Every air traffic frequency at this airport is recorded and kept, and every transmitter is monitored for power and frequency. At ten past three a carrier appears on the ground frequency, three hundred hertz low, for forty seconds, and it is not ours.
I am a communications engineer at an airport. I look after the transmitters, the receivers, the recorders and the aerials, and I have done it here since 2009.
The two bits of equipment this turns on.
The first is a spectrum analyser, which shows you what energy is present across a band of frequencies. A transmitter is a spike on it, and the spike sits at a frequency you can read to a few hertz, and our own ground transmitter sits exactly where its licence says because I align it against a reference every year.
The second is the legal recorder. Every controller position and every frequency is recorded continuously and kept for thirty days, because after an incident somebody will want the tape, and the recorder is not optional and not something anybody is allowed to switch off.
So I have two instruments that do not care what I believe: one that says what frequency a signal is on, and one that says what came out of it.
At ten past three in the morning a carrier appears on the ground frequency.
It lasts forty seconds.
It is three hundred hertz below our assigned frequency, which is outside the tolerance our own transmitter is allowed and is well inside what a receiver will still demodulate.
It does this most nights. I have two hundred and ninety one of them logged since I started logging.
The airport is shut at that hour. The ground frequency has nothing on it. No aircraft, no vehicle, no controller, and our own transmitter is not keyed, which I know because the monitor logs every time it keys and it logs nothing.
What I have eliminated, with the dates, because I have had to write this up twice.
Intermodulation. Two strong signals mixing in a receiver front end will manufacture a third one out of nothing, and that is the professional answer to any unexplained carrier. You test for it by attenuating the input: a real signal drops by what you attenuate, a product drops faster. It drops by what I attenuate. It is real energy arriving at the aerial.
Our own kit. The transmitter has been swapped. The standby was put on line for a month in 2017. Both receivers replaced in 2020 when the site was refurbished.
Somebody else's kit. We are not near a taxi rank or a depot, and a private transmitter on an aviation frequency would be a serious offence, and I reported it as exactly that in 2016 and the investigating officer came out twice with a hand held direction finder and found the same thing I did, which is the next part.
The bearing.
We have two receiver sites on the airfield, about nine hundred metres apart, and I can measure the field strength of that carrier at both of them. With a signal that stable you can do better than that: I borrowed a direction finder and took a bearing from each site on eleven separate nights.
The bearings cross.
They cross on the grass between the taxiway and the perimeter fence on the north side, in an area about fifteen metres across, and the eleven nights agree with each other to within that.
There is nothing there. It is mown grass. No cabinet, no aerial, no manhole, no buried cable on the drawings, and I have the drawings and I have also walked it with a cable locator in daylight and got nothing but the duct that runs parallel to the taxiway, forty metres away.
What I do about it officially.
It is logged as an unidentified carrier, nightly, duration forty seconds, frequency minus three hundred hertz. That log has gone upward twice.
The first response, in 2016, was the investigation, which closed as no source identified.
The second, in 2021, was a reply that since the carrier is unmodulated, does not block a transmission, and occurs when the airport is closed, it poses no operational risk and no further action is required.
That is a reasonable position. I would probably sign it myself if it came to me on a different site.
The recorder.
A carrier with no modulation on it records as forty seconds of nothing, and for years that is what I thought it was, because that is what it sounds like.
Last winter I pulled eleven of them off the archive and put them through a bit of software, not to clean them up, which is how people fool themselves, but to measure what is there.
There is something under it. Very low, at the bottom of what that transmitter path can carry, and it is not voice, because voice has a shape and this does not.
It is a regular event. I measured the spacing because measuring is the only thing I am actually good for.
Nought point nine of a hertz. Fifty four a minute, steady for the whole forty seconds, and then it stops with the carrier.
I know what rate that is. I have not written it down on anything that goes to anybody, and the one change I have made in eleven years is that the drive out to the north receiver site is a daylight job now, and if it has to be done at night it gets done by two of us.