Fire alarm system installation and acceptance testing field guide

What a fire alarm system is, and the four parts

A fire alarm system is the wired network that detects a fire, decides what to do about it, and warns the people and the equipment that need warning, all of it watched for faults around the clock. Strip away the model names and every system is four parts. The fire alarm control panel, the FACP, is the brain. The initiating devices are the inputs, the smoke detectors and heat detectors and pull stations that tell the panel something is wrong. The notification appliances are the outputs, the horns and strobes that tell the building. And the monitoring path reports the whole thing offsite to a central station so help comes when nobody is there to call it in.

What separates a fire alarm system from a string of smoke alarms from the hardware store is supervision. Every circuit is watched. Cut a wire, lose a device, drop the battery, and the panel does not go quiet. It goes into trouble and tells you. That single idea, that a silent failure is not allowed, drives almost every code rule in NFPA 72 and almost every callback on a system that was wired by someone who did not understand it.

The framework is NFPA 72, the National Fire Alarm and Signaling Code. It governs the devices, the circuits, the power, and the testing. The wiring methods come from NFPA 70, the NEC, mostly Article 760. Where a system is required in the first place comes from the building and fire codes, the IBC and IFC, by occupancy. The AHJ, the authority having jurisdiction, adopts an edition, amends it, and has the final say. When this guide gives a number, the adopted edition and the AHJ control it.

What is the difference between conventional and addressable?

A conventional system reports by zone. An addressable system reports by device. That is the whole difference, and it changes how you wire, how you troubleshoot, and what the panel can tell the fire department when it matters.

On a conventional system, devices wire onto an initiating device circuit, an IDC, and the panel sees one zone. When a detector goes into alarm, the panel knows the zone tripped, not which of the dozen detectors on it did. You walk the zone to find the device. A zone usually maps to an area, a floor, or a tenant space, and the granularity is whatever you wired into it. Conventional is cheaper for small jobs and dead simple, and it still has a place on a small building with a handful of devices.

On an addressable system, each device carries its own address and sits on a signaling line circuit, the SLC, a data loop that the panel polls continuously. The panel knows that detector 14 in the east corridor is in alarm, not just that something on the east side tripped. That point-level reporting is the reason addressable won the commercial market. On a large building you cannot walk a zone of forty devices to find the one in alarm while people are evacuating. The SLC also lets the panel read each detector's sensitivity and dirty status, which is how addressable systems handle the annual sensitivity check without pulling every head. For anything beyond a small building, addressable is the default. Conventional survives where the device count is low and the budget is lower.

Smoke detection: photoelectric versus ionization

Photoelectric detectors see slow smoldering fires fastest. Ionization detectors see fast flaming fires fastest. The difference is particle size, and it decides where each one belongs.

A photoelectric chamber works on light scatter. Big smoke particles, the kind a smoldering couch, a mattress, or an overheating wire insulation throws off before it ever flames, bounce light onto a sensor and trip the alarm. An ionization chamber works on a small ionized current that tiny combustion particles disrupt, and a fast flaming fire in paper or a flash kitchen fire makes exactly those small particles. Each technology is faster on the fire it is built for and slower on the other, which is why NFPA guidance leans toward photoelectric, ionization, or a combination dual-sensor head depending on the hazard.

In the field the call is usually made for you. Near cooking, the code effectively forces photoelectric, commonly within about 20 ft of a fixed cooking appliance, because ionization heads nuisance-trip on cooking aerosols and a detector that cries wolf is a detector people learn to ignore. The same logic kills ionization in dusty, humid, or steamy spaces. Smoke detectors carry a nominal spacing of 30 ft on a smooth ceiling, but spacing is a starting point, not a guarantee, and the listing, beam ceilings, airflow, and the room's geometry all pull it in. Air handling is the quiet killer: high airflow sweeps smoke past a detector before it builds, so detector placement near supply diffusers gets adjusted or the detection misses.

Heat detectors: fixed temperature, rate of rise, and spacing

Heat detectors are slower than smoke and that is the point. You put them where smoke detection would nuisance-trip, in a boiler room, a kitchen, a dusty shop, a parking garage, an attic, anywhere smoke is normal and heat is not. They protect property more than they protect escaping people, because by the time a heat detector trips there is real fire.

Two mechanisms, often in one device. A fixed-temperature element trips when it reaches its set point, commonly a 135 degree F or 200 degree F rating. A rate-of-rise element trips when the temperature climbs faster than a set rate, around 15 degrees F per minute, which catches a fast-developing fire before the fixed element ever reaches its number. A combination detector carries both, so the rate-of-rise element answers a fast fire and the fixed element backs it up on a slow one. Rate-of-rise can false on a legitimate fast warmup, a sun-baked attic or a space heater kicking on, so the application has to match the environment.

Spacing is where heat detectors differ hard from smoke. Each heat detector has a listed spacing, not a nominal one, and it can be well under the 30 ft people carry in their head for smoke, often 50 ft or less on a smooth ceiling and frequently tighter. NFPA 72 then applies what the trade calls the point-seven rule: every point on the ceiling has to fall within 0.7 times the listed spacing of a detector, so coverage is real and not just a grid of dots. Ceilings over about 10 ft force the spacing down further, because heat stratifies and spreads on its way up and a detector spaced for a 10 ft ceiling misses on a 20 ft one. Beams, joists, and sloped ceilings all reduce it again. When in doubt, the listed spacing and the NFPA 72 reduction rules govern, not the round number.

Duct detectors, pull stations, waterflow, and tamper

Smoke and heat are not the only inputs. A finished system ties in several others, and getting their signal type right is where installers slip.

Duct smoke detectors sample the air moving through HVAC ductwork and shut the unit down so a fire does not get pushed through the building by its own air handler. The mechanical code commonly triggers them on supply or return systems moving more than 2000 CFM, and they tie to the fire alarm under NFPA 90A. Here is the catch people miss: a duct detector connected to the FACP initiates a supervisory signal, not a general alarm, because it is there to control the fan, not to evacuate the building. It is also not a substitute for area smoke detection. Manual pull stations give a person a way to start the alarm by hand. They go by the exits, commonly within about 5 ft of each exit doorway, with the operable part mounted in the reach range, often 42 in to 48 in above the floor, and placed so travel distance to a station does not exceed roughly 200 ft. Confirm the heights and distances against the adopted edition and the ADA.

From the sprinkler system come two more inputs, and they are not the same signal. A waterflow switch senses water actually moving in a sprinkler line, which means a head has opened and there is a fire, so it initiates an alarm and it is one of the few devices tested quarterly rather than annually. A valve tamper switch senses a control valve being closed, which does not mean fire but does mean the sprinkler system has been disabled, so it initiates a supervisory signal. Wire a tamper as an alarm and you train the building to evacuate every time maintenance closes a valve. Wire a waterflow as supervisory and you have hidden a real fire behind a trouble light.

Why do strobes have candela ratings?

A strobe carries a candela rating because the room sets how bright the flash has to be to reach everyone in it, and candela is the unit of that effective intensity. Put a 15 cd strobe in a space that needs 110 cd and a person at the far wall never sees the warning. NFPA 72 sets the candela by the room size and the mounting, in tables, and a small room may be satisfied by a wall-mounted 15 cd unit while a large open area needs much higher, stepping up through ratings like 30, 60, 75, 95, 110, 135, and 185 cd.

Two rules trip crews on the strobe side. Mounting height matters because the flash has to clear the furniture and the heads in the room, so wall strobes commonly mount 80 in to 96 in above the floor, or just below the ceiling, per the listing and the code. Synchronization matters because strobes flashing out of step can trigger a seizure in a susceptible person, so when more than two strobes fall in one field of view they have to flash in sync, which means a sync module or a panel that synchronizes the NAC.

Audibility is the other half. The horn has to be heard over whatever the space normally sounds like, and NFPA 72 sets the public-mode threshold at 15 dBA above the average ambient sound level, or 5 dBA above the maximum sound level that lasts at least 60 seconds, whichever is greater, measured about 5 ft above the floor on the A-weighted scale. Sleeping areas are stricter, commonly 75 dBA at the pillow. The ADA history is worth knowing because old jobs still carry its ghost: the original 1990 ADA pushed a 15/75 cd strobe before NFPA had researched visible signaling, the 2009 amendment pointed back to NFPA 72, and current visible-notification design follows NFPA 72. The AHJ and the listing govern the final numbers.

The circuits: IDC, SLC, NAC, and Class A versus B

Three circuit types carry the whole system, and you cannot wire a panel without knowing which is which. The IDC, the initiating device circuit, carries conventional inputs back to the panel as a zone. The SLC, the signaling line circuit, is the addressable data loop the panel polls for each device by address. The NAC, the notification appliance circuit, carries power out to the horns and strobes. Inputs come in on IDC or SLC. Outputs go out on NAC. Get the vocabulary straight and the prints read themselves.

Class is about what happens when a wire breaks. A Class B circuit is a single path: the wire leaves the panel, passes through each device, and ends at an end-of-line device that the panel watches. Break the wire and everything past the break goes dark, silently, with only a trouble at the panel to show for it. A Class A circuit is a loop fed from both ends, so a single break annunciates as a trouble but every device stays online because the panel now feeds both sides of the fault. Class A buys you survivability past one break, at the cost of a return path, either a second cable or a four-conductor run, and the extra labor.

Survivability is a separate idea from class, and NFPA 72 treats them in separate places, the pathway class in one part of Chapter 12 and pathway survivability in another. Class is electrical topology, how the circuit behaves with a break. Survivability is physical, how long the wiring keeps working while it is on fire, with levels that step up to two-hour-rated circuit integrity for the cases that demand it. A high-rise or a system where the alarm in one zone depends on wiring passing through another is where survivability rules bite. Where the code or the AHJ calls for Class A, Class X, or rated survivability, that requirement governs over the cheaper Class B default.

Fire alarm cable, separation, and supervision

Fire alarm wiring is power-limited cable run under NEC Article 760, and the markings tell you where it can go. FPLR is the riser rating for vertical runs between floors. FPLP is the plenum rating for air-handling spaces. FPL is the general type, and the more demanding location always wins, so a plenum gets FPLP whether or not the cheaper cable would physically fit. Use the wrong rating in a plenum and you fail inspection on the markings alone, before anyone tests a device.

Keep fire alarm conductors out of the same raceway and enclosures as power and Class 1 conductors, with the separations Article 760 spells out, because power-limited fire alarm circuits are not allowed to share the pipe with line-voltage power except under specific conditions. The practical reason behind the rule is noise and fault energy: you do not want a power fault dumping into the circuit that is supposed to save the building. Support the cable, do not let it lie on the ceiling grid, and keep it out of the way of other trades who will treat anything red as optional.

Every circuit is supervised, and supervision is the feature, not a nuisance. The panel pushes a small monitoring current through each circuit and watches it. An open conductor reads as a break in that current and shows as a trouble. A conductor touching ground reads as a ground fault and shows as a trouble. That is why conventional circuits end in an end-of-line resistor: it completes the supervised path so the panel can tell a healthy circuit from a cut one. Leave the resistor out, or splice the circuit so the supervision loops back early, and you have a circuit that reads fine while half of it is dead. Troubleshooting a fire alarm is mostly hunting opens and grounds, and the panel is already telling you which circuit to walk.

Primary power and the secondary battery

A fire alarm system runs on two sources, and the code requires both. Primary power is the building supply, a dedicated branch circuit, commonly with a marked, locked, or otherwise identified breaker so nobody kills the fire panel thinking it is a receptacle circuit. Secondary power is the backup that carries the system when the building loses power, almost always a sealed battery set inside or beside the panel, sized by calculation, not by guess.

The reason the battery is a calculation and not a catalog number is that the load is two different loads. In standby, the system sits quiet but alive, drawing a small current to power the panel and poll the devices. In alarm, every horn and strobe fires and the current jumps several times over. The battery has to ride out a long power outage in standby and still have enough left to run a full alarm at the end of it. That is the whole logic behind the standby-plus-alarm requirement.

Mark the battery with its install date. Sealed lead-acid batteries lose capacity with age and heat, and a four-year-old battery that calculated fine when new will not hold the alarm load. The ITM program replaces them on a cycle for exactly this reason, and the acceptance test checks that the installed battery actually carries the calculated load, not just that it reads 24 volts at rest.

How do you size the fire alarm battery?

Size the battery for 24 hours of standby followed by 5 minutes of full alarm, then add a margin. NFPA 72 requires the secondary supply to carry the system in standby for at least 24 hours and then operate the alarm for at least 5 minutes, with voice evacuation systems pushed to 15 minutes of alarm. If the secondary power is an engine-driven generator instead of batteries, the standby requirement commonly drops to 4 hours. Confirm the exact figures against the adopted edition.

The math is bookkeeping done carefully. Total the standby current of everything on the panel and multiply by 24 hours. Total the alarm current, using the installed candela setting of each strobe and the real horn draw, not the worst-case number off the spec sheet, and multiply by 0.083 hours, which is 5 minutes. Add the two ampere-hour figures, then multiply the sum by 1.25 to cover battery aging and temperature derating. That product is the minimum battery capacity, and you select the next standard size at or above it.

The error that bites is the alarm current. Crews grab the maximum rated current from the data sheet instead of the current at the candela actually programmed, and the calculation comes out too big, or they use the wrong candela and it comes out too small. The calculation is part of the submittal and the permit record, and the AHJ checks it, so do it on the installed settings and keep it with the as-builts. A battery calc that does not match the devices in the field is a calc that fails its own acceptance test.

I standby

Total current the system draws at rest, powering the panel and polling devices, in amps

I alarm

Total current with all appliances firing, at the installed candela settings, in amps

0.083 h

Five minutes expressed in hours, the standard non-voice alarm duration; use 0.25 h for 15 min voice

1.25

Aging and derating margin applied to the summed ampere-hours

Why do strobes dim or fail at the end of the circuit?

They run out of voltage. A NAC is a power circuit, and like any power circuit it loses voltage along the wire as current flows, so the appliance at the far end sees less than the panel puts out. Push too many horns and strobes down too long a run on too small a wire and the last device drops below the voltage it needs to fire. This is the same voltage-drop physics as any feeder, covered in the voltage-drop field guide, applied to a 24-volt circuit that has almost no margin to give.

The hard number is the floor, not the nominal. A notification appliance typically needs at least 16 VDC to operate reliably, and the panel may only put out around 20.4 V at the worst case of a discharged battery. Subtract the drop along the wire and the appliance draws from whatever is left. Below about 16 V the strobe may not flash, the horn sounds weak, or the device draws erratically and drags the circuit down further. NAC voltage drop is the single most common reason a new system fails its final inspection, so it gets calculated before the pull, not discovered at the test.

When the calculation does not close, you have a short list of fixes. Go up a wire size to cut the resistance. Split one long NAC into two shorter circuits. Move the load center or, more often, add a booster power supply, a NAC power expander, out near the appliances so the long run carries data and a short run carries the power. And switch xenon strobes for LED appliances, which draw far less current and ease both the voltage drop and the battery calc. Do the point-to-point calculation for the honest answer; the end-of-line shortcut is faster but less accurate, and on a tight circuit the difference is the inspection.

Programming the cause and effect

Installing the devices is half the job. The other half is telling the panel what to do when each input trips, and that logic is the cause-and-effect, written as an input-to-output matrix. Every initiating device or zone is a row. Every output, the notification, the relays, the offsite signal, is a column. The cells say what fires when. The matrix is the design intent made testable, and it is the document the acceptance test checks against, cell by cell.

The outputs go well past horns and strobes. The panel drives relays that shut down HVAC fans so smoke does not spread, recall the elevators to a safe floor under firefighters' service, release magnetic door holders so fire and smoke doors close, release egress door locks so people get out, and on systems with suppression, trigger the release. Each of those is a life-safety action with its own timing and its own interlocks, and each one belongs in the matrix with the input that causes it. A relay that closes a damper is only as good as the cause-and-effect that fires it at the right time.

The suppression interface is where this gets unforgiving, and a clean-agent or pre-action data center room is the sharp example, covered in the data center fire and life safety guide. A release decision usually depends on cross-zoned detection, two independent detectors confirming before the agent dumps, exactly so a single false trip does not discharge a cylinder bank onto live equipment. That cross-zone logic lives in the cause-and-effect, and it is the part of the matrix the commissioning agent tests hardest, because the cost of getting it wrong is a six-figure nuisance discharge or, worse, a fire the agent never reached.

Monitoring and the communicator

Most fire alarm systems are required to report offsite to a supervising station, because a building that alarms with nobody in it still needs the fire department dispatched. The communicator is the device at the panel that makes that call, and the path it uses has moved with the phone network.

The old standard was a digital alarm communicator transmitter, a DACT, dialing out over two phone lines. Copper phone service is disappearing, and the trade has moved to cellular and IP communicators, often with two independent paths for redundancy, so the loss of one does not leave the building unmonitored. The communicator and its paths are supervised like everything else: lose a path and the panel shows a trouble, because a monitoring connection that fails silently is a connection that is not there when the fire is.

Confirm what the AHJ and the listing require for the number of paths, the supervision interval, and the acceptable technology, because this is an area the code has revised hard across recent editions. The communicator type that was fine on the last job may not satisfy the current adopted edition.

Grounding the panel

The FACP needs a proper equipment ground and a clean reference, and a fire alarm panel grounded poorly throws ground-fault troubles that chase you for days. Bond the panel to the building grounding system per the manufacturer's instructions and the NEC, and treat the panel ground as part of the whole grounding electrode system rather than a wire landed on the nearest pipe. The grounding and bonding field guide covers how that system is built and why one bonded system, not several separate grounds, is what you want.

On the fire alarm side specifically, a ground fault is not just a code issue, it is a supervised fault the panel reports, so a sloppy ground reads as a recurring trouble. Clean grounding at the panel and clean separation of the circuits is what keeps the system quiet when there is nothing wrong and loud when there is.

What is a fire alarm acceptance test?

A fire alarm acceptance test is the 100 percent test of a new system, every device activated, every output verified, every function checked against the cause-and-effect, and the whole thing documented on the record of completion. New installations get the full test regardless of occupancy. There is no sampling on a new system. If it was installed, it gets tested.

The test runs the matrix. You trip each initiating device and confirm the right notification fires, the right relays operate, the HVAC shuts down, the doors release, the elevators recall, and the offsite signal reaches the supervising station, all on the timing the cause-and-effect calls for. You confirm every notification appliance actually operates and, on the NAC side, that the end-of-line voltage holds with the appliances loaded. The wiring gets a continuity and insulation check, and crews still megger the conductors to catch the ground that will become a trouble next month. Nothing is assumed. The panel may say a device is online; the test confirms it does its job.

When it passes, the installer fills out the NFPA 72 record of completion, the standardized form the code provides, documenting what was installed and how it tested. The AHJ does not have to witness the test under the code itself unless they ask to, but the building code often requires the owner to offer the AHJ the chance, and on most commercial jobs the AHJ is standing there. Either way the record of completion is the document that says the system was installed and tested to the approved plans, and it follows the system for its life.

Reacceptance, sensitivity testing, and the ITM the owner inherits

The acceptance test is day one. After that the system enters inspection, testing, and maintenance, the ITM the owner inherits the day the building opens, and the schedule comes from NFPA 72 Chapter 14. That chapter applies retroactively to every installed system, no matter how old, so an owner who skips it is out of compliance even on a system that passed acceptance years ago.

The frequencies are not uniform, and that is the part owners underestimate. Smoke detectors, heat detectors, pull stations, notification appliances, and control equipment are functionally tested at least annually. Waterflow switches are tested quarterly, the shortest interval of the common devices, because a stuck waterflow switch hides a real fire. Visual inspections run on their own schedule, from monthly to annually depending on the device. Smoke detector sensitivity is its own requirement and the one most often missed: detectors are sensitivity-tested at one year and then commonly every two years after, and a head that drifts outside its listed sensitivity range gets cleaned or replaced. A detector can pass a functional puff test while reading far outside its range, which is exactly how a dirty detector either nuisance-trips or sits too dull to catch a fire.

Reacceptance is the rule people forget on remodels and tenant fit-outs. Any time the system is modified, added to, or reprogrammed, the affected work gets reacceptance tested, and the cause-and-effect gets re-verified for anything the change touched, because a change to one zone's programming can break another's logic. The code commonly allows testing a sample, often 10 percent of the devices up to a cap of around 50 components for a modification, as the AHJ determines, rather than the full 100 percent of a new install. Change the programming and you retest the function, not just the device, because the matrix is what you actually changed.

High-rise survivability and the data center interface

Two scenarios push the system past the baseline, and both are worth knowing before you bid one. High-rise buildings demand survivability, because the alarm and voice signal for an upper floor may depend on wiring that runs through the floors below, and that wiring has to keep working while the lower floor burns. That is where two-hour circuit integrity cable, Class A or Class X pathways, and survivability levels come in, and where a fire command center centralizes the controls for the fire department. The survivability rules are not optional dress-up; they are the difference between a working alarm and a dead one above the fire.

The data center is the other end of the spectrum, and the fire alarm there is the trigger for everything else, covered in depth in the data center fire and life safety guide. The panel reads early detection, often aspirating or cross-zoned spot detection, and on a confirmed event drives the clean-agent or pre-action release through the cause-and-effect, with the cross-zone logic that prevents a single detector from dumping agent onto live racks. The fire alarm does not put the fire out in those rooms; it decides when the suppression does, and it has to coordinate the HVAC shutdown, the damper closure, and the abort logic in the right order. The interface, not the detection alone, is what gets commissioned hardest.

Field checklist

What to document

The record of completion, the as-built drawings, the battery and NAC calculations, and the cause-and-effect matrix are the package the system carries for life. The ITM technician five years out works from these, and the AHJ checks every later inspection against them. A device tested and not recorded is a device nobody can prove works.

For each device, log what it is, where it is, how it tested, and the setting it carries. The candela of every strobe and the sensitivity setting of every detector belong in the record, because the battery calc and the next sensitivity test both depend on them.

Common mistakes

• Undersizing the NAC: too many appliances on too small a wire, so the last strobe drops below 16 VDC and fails inspection.
• Short battery calc: using max rated alarm current instead of the installed candela draw, or skipping the 1.25 margin.
• Wrong device spacing: smoke near supply diffusers, heat spaced for a low ceiling under a high one, ignoring the point-seven rule.
• Wrong signal type: waterflow wired as supervisory, or duct and tamper wired as general alarm.
• Circuits not supervised or wrong class: missing end-of-line resistor, or Class B where the AHJ required Class A or survivability.
• No cause-and-effect test: confirming devices activate but never verifying HVAC shutdown, elevator recall, door release, and relay outputs.
• Record of completion missing or blank: the system passes the test and the document that proves it never gets filled out.

Standards and references

NFPA 72, the National Fire Alarm and Signaling Code, is the controlling document for the devices, the circuits, the power, the cause-and-effect, and the testing. It sets device spacing and placement, the IDC, SLC, and NAC circuit classes and pathway survivability, the secondary-power duration of 24 hours standby plus 5 minutes alarm, the audibility and visible-notification requirements, the 100 percent acceptance test and the record of completion, and the inspection, testing, and maintenance frequencies in its Chapter 14. The specific section numbers shift between editions, so confirm them against the edition the jurisdiction has adopted before citing them on a submittal.

The wiring methods come from NFPA 70, the NEC, mainly Article 760 for power-limited and non-power-limited fire alarm circuits and the FPLR and FPLP cable types. Where a fire alarm system is required at all comes from the building and fire codes, the IBC and IFC, by occupancy and size. The ADA and the federal accessibility guidelines drive visible notification, now harmonized with NFPA 72. The HVAC interface ties to NFPA 90A. Every device and appliance carries a UL listing, and the listing and the manufacturer's published instructions govern the installation details. Above all of it, the AHJ adopts the editions, writes the amendments, and has the final say. Cite the standard that controls the point, and let the AHJ and the project specification override the rule of thumb.

Units, terms, and acronyms

Fire alarm runs on its own shorthand, and the same idea reads three ways across a panel manual, a print, and a spec. Strobe intensity is candela, written cd. Audibility is decibels A-weighted, dBA. Battery capacity is ampere-hours, AH. Notification appliances need a minimum operating voltage in volts DC, commonly 16 VDC.

The circuit and device acronyms are the vocabulary you cannot skip. Learn them once and the prints stop being a foreign language.

FACP / FACU

Fire alarm control panel, or control unit, the system's brain that monitors inputs and drives outputs

IDC

Initiating device circuit, a conventional input circuit reporting by zone

SLC

Signaling line circuit, the addressable data loop the panel polls device by device

NAC

Notification appliance circuit, the power circuit feeding horns and strobes

Candela (cd)

The effective intensity of a strobe flash, set by room size and mounting per NFPA 72 tables

Class A / Class B

Loop fed from both ends that survives one break, versus single path that loses everything past a break

Supervisory vs alarm

A condition that impairs protection, like a closed valve, versus a confirmed fire condition like waterflow

ITM

Inspection, testing, and maintenance, the ongoing program under NFPA 72 Chapter 14

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What fire suppression do data centers use?
• What is the difference between grounding and bonding?
• What is an intrusion alarm system?
• What is aspirating smoke detection?
• What are the three fire pump test points?