Fire alarm commissioning before handover proves that what was installed on site actually works as designed and approved: every detector reports to the right address, every sounder is heard in its area, and every output in the cause and effect matrix happens when it should. A system handed over without this sequence is an installed system that nobody has proven yet.
This article walks through the acceptance test sequence step by step, from cable pre-tests to the witness test and the record of completion, and what should be recorded at each stage. It is the work covered by AZEIDK’s testing, commissioning and handover service.
Key takeaways
- NFPA 72 (2025 edition) is a technical reference adopted by the designer. Chapter 14 covers inspection and testing, and Chapter 7 covers documentation, including completion documents.
- Testing starts with cables and circuits before devices are connected, then each device, then integrated functions.
- The approved cause and effect matrix is the reference for testing interfaces with lifts, HVAC, doors and suppression systems.
- The system is run on secondary power with the primary supply disconnected, and the battery calculation is checked against the batteries actually installed.
- Sound levels are measured in every area with a calibrated meter, not taken from design calculations.
- Any change to devices, wiring or software after acceptance means retesting the affected functions.
In this article
- What fire alarm commissioning and acceptance testing mean
- Before testing: what must be ready
- Stage 1: cable and circuit tests
- Stage 2: device-by-device testing
- Stage 3: cause and effect and interface tests
- Stage 4: secondary power and batteries
- Stage 5: sound level measurements
- The acceptance test sequence at a glance
- Witness testing and the record of completion
- Common fire alarm commissioning mistakes
- How AZEIDK supports testing and commissioning
- Frequently asked questions
- Conclusion
What fire alarm commissioning and acceptance testing mean
Projects use several terms for roughly the same stage. Commissioning is preparing, programming and adjusting the system until it works. The acceptance test is the full test of a new system to confirm it matches the design and the standards. The witness test repeats the tests, or a sample of them, in front of the consultant, the owner’s representative or the authority.
In Saudi Arabia, the Saudi Fire Protection Code SBC 801 applies as part of the Saudi Building Code. Its 2024 edition has been mandatory since 30 June 2025, and the General Directorate of Civil Defense monitors compliance with fire prevention and protection requirements. Detailed test methods usually come from NFPA 72, in the edition set by the project documents or the code’s referenced standards. A published review of the 2025 edition notes that Chapter 7 requirements for design documents, shop drawings and completion documents apply when a code, specification or contract calls for them, so the project specification should say so explicitly.
Before testing: what must be ready
Most handover delays come from starting tests before the system is ready. Before the first test, confirm:
- Approved shop drawings are updated with device locations, addresses and loop routes.
- The cause and effect matrix is approved by the consultant, and the panel program is loaded to match it.
- Location text in the program matches the real floor and room names.
- Work by other trades is complete: lifts, HVAC units, magnetic door holders, and flow and tamper switches on the sprinkler system.
- A written test plan sets out the sequence, the people responsible and the recording forms.
- Building occupants and any monitoring station are notified before testing starts and again when it ends.
Stage 1: cable and circuit tests
According to an Electrical Contractor article based on the NFPA 72 test methods table, metallic conductors are checked for stray voltage, ground faults and short circuits, and loop resistance is measured. Refer to the corresponding table in Chapter 14 of the edition adopted for the project for the details of each test. Measured loop resistance is compared with the manufacturer’s limit for the panel and circuit, as given in the datasheet or installation manual.
Insulation resistance testing with a megger is common practice on projects, but it is done on cables with devices and the panel disconnected, because the test voltage can damage electronics. Record each circuit number, cable type, measured value and the name of the person who tested it.
Once devices are connected, circuit supervision is tested: opening a connection or applying a deliberate short must produce a trouble signal at the panel with the correct location. On addressable loops, each loop isolator is tested with a deliberate short to confirm it isolates only the affected section and keeps the rest of the loop working.
Stage 2: device-by-device testing
In an acceptance test every device is tested, not a sample. Each device must do two things: operate, and appear at the panel with the correct address and text.
- Smoke detectors: according to QRFS, citing the NFPA 72 (2019 edition) test methods table, they are tested in place with smoke or a listed product acceptable to the manufacturer, or per the manufacturer’s published instructions, to confirm smoke enters the sensing chamber. Pressing the test button is not a substitute.
- Heat detectors: restorable types are tested with a suitable heat source. Non-restorable types are not heat tested; follow the manufacturer’s instructions.
- Manual call points: each one is operated with its test key or test element and checked for reset. For installation criteria, see our article on manual call points and notification appliances.
- Notification appliances: confirm every sounder and strobe operates, and that visual and audible signals are synchronised within each zone.
- Modules: input modules are tested by operating the monitored device itself, and output modules by checking the relay change of state on command.
Stage 3: cause and effect and interface tests
With every device proven individually, the system is tested as a whole. An input device is operated in each zone, and every output in the approved cause and effect matrix is checked: which appliances sound, which fans stop, which doors release, and when. Any difference between the matrix and actual behaviour is logged, corrected and retested.
| Interface | What is tested | Evidence |
|---|---|---|
| Lifts | Recall to the designated floor on alarm, and to the alternate floor when the alarm is on the recall floor | Floor where each lift stopped |
| HVAC and smoke management | HVAC shutdown or smoke fan start by zone | Status of each unit before and after alarm |
| Doors and access control | Release of magnetic holders and emergency exit locks | List of doors tested with the result for each |
| Sprinklers and fire pumps | Flow switch (alarm), valve tamper (supervisory) and pump status signals | Address and signal type shown at the panel |
| Gas suppression | First and second stage alarms, delay, hold and abort | Tested with release actuators disconnected per the manufacturer |
| BMS or voice evacuation | Signals received and messages played in the right zones | Event logs from both systems |
Stage 4: secondary power and batteries
The primary supply to the panel is disconnected to confirm that it transfers to batteries without losing any function, and that a power fault is indicated. The system is then put into alarm on batteries to confirm they carry the full alarm load.
This is also the point to check the approved battery calculation against reality: installed battery capacity, date of manufacture, and actual device loads after any changes in quantity during installation. An Electrical Contractor article gives the NFPA 72 secondary power requirement as 24 hours of standby followed by 5 minutes of alarm, or 15 minutes for voice evacuation systems, plus a safety margin added by the designer. Confirm the required duration in the edition adopted for the project. For details, see fire alarm battery calculation.
Stage 5: sound level measurements
Design calculations assume an ambient noise level, but site measurement is what proves compliance. A published review of the 2025 edition of NFPA 72 states that in public mode the alarm signal must exceed the ambient sound level by at least 15 dB in every occupied area. The code sets additional requirements for sleeping areas, and another review of the 2025 edition notes that low-frequency sounders, where required, must produce the 520 Hz signal. The designer applies the values in the edition adopted for the project.
- Use a calibrated sound level meter and record readings in dBA.
- Measure ambient noise first, under normal operating conditions, then the alarm level.
- Record a reading for each zone or room with its measurement location, with doors closed in rooms where they are normally closed.
- Compare results with the design sound level stated in the documents.
The acceptance test sequence at a glance
| Stage | Test | Deliverable |
|---|---|---|
| Pre-tests | Readiness, program and matrix review | Signed readiness checklist |
| Cables and circuits | Stray voltage, ground, short, loop resistance, insulation | Measurement sheet per circuit |
| Supervision | Open circuit, short and isolator tests | Record of trouble signals |
| Devices | 100% test with address and text check | Device-by-device log |
| Integrated functions | Cause and effect and interfaces | Matrix signed with results |
| Power | Battery operation and calculation check | Voltage readings and result |
| Sound | Ambient and alarm level | dBA readings sheet |
| Witness and documents | Full or sample retest with stakeholders | Record of completion and handover minutes |
Witness testing and the record of completion
The consultant or owner’s representative should be invited only after the contractor has passed its own tests in full. The witness test is not the place to find faults. Its scope is set by the consultant, the contract or the authority: a full repeat, or a sample of devices with every matrix scenario.
The record of completion brings system data and test results into one document signed by those responsible, and Chapter 7 of NFPA 72 sets its form and content. It goes with the loaded program version and a backup copy, and the as-built drawings. For as-built drawings and O&M manuals, see fire systems handover documents.
After acceptance, adding or removing devices, changing wiring or changing site-specific software requires, under NFPA 72, reacceptance testing of the functions affected by the change.
Common fire alarm commissioning mistakes
- Starting tests with the cause and effect matrix not approved or not loaded on the panel.
- Testing smoke detectors with the test button only, without smoke or a manufacturer-accepted product.
- Location text that does not match room names, sending responders to the wrong place.
- Insulation testing with devices still connected.
- Relying on sound calculations without measuring after finishes are complete.
- Not rechecking the battery calculation after devices are added during installation.
- Changing the program after the witness test without retesting the affected functions.
- Handing over without a copy of the program and the passwords.
How AZEIDK supports testing and commissioning
Through its testing, commissioning and handover service, AZEIDK reviews system readiness and pre-commissioning checklists, runs function, signal and control tests suited to each system type, checks the response sequence and the interfaces in scope, records results and observations with retests where needed, and prepares the operation and inspection reports required for handover. Under its fire alarm solutions, AZEIDK supplies Context Plus conventional and addressable panels and devices. For Context Plus programming and commissioning steps, see the Context Plus application guide.
Frequently asked questions
Is testing a sample of devices enough at handover?
In the acceptance test of a new system, every device is tested. The witness test may cover a sample chosen by the consultant or authority, provided the contractor has already tested the whole system.
What is the difference between commissioning and acceptance testing?
Commissioning prepares, programs and adjusts the system. Acceptance testing is the documented proof that the complete system matches the design and the standards. The second follows the first.
Can a gas suppression system be tested with the fire alarm?
Yes. The alarm, delay, hold and abort sequence is tested, but with release actuators disconnected per the manufacturer so that no discharge occurs.
When must the system be retested after handover?
When devices are added or removed, wiring is changed or site-specific software is modified, the affected functions are retested, in addition to the periodic inspection programme.
Who signs the record of completion?
Those responsible for installation and testing sign it as the project documents require, the consultant reviews it, and a copy is kept with the system documents held by the owner.
Conclusion
Fire alarm commissioning before handover follows a logical order: sound circuits first, then devices that work and report the right addresses, then a system that behaves as the matrix says, then secondary power and sound levels proven by measurement, then documents everyone signs. Any stage cut short now comes back later as snag lists or alarms nobody can interpret.
Is your fire alarm handover coming up? Send us the approved drawings and the cause and effect matrix, and we will review readiness and the test plan with you before the witness test. Talk to the AZEIDK team
