The Unpolled Outstation: Making Disabled Refuge Intercoms Auditable in 2026

A system hiding in plain sight

A disabled refuge intercom is easy to install and even easier to forget. It sits in a stairwell, lobby or protected corridor, usually as part of an emergency voice communication system, and it exists for a simple reason: to let someone who cannot use the stairs speak to the person managing the evacuation. In many buildings the hardware is present, the commissioning certificate is filed, and the system then drifts outside the routine that keeps fire detection honest.

That is the quiet risk. A weekly fire alarm test proves that a call point operates and the alarm sounds. It does not prove that an outstation in a refuge area can be heard at the master station, that the master station can be heard at the outstation, or that the system reports a fault when a handset is pulled out. Emergency voice communication systems are covered by BS 5839-9, not by the same routine as the fire detection and alarm system. Treating one as evidence for the other is a category error.

Where the routine breaks down

The common failures are rarely dramatic. They are the ones that survive until someone needs the system in anger.

• The master station has a fault indication that nobody has been trained to interpret.
• An outstation is labelled with a room name that no longer matches the floor plan.
• A handset has been removed, damaged or left off the cradle.
• Speech is audible but not intelligible, particularly in a noisy lobby or stair core.
• The battery standby has never been tested under load.
• An IP or networked system has a link that only fails when part of the network is isolated.
• The cause-and-effect matrix has never been checked against the actual building layout.

None of these require a malicious actor. They require time, footfall and the assumption that the fire alarm contractor is already looking after it. In practice, the refuge system may sit between the fire alarm maintainer, the electrical contractor and the facilities team. The more interfaces it has, the more likely it is that no single party owns the full test.

What a defensible test looks like

A defensible routine starts with an inventory, not a clipboard. Each outstation should have a unique identifier, a location that matches the current floor plan, and a record of what it is intended to do. The master station should be tested as a control point, not just as a power supply. The test should confirm two-way speech, not merely that a call is registered.

For each outstation, the operative should confirm:

• the call is received at the master station without ambiguity;
• speech is intelligible in both directions;
• the master station can identify the calling location;
• the fault or reassurance tone operates as designed;
• the handset and cradle are intact and correctly stored;
• the outstation reports a fault when the line or handset is disturbed;
• the master station logs the event with the correct time and location.

The battery standby should be tested under load in line with the manufacturer's instructions and the system design. A voltage reading on a healthy mains supply proves very little. The same applies to networked systems: a ping is not a test of speech quality, and a green status icon is not evidence that the audio path works when the network is under stress.

The test record should be specific. A line that says 'refuge system tested' is nearly useless. A record that names the outstation, the master station, the operative, the date, the result and the fault reference is evidence. If a fault is found, the record should show what was done to make the system safe in the interim, not just that a repair was ordered.

Evidence the auditor will ask for

The responsible person needs to be able to show a chain, not a folder. That chain normally includes the commissioning certificate and design information, the operation and maintenance manuals, the current test schedule, the individual test records, the fault and rectification log, and evidence that the people doing the work are competent for the system they are testing.

Change control matters more than many teams expect. If a floor is refurbished, a refuge area is relocated, or a master station is replaced, the inventory and the cause-and-effect information should change with it. An outstation that still appears on the schedule but no longer exists is not a minor administrative problem. It is a sign that the evidence cannot be trusted.

Questions for the contractor

Before renewing a maintenance contract, ask the provider to walk through their test procedure rather than describe their accreditation. Which outstations will be tested individually? How is speech intelligibility assessed? What battery test is performed, and under what load? How are networked faults simulated? What reporting format is provided, and does it include device-level results? Who is responsible for updating the inventory after a building change?

A competent contractor should welcome those questions. A contractor who answers only in terms of 'annual servicing' and 'fire alarm cover' is probably not testing the refuge system at all.

The silent refuge is not usually silent because the equipment has failed outright. It is silent because nobody has run the test that would make the failure visible. In 2026, with fire safety scrutiny as it is, that is not a gap that can be left to assumption.