As an Amazon Associate, Akermin earns from qualifying purchases. Product links on this page are affiliate links.
Start with power: a battery unit with a flat cell and a mains unit with a tripped circuit both look identical from the front. If power is present and the unit still will not respond to its test button, treat the sensor as suspect. Electrochemical CO sensors consume themselves over their service life and cannot be revived; the listing for the Kidde COBD battery detector and similar units carries an end-of-life signal for exactly this reason. A detector that has reached end of life is not a detector that needs a new battery. It is a detector that needs replacing. Anything that produces a number on a screen, such as the First Alert plug-in alarm with digital display, can still be lying to you if the sensor has drifted, which is why the test button and the peak-level memory are different kinds of evidence.
What the test button actually proves, and what it does not
Every residential CO alarm has a test button, and almost every owner presses it once a month and treats the beep as a clean bill of health. That is a misunderstanding of the circuit. The test button on a typical unit, including the ones described in the listings here, exercises the alarm sounder and the electronics behind it. It does not expose the sensor to carbon monoxide. It cannot tell you whether the electrochemical cell still responds to gas, because no gas is present during the test.
The mechanism matters. An electrochemical CO sensor is a small fuel cell: carbon monoxide reaches a catalytic electrode, is oxidised, and the resulting current is proportional to concentration. The cell consumes a little of its electrolyte and electrode activity over its life, and its output falls. A sensor that has lost most of its sensitivity will still pass a sounder test, because the sounder is not the sensor. This is why the test button is a check on the alarm, not on the detection.
What does prove the sensor is alive is the unit’s own self-diagnostic and its end-of-life signal. The Kidde smart detector listing describes a voice message that announces “Replace Alarm” when the unit has come to its end of life span, which is the manufacturer telling you the sensor has run out of usable life rather than the battery running out of charge. If your unit is silent, unplugged from its battery, and will not respond at all, the test button gives you nothing; you need to establish power first.
Why a detector can be powered, silent, and still not detect
The failure that catches people is the one with no symptom. A unit with a good battery, a steady green LED and a working sounder can still have a sensor that no longer responds to CO. Residential electrochemical cells have a finite service life, and the listing for the Kidde plug-in unit with 9V backup describes LED status indicators and a digital display but does not publish a sensor life figure; Not stated in the listing bullets. The general figure quoted across the category is five to seven years for the sensing element, which is why manufacturers print a replacement date on the back of the housing.
There is a second silent failure: the alarm is mounted where CO never reaches it. Carbon monoxide is slightly lighter than air and mixes readily, so a unit on a shelf behind a curtain, in a closed cupboard, or directly above a cooker hood that extracts air will see lower concentrations than the breathing zone. The listing for the Kidde COBD describes wall mounting or tabletop use and a travel-ready design; a tabletop unit is fine for a hotel room, but in a home it should sit where air circulates and where you sleep, not in a drawer.
A third cause is a unit that has been silenced and forgotten. The hush function on many alarms suppresses the sounder for several minutes while the hazard persists. If someone pressed hush during a real event and then closed the door, the detector is working and the alarm is muted. The listing for the Kidde COBD describes a test and hush button, and the hush state is one of the few failures that is genuinely the user’s, not the instrument’s.
The chirp, the blank screen and the false alarm: reading the symptom
Different symptoms point at different subsystems, and the order you investigate them should follow the symptom rather than the calendar.
| Symptom | Most likely cause | What to do |
|---|---|---|
| No sound, no light, no display | No power: flat battery or unpowered outlet | Replace the battery or test the outlet with another device |
| Regular chirp with a working display | Low battery or end-of-life warning | Check the manual for the chirp pattern; replace the battery, then the unit if the chirp persists |
| Alarm sounds with no obvious source | Real CO, or a sensor at end of life, or a hush that failed | Treat as real until proven otherwise: ventilate, leave, call the gas emergency line |
| Display shows a number but the alarm never sounds | Concentration below the alarm threshold, or a drifted sensor | Compare with a second instrument; a number on the display is not an alarm |
| Unit will not respond to the test button | Power, or a failed circuit board | Replace the battery; if it still will not test, replace the unit |
The blank screen is usually the easiest to diagnose because it is a power symptom. The listing for the First Alert plug-in alarm with digital display describes a battery backup that keeps monitoring during a power failure, so a unit that goes dark when the mains drops has a failed backup battery, not a failed sensor. A unit that is dark with the mains present has lost its supply.
The false alarm is the hardest, because the correct response to an alarm is always to treat it as real. A CO alarm that sounds repeatedly at the same time of day, with no fuel-burning appliance running, is more likely a failing sensor than a leak, but you cannot know that from the alarm alone. The honest answer is that a residential alarm is not a diagnostic instrument; it is a trip. To find out whether CO is present you need a meter that reads concentration, which is a different class of device, as explained in carbon monoxide meter vs CO alarm.
What a replacement has to do that the old unit did not
If the unit is past its sensor life, the replacement decision is about how the alarm reports, not about whether it detects. All four units described here use an electrochemical sensor and an 85 decibel sounder, and the listings do not publish sensitivity or response time figures; Not stated in the listing bullets. The differences are in power, display and connectivity.
A battery unit such as the Kidde COBD works during an outage by definition, because it has no mains connection to lose. A plug-in unit with a backup battery, such as the First Alert plug-in alarm or the Kidde plug-in with 9V backup, works during an outage only for as long as the backup cell holds, and a backup cell that is never tested is the most common reason a plug-in alarm is silent during a storm. A smart unit such as the Kidde smart detector adds voice and app alerts, which matter if the alarm is in a basement or a holiday let where nobody is standing next to it.
The display is worth understanding before you pay for one. A digital readout shows the concentration the sensor is currently reporting, and the peak-level feature on the First Alert plug-in alarm retains the highest level measured. That is useful after an event, because it tells you whether the alarm saw a trace or a dangerous concentration. It is not a substitute for a meter, and it is not a calibration certificate. The listing does not publish the display’s resolution or accuracy; Not stated in the listing bullets.
The maintenance nobody does, and the point where the alarm is the wrong instrument
The maintenance that matters is not the monthly button press. It is the replacement date. Every electrochemical CO alarm has a sensor with a finite life, and the housing carries a date by which the unit must be replaced regardless of whether it still beeps. Owners routinely keep units for a decade because the sounder still works, which is exactly the failure mode described above: a working alarm attached to a dead sensor. If your unit has no date and you cannot remember when it was installed, treat it as due.
The second piece of maintenance is the backup battery in a plug-in unit. A 9V cell that has sat in a plugged-in alarm for years is not being exercised, and it will be flat when the mains fails. The listing for the Kidde plug-in unit describes a 9V battery backup; the cell should be replaced on the schedule in the manual, not left until the chirp.
Where the category stops being the right answer: a residential CO alarm is a trip, not a measurement. It is not designed to tell you the concentration, to find a small leak, or to clear a property for reoccupation. Those tasks need a calibrated CO meter with a published range and accuracy, which is a different instrument at a different level of care, as covered in best carbon monoxide meters. A CO alarm also does not detect natural gas, which is a different hazard with a different sensor; the distinction is set out in gas leak detector vs carbon monoxide detector and in does a carbon monoxide detector detect natural gas. And a CO alarm is not a substitute for a smoke alarm, which responds to a different product of combustion.
What to do before you buy a replacement
Work through power, then the chirp pattern, then the replacement date, in that order. If the unit is within its service life and responds to its test button, the problem is more likely placement or a hush state than a failed sensor. If it is past its date, or it will not test with a known-good battery, replace it and do not try to revive it. When you do replace, choose on power source and on whether you need remote alerts, not on the number of features; the detection itself is the same electrochemical method across the units described here. For a wider comparison of plug-in options, see best plug-in carbon monoxide detectors.
Frequently Asked Questions
My detector chirps even after I replaced the battery. What does that mean?
A chirp that persists with a fresh battery is usually the end-of-life signal rather than a low-battery signal. The two can sound similar, so check the manual for the pattern. If the unit is past its replacement date, it is telling you the sensor has expired and the unit should be replaced, not repaired.
Can I test a CO detector with a gas stove or a car exhaust?
No. Deliberately producing carbon monoxide indoors to test an alarm is dangerous, and the concentration is unpredictable. The test button checks the sounder and electronics, and that is the only test a homeowner should perform. Sensor verification requires calibrated gas and is a job for a service technician.
Why does my plug-in alarm go silent during a power cut?
Because its backup battery is flat or missing. A plug-in unit with a backup cell, such as the units described here, monitors during an outage only while that cell holds. Replace the backup battery on the schedule in the manual rather than waiting for the chirp.
The display shows a number but the alarm never sounds. Is it broken?
Not necessarily. Residential alarms are designed to sound at sustained concentrations above a threshold, not at every reading. A low number that never rises may be normal. But a display number is not proof the sensor is accurate, and the listing does not publish the display's accuracy; if you need a measurement, use a calibrated meter.
How long does a carbon monoxide detector last?
The sensing cell has a finite life, and the replacement date is printed on the unit. The listings here do not publish a service life figure; Not stated in the listing bullets. The general category figure is five to seven years for the sensor, which is why the date on the housing, not the beep, is the thing to check.
Does a CO detector also detect a gas leak?
No. A carbon monoxide alarm responds to CO, a product of incomplete combustion. Natural gas and propane are different gases and need a different sensor. The two hazards are covered separately in gas leak detector vs carbon monoxide detector.