Your Smoke Alarm Has an Expiration Date — and It's Printed on the Back

The label on the back says it plainly

The alarm in the photo is a Firex model i4618A — a hardwired ionization alarm with a 9-volt battery backup, the kind installed in thousands of houses built or remodeled in the 2010s. Its back label reads "2015 Jun.26" and, right next to it, "REPLACE IN 10 YEARS." Ten years from that date was June 2025. We found it in service more than a year later.

That's not unusual, and it isn't a sign of neglect. The battery had been changed. The test button worked. Nothing about the unit looked old from the floor. The replacement date is printed on the one surface nobody sees.

Where the 10-year rule comes from

The ten-year limit isn't a manufacturer's marketing number. NFPA 72, the National Fire Alarm and Signaling Code, is the standard that governs smoke alarms, and NFPA's guidance states it without hedging: "All smoke alarms must be replaced 10 years after its manufacturing date." The U.S. Consumer Product Safety Commission says the same thing in its homeowner guidance: "Replace smoke alarms that are more than 10 years old. Smoke alarms don't last forever."

Two details matter here, and both get missed.

First, the clock starts at manufacture, not installation. An alarm that sat in a warehouse and on a shelf for a year before it went up on your ceiling has a year less of service life than you'd assume. The only date that counts is the one on the back.

Second, this applies to every smoke alarm — hardwired units with battery backup, plain battery units, and the newer sealed "10-year battery" models. The sealed ones don't escape the rule; they're built so the battery and the alarm expire together, and CPSC's guidance is that "after 10 years, the entire unit is disposable."

For a Tennessee homeowner, the edition of NFPA 72 you'll most often see referenced is the 2013 edition, because that's the one the 2017 National Electrical Code points to — more on the NEC below.

Why ten years — what the research actually documents

Here's the part that's usually explained badly, including in an earlier version of our own social post on this subject. The popular version is "the sensor wears out and gets less sensitive." The documented picture is more specific than that.

The most direct evidence comes from a CPSC engineering study that collected 155 smoke detectors from real homes — units that had failed a smoke test, failed a test button, or been disconnected by the homeowner — and bench-tested them. The lab found four recurring problems, and two of them are why age matters:

Contamination in the sensing chamber. In the lab's words, "Dust, cooking contaminants, high humidity, debris, and insects in the sensing chamber can increase the sensitivity level of the detector, which leads to excessive nuisance alarms." Almost one-third of the units collected for nuisance alarming had "significant accumulations of debris," and small insects were found inside ionization chambers, altering the sensor's reading. This is the real mechanism: what gets into the chamber over years of service changes how the sensor reads the air. Both sensor types are affected — the CPSC/USFA technology roadmap notes that ionization chambers respond to "water droplets, dust, and other aerosols" and that photoelectric chambers can be triggered by "backscattered light from dust or other aerosol particles that enter the detector chamber."

Deterioration of components. The same lab traced a set of smoke-test failures to corrosion on the horn contacts, caused by ordinary household contaminants — cooking fumes and cleaning products are named specifically. Those alarms had working power and a working sensor circuit; the part that makes the noise had quietly failed. Others had burned-out light sources, failed resistors, or failed capacitors. One photoelectric unit, after its light source was replaced, could still respond to heavy smoke but no longer met the sensitivity limit in the UL 217 test standard — a working-looking alarm that was out of spec.

Notice what this means in practice. Contamination usually pushes an alarm toward false alarms first, and the NFPA's national fire data shows what happens next: the leading reasons alarms fail in real fires are a "missing or disconnected battery" (35 percent) and a "dead or discharged battery" (24 percent), with the report noting that "power sources are often disabled because of unwanted alarms." "Lack of cleaning" is itself listed as the cause in 8 percent of failures across all alarm types — and 14 percent for hardwired-only alarms — with "defective unit" at 8 percent overall and 19 percent for hardwired-only units. An alarm that has spent a decade collecting what's in your air is more likely to be either crying wolf or silent, and either way it ends up not protecting you.

One thing the ten-year rule is not about: the radioactive source. The back of an ionization alarm discloses a tiny amount of Americium 241 — the label in our photo lists 0.9 microcuries, which the CPSC/USFA roadmap confirms is what "virtually all" residential ionization alarms use. That source is not what wears out over a decade. The chamber around it, the electronics, and the horn are.

The test button isn't a smoke test

This is the point most worth remembering. Pressing the test button confirms the alarm has power and the horn works. It does not put smoke in the chamber.

The CPSC's current national in-home study — 1,314 alarms tested in 985 households, completed in 2024 — tested alarms with an aerosol smoke spray first, and only pressed the test button if the alarm didn't respond to smoke. An alarm whose button sounded but which didn't respond to smoke was classified as having an inoperable sensor and tagged for replacement. That ordering is deliberate: the button can pass on an alarm that can't do its job.

What that testing found, per the report:

• 79 percent of alarms worked on the first try.
• 13 percent didn't work until a battery was replaced or installed.
• 9 percent — about 1 in 11 — were "completely non-functioning, even when power was restored." Their manufacture dates ran from 1990 to 2022, and the report notes some "were relatively new and stopped working before their expected replacement date."
• Among households that believed all their alarms were working, 16 percent had at least one that failed the initial test.
• Only 51 percent of households correctly knew the age of all their alarms; 24 percent were never right, and 13 percent couldn't guess. The report's conclusion: "the appearance of the alarms is not enough information for householders to assess or even guess their age."
• Only 13 percent of households knew that a smoke alarm should be replaced every 10 years.
• 33 percent said they never use the test button at all, and only about 7 percent test monthly as recommended.

None of those households were careless in any obvious way. They had alarms. Most had changed batteries. The failures were invisible without either a smoke test or reading the date.

What a working alarm is actually worth

NFPA's June 2024 analysis of U.S. home fires reported to fire departments (2018–2022) is the source for the numbers that matter:

• Nearly three out of five home fire deaths — 59 percent — were caused by fires in homes with no smoke alarm (43 percent) or an alarm that failed to operate (16 percent).
• The death rate per 1,000 reported home fires was approximately 60 percent lower in homes with working smoke alarms (5.8) than in homes with no alarm or none that operated (14.4).
• Each year, an estimated 19,119 home fires occurred where an alarm should have operated but didn't, causing an average of 403 deaths and 1,052 injuries annually.
• When present, hardwired alarms operated in 94 percent of fires large enough to trigger them; battery-only alarms operated 85 percent of the time. Two-thirds of the fatal injuries in homes that had alarms happened in homes with battery-only units.

That last point is worth sitting with if your home still relies on standalone battery alarms: the power source that depends on someone remembering is the one that fails more.

How to check yours — about a minute per alarm

Twist the alarm off its bracket. Hold the body of the alarm and rotate it — usually counterclockwise — until it releases from the mounting plate. On a hardwired unit it will stay tethered by a wiring plug; that's normal, and you don't need to unplug it to read the label. The manufacture date is on the back plate, not inside the battery door. Opening the small battery compartment shows you the battery and nothing else.

Find the date. Look for a printed or stamped date — "2015 Jun.26" in our example — often near a "replace by" or "replace in 10 years" line. Add ten years. If that date has passed, or there's no date at all, the alarm is due.

Check the type while you're there. The label will say ionization or photoelectric (or both), and whether it's a "single and/or multiple station" unit — meaning it can be interconnected with others. If it says 120 VAC, it's hardwired.

Test with the button monthly — NFPA's guidance is at least once a month — and take a chirp seriously. A single chirp every 30 to 60 seconds is a low battery. Chirping that continues after a fresh battery means, in NFPA's words, "the alarm is at the end of its life and the unit must be replaced."

Clean it per the label — this unit's says annually, with compressed air or a vacuum around the perimeter — and read the sentence that follows: "If cleaning does not restore your alarm to normal operation, the alarm should be replaced."

Ionization, photoelectric, or both

When you do replace an alarm, you get to choose the sensor type, and the research is clear about the trade-off. In NIST's full-scale fire tests, ionization alarms responded before photoelectric alarms in every flaming-fire trial, and photoelectric alarms responded first in smoldering fires — the slow, smoky kind that can burn for an hour or more before flames appear. Dual-sensor alarms, which carry both, had the fastest average response across the board in one test series — activating, on average, 539 seconds ahead of ionization alarms and 79 seconds ahead of photoelectric alarms.

NFPA's recommendation follows from that: use both types in the home, or combination ionization-photoelectric alarms, and put photoelectric units near kitchens and bathrooms where nuisance alarms are the problem. Keep any alarm at least 10 feet from a cooking appliance.

One more thing NIST found that's useful when you're comparing boxes on a shelf: the sensitivity range printed on the back of an alarm — "0.61 – 1.13 %/ft" on ours — is the allowable production window for a listed alarm, not a performance score, and it "is not predictive of relative alarm performance when comparing any two alarms." Pick by sensor type and listing, not by that number.

Hardwired alarms, interconnection, and the code year

If your alarms are hardwired — and in most homes built since the early 1990s they are — replacing them is an electrical job, and NFPA's guidance is direct: "Only qualified electricians should install hardwired smoke alarms." CPSC notes that new homes have included hardwired, interconnected alarms since 1989, and NFPA's data shows hardwired units are both more likely to operate in a fire and more likely to be interconnected. For older homes that were never wired for it, wireless-interconnected battery alarms now exist and can get you the same "one sounds, all sound" behavior without opening walls — when interconnecting, use alarms from the same manufacturer.

On the code side, a few things worth knowing, stated against the 2017 National Electrical Code, the edition Tennessee and the jurisdictions Red Cedar Electric works in are currently enforcing:

• The NEC does not tell you how many smoke alarms you need or where they go. The phrase "smoke alarm" appears in the 2017 NEC exactly once — in an informational note under Article 210.12 that points to NFPA 72-2013 for the power-supply requirements of smoke alarms in dwellings. Count and placement come from NFPA 72 and the residential building code your jurisdiction enforces, not from the electrical code.
• The circuits that feed alarms are AFCI territory. NEC 2017 210.12(A) requires arc-fault circuit-interrupter protection for 15- and 20-amp, 120-volt branch circuits serving bedrooms, hallways, and most living areas — exactly where smoke alarms live.
• A straight swap is different from adding alarms. Replacing an existing hardwired alarm with a compatible new one on the existing wiring does not modify the branch circuit. But under 210.12(D), if the wiring in those areas is "modified, replaced, or extended" — say, running cable to add an alarm where there wasn't one — the circuit has to be brought up to AFCI protection, with one exception for an extension of six feet or less that adds no new outlets or devices. A new alarm location is a new device, so that exception won't cover it. This is why "just add one more alarm in the hall" can turn into a breaker change, and it's worth knowing before you get a quote.

If you're reading this from a jurisdiction on a newer code cycle, confirm which edition your inspector is enforcing before relying on any section number, including ours.

When to call

Call a licensed electrician when your hardwired alarms are past their date, when alarms have been disconnected because of nuisance tripping, when you're adding bedrooms or finishing a basement and need alarms where none exist, or when you want standalone battery alarms brought into an interconnected system. An electrician can replace the units with compatible listed alarms, confirm interconnection actually works across all of them, check the circuit's AFCI status under 210.12, and leave you with a written record of manufacture dates so the next ten-year check isn't a mystery.

Stop and call right away if an alarm has visible scorching or melting, or if a hardwired alarm has gone completely dead — no light, no chirp — which can mean a circuit or wiring problem rather than an alarm problem.