How to Diagnose Common Fire Alarm Faults on BS 5839 Systems: A Field Engineer's Guide

If you've been maintaining fire alarm systems long enough, you know the call. Panel's in fault, client's on the phone, and you've got thirty minutes between jobs to figure it out. BS 5839-1 lays out the framework — detection, signalling, control — but it doesn't write your diagnostic checklist for you. That's what this guide is for.

This isn't a rundown of what the standard says. You already know it. This is a practical breakdown of the five fault types that eat the most engineer hours, how to isolate them quickly on site, and when to walk away versus fix in place.

The Five Fault Types That Account for Most Callouts

1. Earth Faults

An earth fault means the system wiring has developed an unintended connection to earth somewhere in the installation. On addressable systems, the panel will usually tell you which loop is affected. On conventional systems, you're isolating zones one by one.

What causes it: Damaged cable insulation (often mechanical damage at trunking entry points), moisture ingress into a device or junction box, or a device that's developed an internal fault.

How to narrow it down on site: Disconnect the loop at the panel. If the fault clears, start reconnecting sections from the panel outward until the fault reappears — that's your section. On large loops, use the T-branch method: disconnect at midpoints to halve the search area. Don't forget to check any repeater panels or ancillary interfaces on the loop.

What the panel reports: "Earth fault — Loop 1" or similar. On some panels (Hochiki especially) the address of the nearest device before the fault location will log, which saves time. Check the panel event log, not just the current fault display.

2. Open Circuit Faults

An open circuit is a break in the loop wiring — the panel can't communicate beyond a certain point. On addressable loops, everything downstream of the break goes missing. On conventional zones, the zone goes off-scan.

What causes it: Cable pulled or cut during other trades' work, a corroded or loose terminal, a failed device (rare but it happens), or a door that's been slammed repeatedly on a surface-run cable.

How to narrow it down on site: On addressable systems, the panel should report the last known address before the break — start your search there. Check the termination at that device first; poor crimps and loose screw terminals account for a disproportionate number of these. On conventional zones, use a test meter at the EOL resistor position to confirm continuity is lost, then work back toward the panel.

What the panel reports: On addressable systems: missing device addresses or "open circuit — Loop X after device Y." On conventional: zone fault or zone off-scan.

3. Zone Short Circuit Faults

A short circuit is the opposite problem — two conductors that shouldn't be touching, are. On conventional systems this takes a zone out immediately. On addressable systems with loop isolators, the panel should isolate the short and maintain the rest of the loop.

What causes it: Cable damage, a device with a shorted terminal board (often after water ingress), someone who's accidentally bridged terminals while working nearby.

How to narrow it down on site: Disconnect devices from the loop in sections. When the short clears on your meter, the problem is in the section you just disconnected. On addressable loops, check whether the built-in loop isolators have operated — the panel event log should show which isolators activated, which narrows the search area considerably.

What the panel reports: "Short circuit — Loop X" with isolator activation events, or conventional zone fault with 0Ω reading.

4. Device Poll Failures (Addressable Systems)

The loop is intact but a specific device isn't responding to polls from the panel. You'll see the address logged as a fault or missing.

What causes it: The device has failed internally, the device address has been set incorrectly (common after a device swap), connection issue at the device base, or a firmware compatibility issue on older panels that've had devices added outside the original spec.

How to narrow it down on site: Go to the device and check the base connection first — a dirty or corroded contact here is responsible for more poll failures than actual device failures. Swap the head/detector with a known-good unit. If the fault follows the unit, it's the device. If the fault stays at the address, it's the base or wiring.

What the panel reports: "Device fault — Address XX" or "Missing device — Loop X, Address XX."

5. PSU and Battery Faults

The panel's power supply or standby battery is outside acceptable parameters. This one gets ignored more than it should — BS 5839-1 requires 24-hour standby and 30-minute evacuation capacity. A degraded battery is a compliance issue, not just a nuisance fault.

What causes it: Battery age (typical life 3–5 years, less in warm plant rooms), a failing charger circuit, mains supply interruption that wasn't recorded, or a PSU output that's drifted out of tolerance.

How to narrow it down on site: Check mains input voltage first. Then measure battery terminal voltage under load if you can. Most modern panels will log a low battery event before it reaches fault threshold — pull the event log and check how long it's been degrading. If the battery's more than 4 years old and showing any fault, replace it regardless.

What the panel reports: "Battery fault," "PSU fault," "Charger fault," or "Mains fail" (the last should have auto-restored if mains is present).

Intermittent Faults: The Hard Ones

Intermittent faults — ones that clear before you arrive or come and go without pattern — are the most time-consuming to diagnose. A few principles that save time:

Pull the event log first, every time. The fault may be clearing before you arrive, but the panel has a record of when it occurred, how many times, and what address was involved. Pattern analysis from the log tells you more than anything you'll find by visual inspection on a single visit.

Check environmental factors. Does the fault correlate with temperature changes, HVAC cycles, or times of day? A detector above a kitchen extraction unit, a cable run through an unheated plant room — these produce intermittent faults that look random until you map the timing.

Document and monitor. If you can't reproduce the fault on site, set the panel to log aggressively, leave a fault record with the client, and return. Trying to fix an intermittent fault you can't reproduce on site often results in unnecessary replacements.

When to Fix on Site vs. When to Escalate

BS 5839-1 Clause 45 on maintenance distinguishes between attended faults (AS 1 — respond and fix within 24 hours) and unattended faults in different risk categories. For most commercial premises:

• If the system is operating in a degraded state (zones isolated, devices missing) and you can't restore full coverage in the visit, document exactly what's impaired and notify the responsible person in writing.
• Any fault affecting more than one zone, the main panel, or the standby power supply should be treated as urgent (AS 1 response).
• Intermittent faults with no immediate cause found should be logged, monitored, and returned to — not ignored.

When you can't restore the system to full operational status, the client needs to know what compensating measures are appropriate while the fault is outstanding. That's part of your job, not just the repair.

A Note on Fault Code Libraries

Panel-specific fault codes vary significantly between manufacturers. A Hochiki ESP "Type 45" event means something different from an Advanced MxPro 5 "Fault 45." If you're working across multiple panel types — and most engineers are — a fault code reference that covers the specific panel you're on saves time and reduces the risk of misinterpreting a log entry.

IFS Pro includes a fault code library covering 900+ codes across Hochiki, Advanced, Kentec, Notifier, and Morley panels, alongside BS 5839-1 calculators and digital site logbooks. If that's useful to you, try it free at incognitofiresecurity.com.