The Dimming Point: What Emergency Lighting Duration Evidence Must Actually Show

Emergency lighting is one of the quieter areas of a building's compliance file. It is rarely the subject of a dramatic enforcement story, and it is often managed through a familiar annual routine: book the contractor, isolate the circuit, operate the test switch, complete the worksheet, file the certificate. The difficulty is that this routine can produce a record of activity rather than a record of performance. When an auditor asks whether the premises can demonstrate that escape lighting remained effective for the required duration, a tick in a box is not an answer. The file has to show what was tested, what was measured, and what happened when the fitting began to dim.

A switch operation is not a lighting outcome

Monthly function tests have a legitimate role. They confirm that the circuit operates, that lamps illuminate and that obvious failures are identified between annual checks. They do not establish duration. A fitting can illuminate for thirty seconds and still fail the substantive claim that it will provide usable light for the full rated period in an emergency.

The annual duration test is therefore a performance test, not an administrative exercise. The record should identify the test method, the start and end times, the fittings included and any fittings excluded with a reason. It should also distinguish between a test that was completed and a test that was passed. Those are different statements. A completed worksheet may show that someone operated the system; a passed test requires evidence that output remained adequate at the end of the duration period.

Lux readings are the point of the exercise

Where the design and the relevant standards require specified illuminance levels, the audit file should contain measurements rather than assurances. That means identifying the measurement points, the instrument used, its calibration status and the conditions under which readings were taken. A reading taken in ambient daylight is weak evidence. A reading taken at the start of the test, with no reading near the end, does not demonstrate duration.

The useful pattern is a before-and-after record: initial illuminance at defined points, then illuminance at or near the end of the rated duration. If the fitting is designed to provide a minimum level on an escape route or in an open area, the record should show that the level was maintained for the period claimed. Where measurements are not taken, the file should say why and what alternative evidence supports the conclusion. A certificate that simply states 'duration test satisfactory' invites the auditor to ask what satisfactory meant in practice.

Battery age belongs in the asset record

Emergency lighting batteries degrade. A site can pass an annual test one year and show widespread failures the next if the battery population is ageing together. The audit file should therefore connect the test result to the asset register. Each fitting, or at least each relevant type and batch, should carry an installation date, battery replacement date, battery type and expected service life.

This is not merely an engineering nicety. It allows the responsible person to see whether a good result is credible and whether a poor result is isolated or systemic. A site with a large cohort of batteries approaching the end of their service life should be able to show a replacement programme, not just a reassuring test sheet. Conversely, a cluster of failures in fittings of a similar age is a finding that should trigger a planned response rather than a series of disconnected repairs.

Subcontractor evidence needs verification

Specialist contractors are often necessary for duration testing, particularly where central battery systems or complex control equipment are involved. The certificate they issue is a summary. The underlying evidence should still belong to the site. That includes the method statement, the competence and calibration details for the instruments used, the raw readings, the list of defects and the remedial actions proposed.

Verification is not an accusation of bad faith; it is part of the evidence chain. Attendance should reconcile to site access records, permits or work orders. The scope should reconcile to the asset list. If the contractor tested a sample, the file should explain the sampling rationale and show how the sample rotates over time. If remedial work was carried out, the file should contain the retest evidence, not merely an invoice reference. An auditor is entitled to ask how a certificate can be relied upon if the site cannot show which fittings it covered.

The worksheet is a claim; the record is corroborated

A completed worksheet is a claim made by the person completing it. An auditable record is a claim that can be reconstructed by someone else. The difference is usually a small number of items: unique fitting references, measured readings, dates, signatures, exceptions, remedial actions and closure evidence.

A useful test is to select a fitting at random and ask whether the file can answer five questions. Where is it? When was its battery last replaced? When was its last monthly function test? When was its last duration test and what were the readings? If it failed, what was done and when was it retested? If those questions cannot be answered without a telephone call, the record is not yet in audit condition.

Preparing the file before the auditor arrives

A practical pre-audit review should reconcile the emergency lighting asset register to the test records, check that every recorded defect has a closure date, confirm that instrument calibration certificates are current and ensure that any excluded areas are documented with justification. It should also look for patterns: repeated failures by manufacturer, age or location; missed monthly tests; certificates that arrived without readings; and remedial works that were invoiced but not retested.

The ninety-minute claim is ultimately a performance claim. It is not proved by a flick of a switch or a completed form. It is proved by a file that lets a third party see the fitting, the measurement, the battery history, the contractor's evidence and the response to failure. Where that chain exists, the audit is straightforward. Where it does not, the record is simply a record of someone having tried the lights.