Fire Alarm Battery Sizing: NFPA 72 §10.6.7 Explained

Why fire alarm batteries exist

Every commercial fire alarm system in the U.S. is required to have a secondary power supply. The primary supply is utility power. The secondary supply is a battery bank (or, less commonly, a generator sized specifically for fire alarm). The battery has to run the entire system — panel, initiating devices, notification appliances — for a specified duration during a utility outage AND then still have enough capacity left to run the system in ALARM MODE for a specified period.

That two-phase requirement is what §10.6.7 encodes. It is not "run the system for 24 hours" — it is "run standby for 24 hours AND alarm for 5 minutes AFTER those 24 hours." The battery has to be sized for both phases, plus a mandatory derating factor, plus (for many designers) an additional aging factor.

What NFPA 72 §10.6.7 actually requires

NFPA 72-2022 §10.6.7 establishes the secondary power supply requirements for fire alarm systems. The core provisions:

• §10.6.7.1 — Standby duration. For local (protected-premises) fire alarm systems, the secondary supply must sustain the system in normal (non-alarm) operation for 24 hours. For auxiliary, remote supervising, or proprietary supervising station systems, the duration extends to 60 hours. Engineered smoke-control systems have a separate 12-hour requirement under §10.6.7.1.5.
• §10.6.7.2 — Alarm-mode duration. After the standby period, the system must operate in alarm mode for at least 5 minutes. Voice / emergency communications systems require 15 minutes per Chapter 24; mass notification and ECS systems can require 30 minutes.
• §10.6.7.2.1 — 20% derating factor. The calculated capacity must be multiplied by 1.20 to account for battery-capacity variance, temperature, and end-of-life margins.
• The battery must be sized to at least the calculated + derated value, rounded up to a commercially available size.

Different NFPA 72 editions have shuffled the sub-numbering slightly, but the substance has been stable across the 2016, 2019, and 2022 editions. Confirm the edition your AHJ has adopted before citing exact sub-numbering in a submittal.

The math, piece by piece

The full battery-sizing equation:

Every input is a value you look up on device data sheets or from the fire alarm panel's own current worksheet. The math itself is trivial. The mistakes happen in what people include vs. leave out.

Standby load — what's included

Standby current is the total DC current the panel draws when everything is normal — no alarms, no supervisory conditions, no trouble signals. It includes:

• The FACP itself — control panel processor, display, and any onboard networking
• Every SLC-connected device — smoke detectors, heat detectors, monitor modules, control modules, manual pull stations. Each has a data-sheet standby current in the microamp-to-milliamp range; sum them.
• All notification appliance circuits (NACs) — even though NACs are not active in standby, they still draw supervision current on the wiring itself. Panel data sheets list this per-NAC standby draw.
• The communicator — DACT, cellular, or IP path to the monitoring company. IP paths often draw more than DACT.
• Any auxiliary devices — LED annunciators, remote annunciators, releasing panels, aspirating detectors (VESDA), gas detectors, waterflow supervisory switches.
• Voice/ECS amplifiers in standby — the amps draw idle current even when no message is playing.

The single most-common mistake is forgetting the panel data sheet's "worst case standby" line — that line already sums the FACP + NACs + typical SLC populations for you. Then you add your specific device SLC load on top. Do not double-count.

Alarm load — what's included

Alarm current is what the panel draws when EVERY notification circuit is active. That means:

• Full NAC current — every horn, strobe, and horn-strobe on every NAC drawing its full-fire load. Look up the appliance data sheet's current for the candela setting used (higher candela = higher current).
• Releasing outputs firing — if the panel drives clean-agent release solenoids or deluge solenoids, add those in alarm.
• Voice/ECS amps in active broadcast — the amps go from idle to active, drawing significantly more current.
• Panel alarm current — the FACP itself draws more in alarm mode (relays energized, display active, communicator transmitting).
• Any addressable module in alarm state — includes both initiating modules (in latched-alarm) and control modules (activated).

The NAC amperage worksheet is the single most-important input to alarm-load sizing. Every notification device on every NAC has to be tallied for both worst-case candela and current at that candela. Underestimating alarm current is where fire alarm submittals get bounced back with corrections.

The 20% derating rule

§10.6.7.2.1 requires the calculated Ah to be multiplied by 1.20. This is not optional and it is not negotiable. The derating accounts for:

• Manufacturing variance — a "12 Ah" battery may deliver 11.5 Ah in one unit and 12.5 in another
• Temperature — battery capacity drops in cold conditions and in hot conditions relative to the 25°C nameplate rating
• End-of-life — batteries are typically replaced at ~80% of nameplate capacity, so the "usable" capacity is derated from day one

Some designers add an additional 10–15% for aging beyond the code minimum. The tool exposes this as an optional aging-factor input. Adding it is a defensive posture: your calculation still passes AHJ review if the code minimum is met, but adding aging factor gives you buffer for hot-Texas-summer performance and for the batteries' 4-5 year replacement cycle.

Worked example

A typical mid-sized commercial building. Fire alarm system with:

• Standby load: 320 mA (0.320 A) — measured off the panel worksheet
• Alarm load: 2400 mA (2.4 A) — from NAC amperage worksheet with all 8 horns/strobes active
• Standby duration: 24 hours (local protected-premises system)
• Alarm duration: 5 minutes
• Aging factor: 0% (code minimum only)

Practical result: two 12V/12Ah SLA batteries wired in series to make a 24V/12Ah pack (fire alarm panels operate at 24VDC nominally, so most panels use two 12V batteries in series).

If you add a 10% aging factor:

In this case the aging factor didn't change the required battery size — but in an edge case where the calculation lands at 11.8 Ah, adding aging pushes to 12.98 Ah and forces the next-standard size (18 Ah). That's the design margin discussion.

Selecting the next-size-up battery

Sealed lead-acid (SLA) batteries commonly used in fire alarm panels come in standardized sizes. The most common:

• 7 Ah — smallest common size, used only in tiny systems
• 12 Ah — the workhorse for mid-size commercial systems
• 18 Ah — larger addressable systems or systems with heavier NAC loads
• 26 Ah, 35 Ah, 55 Ah — larger buildings, voice/ECS systems, high-rise
• 65, 75, 100, 200 Ah — enterprise / campus, high-rise, and voice systems with 15+ minute alarm requirements

Round UP to the next standard size. Never round down — a 9.46 Ah requirement does not get a 7 Ah battery. AHJs will fail the submittal on that alone.

Common mistakes

• Skipping the 20% derating. The calculation without derating passes math but fails code. AHJs check this specifically.
• Using the panel's nameplate current instead of the actual system load. The panel spec sheet gives a "typical" standby — but your actual system may have 200 devices when the "typical" was based on 50.
• Sizing for 5 minutes when voice/ECS requires 15. Chapter 24 requires 15 minutes for voice fire evacuation systems. Missing this is a common submittal defect.
• Assuming 24-hour standby when the system is actually auxiliary or supervising station. A DACT-connected system reporting to a central station requires 60 hours per §10.6.7.1.2, not 24.
• Missing the NAC standby draw. Notification circuits draw current even when not active — panel supervision current on the wiring itself. Sum these from the NAC data.
• Not checking the physical fit. The 12 Ah calculation is right; the 12 Ah battery is too tall for the cabinet. Check panel specs before finalizing.
• Reusing a design from a similar building without re-running the math. A 30,000-square-foot warehouse and a 30,000-square-foot medical office building have very different NAC loads.

Frequently asked