How to Test Carbon Monoxide Detector? | The Button Test Isn’t Enough

That first tap tells you the alarm circuitry works, but it doesn’t prove the sensor will detect CO gas. For system-connected units wired into a commercial fire alarm panel, the bar is much higher: those need functional testing with real CO gas at least once a year. Here’s the exact routine for each type, plus the methods you should never use.

How the Built-in Test Button Works

Every home CO alarm on the US market includes a test button. Pressing and holding it — typically for a few seconds — makes the horn sound and the LEDs flash. The Honeywell CO1224T manual spells out the sequence: after power-up, wait 80 seconds for the detector to stabilize, then press and hold the test button for 1-4 seconds. If the horn beeps twice in a Temporal 4 pattern and the LEDs light up, the unit is operational.

The catch: that test verifies the alarm circuit, the horn, and the battery connection — not the electrochemical sensor itself. A detector can pass its button test while its sensor has degraded to the point it won’t respond to CO. That’s why a purely button-based routine leaves a gap.

The Functional Gas Test for Verifying the Sensor

To prove the sensor actually detects CO, you need a functional test with real gas. For most home detectors, the practical route is a UL-approved canned CO test agent. The Honeywell manual permits its use with the RealTest feature: spray the agent near the sensor port, and if gas entry succeeds, the detector alarms and exits its speed-up mode after 20-60 seconds.

Some detectors skip the spray entirely. The SD365CO manual describes a magnet test mode: hold a test magnet to the marked area for 6-12 seconds, then spray a UL-approved CO test agent, with the alarm sounding in up to one minute. For system-connected detectors, the NFPA 72 code guidance is explicit — they require functional testing with CO gas at acceptance and annually, plus broader system testing by a qualified technician at least every three years.

Before any gas test, Honeywell says to notify the proper authorities to avoid nuisance alarms and to check local codes to confirm whether the test is even required. That warning exists because a functional test will trip the alarm, and on a monitored system, that alarm goes somewhere.

What Not to Do

Never test a CO detector by holding it near a car tailpipe or other engine exhaust. The manufacturer’s own instructions are blunt about why: engine exhaust is hazardous to you, can damage the detector, and contains other combustion products that skew the sensor.

The weekly button test remains the baseline.

Common Mistakes That Lead to False Results

• Skipping the stabilization wait. Testing immediately after power-up yields a false failure. Honeywell’s 80-second wait is typical.
• Testing with exhaust. Unsafe for you, destructive to the detector, and not a valid method.
• Assuming the button proves the sensor. It proves the electronics, not the sensor.
• Ignoring the manual. Test timing, tone patterns, and allowed test agents vary by model.

Testing with the wrong method doesn’t just waste time — it tells you nothing. The detector’s manual is the only authoritative source for the exact sequence your model expects. When the test agent method is permitted, use exactly what the manual names; for system-connected units, schedule the annual gas test with a qualified technician before the code deadline catches you. If you’re working in confined spaces where CO risk is acute, our tested picks for confined-space CO detection gear cover monitors built for that job.