How to Run a Photometric Measurement of Airfield Lighting

What you need before you start

A photometric campaign fails on preparation far more often than on measurement. Everything in this list has to exist before a date is booked, and none of it is a step in the procedure.

The specification the fixtures were installed to. Intensity minima and colour boundaries are properties of the fixture class, and they live in the fixture specification and in the aerodrome design standards. Without them there is no criterion to judge a reading against, and the campaign produces numbers with no verdict attached.

A calibrated instrument with a current record. Whatever the carrier, an absolute value of light intensity in candelas has to trace back to a calibrated photometer with a valid calibration. Check the certificate date before the visit rather than after the report is written.

Control of the intensity steps. Somebody has to be able to drive the constant current regulator to a commanded step and hold it there for the duration of a run. If that person is not available, the campaign cannot happen.

An agreed fixture inventory. A list of the fixtures to be measured, such as the runway edge lights of one circuit, with identifiers that will still mean the same thing next year. Measuring a row and reporting positions rather than identifiers makes the second campaign incomparable with the first.

Airside access and coordination. Tower coordination, an escort where required, and a window. For a ground run this is a booked occupancy of the surface.

A dark enough sky, if the method needs one. Some photometric methods require low ambient light, which pushes the work into a night window and changes the staffing.

What the campaign produces

A dated record for one runway’s lighting in which every measured fixture carries an intensity value, a colour assessment, the intensity step the reading was taken at, the instrument that took it, and the criterion it was judged against. Fixtures outside their criteria are named with the margin by which they miss. The record is the deliverable, and the verdict is a line inside it rather than the whole of it.

Checking the campaign actually worked

The completion check is not whether the report arrived. It is whether the record answers the reconstruction questions on its own.

Pick one fixture from the middle of the inventory, at random, and try to answer five questions from the record alone. What did it measure. At which intensity step. With which instrument, calibrated when. Under what conditions and taken by whom. And against which criterion was it judged. If any answer requires asking somebody, the record is incomplete and the gap should be closed while the team is still assembled.

Then do the same for a fixture reported as outside criteria, and check that the margin is stated rather than the verdict alone. A record that says a fixture failed is less useful than one that says by how much, because the second one tells a maintenance team whether it is chasing a cleaning job or a replacement.

When it goes wrong

Readings vary along the array with no pattern. Suspect the pass geometry before the fixtures. Distance, alignment and speed all change what the instrument sees, and drift in any of them scatters results in a way that looks like fixture variation. Re-run one circuit deliberately holding geometry and compare.

Everything reads low. Check the intensity step first and the instrument second. A circuit sitting one step lower than the campaign specified produces a uniformly low set of readings that looks alarming and means nothing.

Colour readings disagree with the eye. The eye is a poor absolute judge and a good relative one. If measured chromaticity says a red unit is drifting while the eye says it looks fine, trust the instrument, because filter ageing moves colour slowly and perception adapts to it.

One fixture reads far above its neighbours. Check for a replaced unit of a different class or a different age in an otherwise uniform row. A new fixture among old ones is not an error, and it is a useful indication of how far the rest have degraded.

The regulator will not hold a step. That is a finding, not an obstacle. Log it, stop the affected circuit, and pass it to the electrical team. A campaign that measures through an unstable supply produces numbers nobody can defend.

The window closes early. Stop cleanly at a circuit boundary and log where you stopped. A campaign completed across two windows with clean boundaries is fine. A campaign with a circuit measured half in one window and half in another is not comparable with itself.

Variations on the method

Ground-based mobile measurement. A photometer on a vehicle traverses the surface and measures fixtures directly in candelas. This is the method the acceptance criteria were written for, and it is a booked runway occupancy. Published vendor material describes covering a full precision approach runway in both directions in under two hours.

Airborne relative measurement. An airborne camera flown between aircraft movements measures colour ratios and relative output across every unit in an array without booking a closure, and needs step 6 above to state anything in candelas. It covers more fixtures per hour and makes a weaker absolute claim, which is the trade set out in airfield lighting inspection .

Fixture-by-fixture tripod measurement. The most controlled option and the slowest. Useful for resolving a disputed result on a small number of units, and impractical as a way to cover a whole array.

Approach aid measurement. Glide slope indicators are a special case, because the parameter that matters most is an angle rather than an intensity, and an angle is geometry. The procedure for those is closer to the one on the PAPI inspection page than to this one.

Whichever variation you use, the record requirements do not change. If you would rather have the campaign run for you, TarmacView scopes one against your runway and circuit count.