How Many Runway Lights Can Be Unserviceable?

Ask how many lights can be out and the honest reply is another question: which system, and for which operation. The rules do not give a runway a single pass mark. They give each lighting system a floor for each category of approach or take-off, and they define what an unserviceable light is in terms of the light that leaves the fitting, not the lamp inside it. This page sets out those floors from the clauses, then what happens when a system drops below one, how often the lights should be measured, and how aerodromes produce the figure.

Which rule applies

In the EU the binding text is Regulation (EU) No 139/2014 , point ADR.OPS.C.015(b). It requires the aerodrome operator to run a preventive and corrective maintenance programme for the individual lights and the lighting systems, and it states the maintenance objectives system by system. EASA’s guidance material, GM1 to ADR.OPS.C.015(b), repeats the figures as one table, and GM3 explains that the serviceability levels define maintenance performance objectives.

ICAO Annex 14, Volume I , Section 10.5, holds the same objectives as Standards in 10.5.7 to 10.5.12. The measurement methods and intervals around them, in 10.5.3 to 10.5.6, are Recommendations. Outside the EU those Annex values reach an aerodrome through its national regulations.

In the United States, AC 150/5340-26C , Appendix A, Table A-8, gives operating limits for each system. Its CAT II/III figures for the runway lights match the European ones, 95 % for the edge and centre line lights and 90 % for the touchdown zone. Its other limits are set differently, and they are covered further down.

When a light counts as unserviceable

ADR.OPS.C.015(c) makes a light unserviceable in two cases.

• Output. The average intensity of the main beam is below 50 % of the value in the certification specifications. Where a unit is designed to a higher main beam intensity than the specification asks for, the 50 % applies to that design value. Annex 14 10.5.1 says the same against the figures in its Appendix 2.
• Colour. The filter is missing or damaged, or the light does not produce the correct colour.

The definition is optical. A lamp that draws its current and lights up can still be an unserviceable light if rubber, dust or de-icing residue has cut its output by half, or if a red filter has faded or cracked. In-pavement lights on the centre line and in the touchdown zone are the usual case, because every landing runs over them. EASA’s GM2 to ADR.OPS.C.015(b) adds that LED output falls and its colour shifts over time, so a long-life LED is not a reason to “install and forget”.

GM1 to ADR.OPS.C.015(b) and (c) recommends a second, higher threshold. The maintenance programme can define a maintenance level, always above 50 % of the specified intensity, at which corrective work is scheduled before the light reaches the point where it has failed. The FAA sets its own trigger in AC 150/5340-26C. Paragraph 5.3.4.1.3 says an in-pavement runway light giving less than 70 % of the minimum output required at maximum intensity should be targeted for cleaning or maintenance, and paragraph 5.3.2.1.2 applies the same 70 % figure to high intensity elevated edge lights.

The figures per system

Each percentage is a floor, not a target. Every one of these points opens with the same objective: during any period of the operation, all approach and runway lights are serviceable. The percentage is the least that is acceptable while that objective is being worked towards.

The floors are per system, and that is where a whole-runway count misleads. A runway with long centre line and edge rows has hundreds of units, while the runway end bar has a handful. Merge them and a few dark or discoloured end lights vanish inside a healthy average, while the runway end lights on their own sit below 75 %. The touchdown zone can hide the same way.

For CAT I, GM1’s table shows the touchdown zone lights at 85 % in brackets where they are fitted, while the regulation’s list in (b)(6) does not name them.

Adjacency and the pattern

The percentage is not the only test. Under CAT II/III, ADR.OPS.C.015(b)(2) says the lights allowed to be unserviceable must not alter the basic pattern of the system, and (b)(3) says an unserviceable light may not sit next to another one. The exception is a barrette or a crossbar, where two adjacent unserviceable lights are allowed. GM3 defines adjacent for barrettes, crossbars and edge lights: consecutive lights in the same barrette or crossbar laterally, or in the same row of edge lights or barrettes longitudinally. Annex 14 10.5.7 carries the same rule and the same note.

For CAT I, (b)(6)(ii) forbids adjacent unserviceable lights unless the spacing is significantly less than required. For take-offs, both below and at or above 550 m RVR, (b)(7)(ii) and (b)(8)(ii) forbid adjacent unserviceable lights with no exception.

Two ground systems carry limits of their own. A stop bar at a runway-holding position used with a runway for operations below 550 m RVR may have no more than two unserviceable lights, and not two adjacent ones unless the spacing is significantly tighter than required, under (b)(4). On a taxiway used below 550 m RVR, no two adjacent centre line lights may be unserviceable, under (b)(5). Annex 14 10.5.8 and 10.5.9 set the same objectives but tie them to 300 m RVR, so the EU rule reaches further.

A PAPI is judged by unit rather than by percentage. AMC1 to ADR.OPS.C.015(b) and (f) treats a unit as unserviceable when more than one light is out in a unit of three or more lights, or any light is out in a unit of fewer than three. When a red filter is missing, damaged or the wrong colour, all the lights behind it should be switched off until it is fixed.

Take-off below and above 550 m RVR

A runway can be a low visibility runway without any CAT II/III approach. If take-offs are allowed in runway visual range below 550 m, (b)(7) applies: 95 % in the runway centre line lights where provided, 95 % in the edge lights, 75 % in the end lights, and no two adjacent unserviceable lights. At 550 m RVR or more, (b)(8) asks for 85 % in the edge and end lights, again with no adjacent pair.

So a CAT I runway that also hosts low-RVR take-offs is held to 95 % in its centre line and edge lights while those take-offs run, whatever its approach category.

The FAA table is laid out differently. Table A-8 gives 95 % for centre line lights and 90 % for touchdown zone lights, 85 % for edge lights except on CAT II and CAT III runways, which need 95 %, 75 % for end lights, and 75 % for threshold lights on VFR and non-precision IFR runways. Its note 1 repeats the pattern and adjacency rule in the same terms as Annex 14.

What happens when a system falls below its objective

Nothing in ADR.OPS.C.015 downgrades a runway automatically. The consequence comes through three other provisions.

The operator reports it. ADR.OPS.B.045(c) requires the aerodrome operator to inform the aeronautical information services provider and the air traffic services provider of any change in the status of equipment and facilities that affects low visibility operations. ADR.OPS.A.057(b)(5) and (6) require a NOTAM for significant changes to visual aids and for the interruption of, or return to operation of, major components of the lighting systems. Under (c)(2) a NOTAM about an unserviceability should estimate how long it will last.

The flight operations rules set the restriction. AMC1 to ADR.OPS.B.045(c) lists the equipment failures to report under low visibility procedures and their effect. For approach lighting, runway lighting, runway centre line lighting, runway edge lighting, touchdown zone lighting and taxiway lighting the effect given is a restriction depending on flight operations rules. In practice that means the category or the RVR minimum the crews can use is limited, and the limit is set in the flight operations rules, not in the aerodrome rules.

Monitoring tells ATS. CS ADR-DSN.S.890(d) and (e), matching Annex 14 8.3.4 and 8.3.5, recommend that on runways used below 550 m RVR the monitoring shows when any element falls below its maintenance objective, relayed to the maintenance crew, and when it falls below the level below which operations should not continue, relayed to air traffic services and displayed prominently. GM1 to CS ADR-DSN.S.890 says the competent authority sets that second level. Ask your authority for it, because it is the figure that actually stops the operation.

How often the lights should be measured

The regulation itself does not print an interval. ADR.OPS.C.015(f)(1) requires the maintenance programme to include inspections and checks of each element and each system at defined intervals appropriate to the operation. The intervals below are what EASA and ICAO offer as the way to meet that.

AMC1 to ADR.OPS.C.015(b) and (f) sets:

• Daily ground checks, at least for light failures, breakage, gross misalignment and correct operation of the intensity control.
• Photometric measurements at appropriate intervals, for every runway type, set in the maintenance programme.
• For CAT II/III, visual inspection and in-field measurement of intensity, beam spread and orientation of the approach and runway lights, a check of the electrical characteristics of each circuit and a check of the intensity settings used by ATC. All lights should be measured as far as practicable, with a measuring unit accurate enough to analyse individual lights.
• A measurement frequency based on traffic density, local pollution, the reliability of the equipment and the results of earlier measurements, and not less than twice a year for in-pavement lights and once a year for other lights.
• A flight check of the approach, runway and taxiway lighting, the VASIS, the aerodrome beacon and the intensity control at least yearly, with special checks after major maintenance and before commissioning.

Annex 14 recommends the same content and the same twice-a-year and once-a-year floors in 10.5.3 to 10.5.6, naming a mobile measuring unit. Being AMC and Recommendations, these are not the binding interval, but an authority will expect a reason for anything less frequent.

The FAA asks for more on in-pavement lights. AC 150/5340-26C paragraph 5.3.4.1.3 notes that weekly checks may be necessary for centre line lights because of rubber deposits, especially on CAT II and CAT III runways. Paragraph 5.3.2.1.2 asks for photometric testing of elevated lights on a precision approach runway not less than twice a year. Both paragraphs ask for testing before cleaning or relamping and again afterwards, to show the work had an effect.

How the figure is evidenced

No single method produces the per-system figure. Most aerodromes combine several.

Patrols find dark lights, broken units and gross faults, and they meet the daily check. They cannot put a number on a light that is lit but dim.

Circuit monitoring is the live alarm while the operation runs, and it is what reports status to ATS. It works from the electrical side. The FAA regulator monitor specification lists alarms such as a drop of 10 % or more in the volt-amperes delivered to the circuit and the failure of a preset number of lamps, adjustable from 4 to 10. A lamp behind a dirty lens or a faded filter draws its current and reads as healthy on the constant current regulator .

Ground photometric vehicles drive over or alongside the lights and measure the intensity of inset and elevated lights with a calibrated photometer as they pass. This is the instrument AMC1 and Annex 14 10.5.5 describe, and its figures are absolute, so they can be compared directly with the 50 % threshold. It needs the runway surface for the run. The photometric measurement guide covers how it works.

Flight checks look at the approach and runway lighting and the PAPI from the approach, as the pilot sees them. They judge the pattern and the systems as a whole rather than every unit.

A drone survey records every unit in a system as lit or dark, its colour against what its position calls for, its output relative to its neighbours and to the commanded intensity step, and its RTK-surveyed position. It is flown between movements in coordination with the tower, with the lights switched on at the steps being checked. That finds low and wrong-colour units and gives the maintenance team their positions. It does not measure intensity in candela. If every light in a row has dimmed evenly, a comparison between neighbours will not show it, and a statement against the specified value needs a calibrated photometer.

A sound programme uses the drone or the patrol to find suspect units, the photometer to put absolute figures on the systems that need them, and monitoring for the live state. The airfield lighting inspection page sets out how a drone survey fits that programme, and the per-system services are runway edge lights , runway centre line lights , threshold and runway end lights , approach lights , taxiway lights and the PAPI . For the trade-off with an aircraft, see drone survey versus flight inspection .