The 90-Minute Rule: Why Emergency Lighting Fails Fire Marshal Audits

The 90-Minute Rule: Why Emergency Lighting Fails Fire Marshal Audits (and How DFW Facilities Can Avoid It)

A maintenance tech walks the building with a screwdriver, presses the little test button on each emergency light, watches the lamps come on for a second or two, and moves on. Every unit "passed."

Six weeks later the fire marshal shows up, asks for the annual load test records, and the inspection goes sideways.

This is the single most common emergency lighting finding we see in commercial buildings across the Fort Worth and Dallas metro, and it comes down to one number: 90 minutes.

What the Code Actually Requires

NFPA 101, the Life Safety Code, doesn't ask whether your emergency lights turn on. It asks whether they will still be lighting the exit path an hour and a half after the power drops. There are two separate tests, and doing one does not cover the other.

The monthly test — 30 seconds. At intervals not exceeding 30 days, every emergency lighting unit and battery-backed exit sign gets a functional test for a minimum of 30 seconds. This is the button-push test. It confirms the transfer relay works, the lamps are intact, and the charger isn't dead. That's all it confirms.

The annual test — 90 minutes. Once a year, every unit gets a full 1.5-hour discharge test under load. The unit runs on battery alone for the entire 90 minutes, and the lamps have to still be producing usable illumination at the end of it. This is the test that tells you whether the battery has any real capacity left.

If your building has self-testing/self-diagnostic units, the monthly manual test can be replaced by a visual inspection of the status indicator, and many of those units run their own annual 90-minute cycle. You still need the records. A green LED nobody has ever written down is not documentation.

One practical note: after a 90-minute discharge, the batteries need roughly 24 hours to recharge. Don't schedule that test on a day when the building has to be fully protected an hour later, and don't let a contractor run it on his way out the door at 4:45 p.m.

Why a Unit Passes at 30 Seconds and Fails at 90 Minutes

A sealed lead-acid battery doesn't quit all at once. It loses capacity gradually, and a battery down to 15% of its rated capacity will still fire the lamps enthusiastically for the first thirty seconds. It just can't hold them for an hour and a half.

Here's what actually causes failures in the field:

Battery age. Sealed lead-acid batteries in emergency lighting units typically last four to five years, and heat shortens that considerably. A unit mounted near the ceiling of an un-conditioned warehouse in a North Texas August is living in an oven. If you don't know how old your batteries are, they're probably older than you'd guess.

Remote heads. A single unit powering two or three remote lamp heads is drawing far more current than the same unit powering its own two lamps. Remote heads get disconnected during remodels, aimed at a wall, or added to a unit that was never sized for the extra load. They are the most frequently skipped item on a monthly walkthrough because they're across the room from the unit that feeds them.

Aging exit signs with battery backup. Everyone remembers to look at whether the exit sign is lit. Almost nobody kills power to it and watches it for 90 minutes. LED exit signs last a very long time; the little battery pack inside them does not.

Chargers and pilot lights. A charge indicator confirms the charger is energized. It does not confirm the battery accepts or holds a charge. Plenty of failed units have a cheerful green light on the front.

What the Fire Marshal Is Actually Reading

When an AHJ or an insurance carrier asks for your emergency lighting records, they are not looking for a sticker on the wall. NFPA 101 requires written records of your inspections and tests to be kept and made available. What separates a clean file from a problem file:

• Unit-by-unit detail, not a building-level "pass." A report that says "emergency lighting tested — satisfactory" tells an inspector nothing. Records should identify each unit by location or ID number.
• Test type and duration recorded separately. The monthly 30-second checks and the annual 90-minute discharge test should be distinguishable on the record. If everything looks the same, the inspector assumes the 90-minute test never happened.
• Explicit pass/fail per unit. Including the failures. A record showing three failed units and a repair date is far more credible than a record showing 100% pass rates year after year.
• Deficiency notes with enough specificity to act on. "Unit 14, west dock — battery failed at 22 minutes" is a finding. "Needs service" is not.
• Corrective action and a date. The question after a failure is always: what did you do about it, and when? A file that answers that question closes an audit fast.
• Who performed the work. Name, license, company.

Insurance carriers apply the same logic, often harder. Loss control reports routinely flag emergency lighting records that show no annual load testing, and that flag lands on your renewal.

The Straightforward Fix

Emergency lighting is one of the least expensive life safety systems in a commercial building and one of the easiest to let drift, precisely because the button-push test feels like a real test. The correction is procedural, not technical:

1. Inventory every unit, remote head, and battery-backed exit sign in the building, including the ones added during tenant improvements.
2. Do the monthly 30-second checks in-house and write them down. A simple log with dates, unit IDs, and initials is sufficient.
3. Schedule the annual 90-minute load test as a real appointment, with a report that documents each unit individually.
4. Replace failed batteries and units promptly, and keep the invoice with the test record.

Do It in One Visit

If you're already scheduling an annual NFPA 10 fire extinguisher inspection or NFPA 72 fire alarm ITM, the emergency lighting load test belongs on the same visit. Same building walk, same access coordination, one set of records, one point of contact when the fire marshal asks questions.

Fortress Fire Protection provides inspection, testing, and maintenance for portable fire extinguishers, fire alarm systems, and emergency lighting across the Dallas–Fort Worth metroplex, including Fort Worth, Dallas, Arlington, Irving, Grand Prairie, Grapevine, Denton, Flower Mound, Weatherford, Decatur, Azle, and Boyd. Our reports are built to be read by someone who is looking for problems: unit-by-unit findings, clear pass/fail results, and deficiencies documented and rated by severity.

Call 817-527-9900 or email info@fortressfireprotection.org to schedule combined life safety testing for your facility.