing the level the alarm is sensing. Neither listing publishes an accuracy figure, and neither claims the display is a calibrated measurement; treat the number as an indicator, not a reading.
Battery backup changes the testing routine. The First Alert Plug Carbon and the Kidde 3rd both plug in and carry a battery backup, so the button test can be run with mains power present and the backup checked separately. The Kidde Carbon Monoxide is battery-only, which makes the battery the single point of failure and the test more important.
Smart units add remote status but not remote gas testing. The WiFi CO monitor with app alerts can tell you the alarm is powered and connected; it cannot tell you the sensor responded to gas. Do not let an app green tick replace the button press and the end-of-life check.
Whatever you choose, the test regime is the same: button as instructed, vents clean, date checked, replace at end of life. The carbon monoxide detector roundup covers the models in more detail, and the plug-in CO detectors page covers the mains-powered group.
The test button is a check, not a certificate
A carbon monoxide detector test button confirms the alarm can make a noise and run its electronics. It does not confirm the sensor can still see gas. Treat the button as one part of a routine that includes clean vents, a known manufacture date, a fresh battery where fitted, and replacement at the end of service life. If you need proof of response to gas, that is a job for a calibrated meter and test gas, not a household alarm.
Frequently Asked Questions
Does pressing the test button prove my CO detector will detect carbon monoxide?
No. The button exercises the circuit, sounder and display, and often the battery. It does not expose the electrochemical sensor to carbon monoxide, so it cannot show that the sensor still responds to gas. A unit can pass the button test and still be near the end of its sensing life.
How often should I test a carbon monoxide alarm?
Follow the manufacturer's schedule for the model. Many specify a weekly or monthly button press. The annual tasks are separate: check the manufacture date, clean the vents, confirm the unit is inside its stated service life, and replace the battery on schedule on battery models.
Can I test a CO alarm with a gas stove or a car exhaust?
No. Those produce carbon monoxide along with other combustion products, and the concentration at the sensor is uncontrolled. You can set off the alarm, but you cannot tell whether it responded at the right threshold, and you expose people to gas. Use the button test and the manufacturer's diagnostics, and leave gas testing to a technician with calibrated gas.
What does the end-of-life signal mean, and can I ignore it?
It means the sensor has reached the end of its service life and the unit can no longer be relied on to detect gas. Do not ignore it. Replace the alarm. The listing for the Kidde Smart Carbon states that the voice alarm announces Replace Alarm at end of life, which is one of the clearer ways a unit communicates this.
Is a digital display on a CO alarm a measurement I can trust?
Treat it as an indicator, not a calibrated reading. The listings for the First Alert Plug Carbon and the Kidde 3rd state a digital display showing the sensed level, but neither publishes an accuracy figure. A CO alarm is a threshold device; if you need a number you can act on, that is a carbon monoxide meter, which is a different instrument.
Do I need a professional to test my home's CO alarms?
Household alarms are designed for owner testing by button and by the unit's own diagnostics. A professional with calibrated test gas is needed when you want to verify response at a known concentration, for example in a rental property or after a suspected incident. That service uses a meter and gas, not the alarm's button.