8.2 Acceptance Testing and Witness Coordination
Key Takeaways
- On a new installation, NFPA 72 acceptance testing requires 100% of initiating devices, notification appliances, and controlled devices to be tested and verified.
- Reacceptance testing after a modification tests every function affected by the change plus a sample of unaffected devices; site-specific software changes require 100% of affected functions verified.
- Acceptance is a structured, witnessed demonstration that the installed system meets the approved cause-and-effect, with results recorded on the NFPA 72 Record of Completion.
- Witness coordination may include the AHJ, owner, general contractor, fire alarm contractor, supervising/monitoring station, and the trades for elevators, sprinkler, HVAC, and smoke control.
- The exam trap is treating acceptance as a random device walk instead of a documented demonstration with deficiency tracking and restoration.
Acceptance Test Scope: 100% on New Systems
Acceptance testing is where the commissioned system is demonstrated to the people who need confidence in its operation. NFPA 72 sets the scope clearly for a new installation: 100 percent of all initiating devices, notification appliances, and controlled outputs must be tested and verified for correct operation, along with primary and secondary power, supervision, off-premises signal transmission, and every relationship in the cause-and-effect matrix. There is no 10 percent sampling on a brand-new system; sampling appears later, in reacceptance and in periodic visual inspection allowances.
Typical witness coordination includes the AHJ, owner or owner's representative, general contractor, fire alarm contractor, the supervising (monitoring) station, and representatives for any controlled building systems. NICET does not ask candidates to invent a universal script — it expects them to understand the technician workflow and the documentation that proves it.
Reacceptance Testing After Modifications
When an existing system is modified, NFPA 72 requires reacceptance testing scoped to the change. The core rules candidates should know:
| Situation | Required reacceptance scope |
|---|---|
| New system (initial acceptance) | 100% of initiating devices, notification appliances, and controlled devices, plus full cause-and-effect. |
| Device added, deleted, or relocated | The added/affected device and all functions known to be affected are 100% tested. |
| Site-specific software change | 100% of all functions known to be affected by the change are tested. |
| Software change, unaffected devices | 10% of initiating devices not directly affected (up to a maximum of 50 devices) are tested to confirm the change did not disturb them. |
This 100%-affected-plus-10%-sample rule for software is a favorite exam point: a programming edit that changes one zone still requires retest of every function that edit could touch, plus a 10 percent sample (capped at 50 devices) of the rest to prove nothing else broke.
A Structured, Documented Demonstration
| Acceptance phase | Technician focus |
|---|---|
| Pre-test planning | Confirm approved drawings, sequence, Record-of-Completion form, contacts, access, and readiness. |
| Witness briefing | Explain what will be tested, in what order, and how results are recorded. |
| Functional testing | Verify initiating devices, notification, supervisory inputs, troubles, power, and every interface. |
| Deficiency handling | Record each failure clearly and assign correction and retest. |
| Restoration | Return devices, circuits, interfaces, and the monitoring account to normal. |
| Final records | Complete the NFPA 72 Record of Completion and align as-builts. |
Applied NICET FAS scenario guidance: a question may describe acceptance for a renovated area where the elevator contractor is absent for recall functions. The best answer coordinates the missing party and sequences or reschedules that portion — a technician should not simulate or skip an interface when the required representative or access is missing. Another scenario involves a failed notification circuit during witness testing: the strong answer is to document the deficiency, correct the cause, retest the affected function, and ensure the Record of Completion shows the verified result.
Exam trap: random testing feels active but is often wrong. If an answer skips the approved sequence, ignores test records, or fails to restore the system, it is not the best choice. NICET exams can include exhibits and multi-answer items — read the labels, sequence notes, and deficiency log before choosing. The correct answer may be the one that coordinates people and records, not only the one that names a device. A strong acceptance mindset is simple: prove the intended operation, record it on the Record of Completion, correct exceptions, and restore normal service — a frame that scales across Levels II, III, and IV.
What the Acceptance Test Actually Demonstrates
A full acceptance test is broader than activating each device. NFPA 72 expects verification of the primary (operating) power behavior and the transfer to secondary (standby) power, including the trouble signal that announces loss of AC and the system's continued operation on batteries.
It expects confirmation that the panel correctly annunciates alarm, supervisory, and trouble conditions as separate states with distinct indications, that off-premises signals reach the supervising station with the correct point identity, and that every fault type the supervision is designed to catch — open, short, and ground — produces the proper trouble response on each circuit.
On audible notification, the technician confirms the temporal-three evacuation pattern and adequate sound level; on visible notification, the technician confirms strobe operation and, where required, synchronization. The acceptance test is therefore a demonstration of the entire designed behavior, not merely of device-by-device activation.
Why Documentation Drives the Score
NICET writes acceptance scenarios so that the technically correct field action and the correct paperwork action are inseparable. A choice that activates the right device but never records the result, or never restores the circuit and monitoring account, is incomplete. The Record of Completion exists precisely so that, months later, anyone can read which functions were verified, on what date, by whom, and what deficiencies remained open.
When a witnessed test reveals a failure, the disciplined loop is: log the deficiency with its location and symptom, correct the physical or programming cause, retest the affected function, and update the record so the verified result — not the original failure — is what the closeout reflects. 6 uses for closeout, which is why the exam keeps rewarding it across the whole commissioning-to-turnover cycle.
The candidate who internalizes "demonstrate, record, correct, retest, restore" answers a large share of these items correctly without memorizing each scenario.
What percentage of initiating devices, notification appliances, and controlled outputs must be tested during acceptance testing of a brand-new fire alarm system under NFPA 72?
After a site-specific software change, NFPA 72 reacceptance requires testing 100% of affected functions PLUS what sample of the unaffected initiating devices?
During acceptance testing, a required building-system representative (for example, the elevator contractor) is not available for an interface test. What is the best response?