I Retracted This Once. Nine Receivers On Four Continents Recorded It In June.
An unmodulated carrier at 1,347.29 megahertz sits in all nine raw files, across nine consecutive passes, and its Doppler curve tracks the unnumbered object to within three hertz. My June retraction stands. So does this.
A regular column. The views expressed are the writer's own.
LEIDEN, Netherlands — Nine receivers. Nine consecutive passes. One unmodulated carrier at 1,347.29 megahertz, in every raw file.
I am the last person who should be telling you that, which is why the first half of this column is about June.
What I got wrong
Between 2 and 5 June I reported, on four consecutive nights, a repeating twelve-bit sequence on the descending pass of the object this desk has been calling the unnumbered one since the transponder story.
One receiver. One analysis chain. One operator. All three of them were me.
On 6 June a reader in Lyon, Casimir Delahaye-Ntuli, asked for the raw samples rather than the spectrogram. I sent him 19 gigabytes. He took three days.
The twelve bits were the low bits of my own sample counter, folded into the spectrum by an overlap-add I wrote in 2019 and never re-derived. They are present in a file I recorded on 9 June with the antenna disconnected and a resistor across the input.
We retracted on 12 June, in full, at the top of the page and in the error record. Nothing below that line changes it.
The standard
On 15 June I wrote a rule and the standards editor made it binding on this desk: this newspaper will not claim an orbital signal on the strength of one receiver, one software chain or one operator. Ever again.
It reads like housekeeping. It cost me ten weeks and the story I wanted.
Nine strangers
On 17 June we found nine amateur operators on four continents who between them run five receiver architectures: two different commodity chipsets, a converted C-band feed on a garden dish, a university-surplus telemetry receiver built in 1984, and one homebuilt direct-conversion front end.
Each was given a pass prediction — six numbers and a time — and a four-megahertz window, centred 1.29 megahertz off the line we cared about, because a receiver's worst self-generated spur sits at the middle of its own band and I have been caught by that once in public.
They were told to record raw and hand it over unprocessed. No spectrograms, no averaging, nothing opened. Nine drives came back in the post, 214 gigabytes. The nearest thing this desk has done before is in our vault as PX-2191, two amateur trackers' working published entire, and it is why we wanted samples, not pictures.
Three chains then ran on them, written independently by people who have never met, none told what the recordings were of, with forty hours of off-pass recording from the same receivers mixed in unflagged.
What is in the files
A carrier at 1,347.29 megahertz, in nine files of nine, on nine consecutive passes between 19 and 27 June.
It rises and falls through each pass. On 24 June, at closest approach, it slides at 387 hertz a second, a rate no mast on this planet can produce. Fitted against the object's ephemeris, the residual stays under three hertz across all nine passes.
It is in none of the forty off-pass hours. It is in none of the three chains' own test vectors. It is not the artefact I published in June, which lived at baseband and never left the computer.
The three chains disagree about its power by four decibels, and the author of the low reading argues at length that the other two are wrong. It touches neither the frequency nor the curve.
What this is not
We cannot demodulate this, because there is nothing on it to demodulate: it is unmodulated to 46 decibels below the carrier within 25 kilohertz either side, and something quieter or further out would have gone past all three chains. Nor can we put the transmitter on the body. The Doppler fits the object's ephemeris beautifully, and a fit is not a position: anything holding station within about forty kilometres along-track would fit as well. Both sentences belong here rather than further down, where they would read as an afterthought.
An angular fix would settle it in an hour. We have asked twice and been declined twice.
The office that says there is nothing there
Dr. Étienne Harcourt-Bello coordinates this block of spectrum. His reply runs unedited at the foot of this page.
He says there is no assignment at 1,347.29 megahertz, and I believe him. He says nine receivers can share a fault, which is fair, and is why the forty off-pass hours are in the design.
He says the object is uncatalogued because its element set does not stay good for long. That is the only line in his letter I can check against a public document, and the document disagrees: the released summaries hold its perigee inside 1.1 kilometres across nine years and a full solar maximum. Nothing up there is steadier. In telling me why there is no number he has given me a perigee, an inclination and a tracking history back to 1998 that nobody has published anywhere.
Then he asks for our nine raw recordings, complete, for the purpose of interference resolution.
You do not resolve interference that is not there. He knows that, and wrote it anyway.
The drives are his. All nine went out on Thursday, and we asked nothing back.
The retraction stands. So does this.
The six o'clock edition
Three stories a morning, in your inbox before the coffee is made.
How Others Covered This
The same events, as reported elsewhere on the same day. We list what each outlet had that we did not, as well as what we had that they did not — including where we come off worse. Why we print this.
-
The Meridian TelegraphIT IS TALKING TO US: THE 1,347 SIGNAL DECODED AT LAST
Ran a decode of the carrier and printed 41 characters of recovered text within nine hours of our figure appearing.
Had that we did not
The frequency, correctly, and the nine pass times, correctly, both lifted from us with attribution.
Left out
That the carrier is unmodulated, which means there is nothing in it to decode; and that they decoded a JPEG of our spectrogram rather than the samples, which we had offered them in the same email.
-
Signal & LedgerUnassigned Emitter In A Protected Aeronautical Band; Coordinators Will Not Say
Read it as a spectrum-management story for readers who hold licences in the neighbouring block.
Had that we did not
The assignment table for 1,340–1,350 megahertz, set out more clearly than we set it out.
Left out
The Doppler. Their own readers would have wanted the one fact that ties the emitter to a moving object rather than to a mast, and it is not in the piece.
-
They Buriedthis newspaperI Retracted This Once. Nine Receivers On Four Continents Recorded It In June.
Wrote the standard first, then met it: nine strangers, five receiver designs, three analysis chains, raw samples only.
Had that we did not
The June error in full, the nine raw files, the three chains' disagreements, and the office's reply entire.
Left out
This column is built on a correspondent's mistake and then hands that same correspondent the byline on the vindication. We did not give the story to anyone else, and the person best placed to argue that we should have is the person who wrote it. — V. Ashcombe-Doyle, standards editor
Right of Reply
They Buried contacted Dr. Étienne Harcourt-Bello, director of orbital frequency coordination, the Interagency Spectrum Board on 12 August 2026, in writing, with the full draft, the frequency, the nine timestamps and an offer of the samples. Replied 18 August 2026. Printed in full and unedited, including the final two paragraphs.
There is no assignment at 1,347.29 megahertz. There has never been an assignment at 1,347.29 megahertz. The block from 1,340 to 1,350 is aeronautical telemetry, it is protected, and this Board coordinates it, so I am able to tell you with complete confidence that what your correspondents have recorded is not an authorised emission.
What it is, I would suggest, is what it was in June. You published a twelve-bit pattern that turned out to be your own arithmetic. You then retracted it, which was creditable, and you have now built an elaborate procedure whose purpose is to produce a different answer. Nine receivers can share a fault. They share the sky, they share a handful of chipset designs, and they share your pass predictions, which is to say they were all told when to expect something. I have read your description of your three chains and I do not doubt the good faith of any of it. Good faith is not the issue. The issue is that you have designed a study to look for a line and you have found a line.
On the object. It does not carry a catalogue number because it does not meet the criteria for one, and I would ask you not to write that as though a number were being withheld. Uncorrelated targets are common. This one has been tracked intermittently since 1998 under a working designation, it is in a moderately eccentric orbit with a perigee under nine hundred kilometres, and the reason it has never been catalogued is that its element set does not stay good for long enough to be useful to anybody. That is an administrative fact about our thresholds, not a decision about your readers.
I would also ask you not to publish the pass predictions. Coordinated observation of an uncorrelated target by unlicensed persons is a genuine interference risk in a protected aeronautical band, and if this Board is later asked why nine private stations were pointed into telemetry space on nine consecutive nights, the honest answer will be that a newspaper organised it.
Two requests, and then I have finished. First, I would like the nine raw recordings, complete, with the receiver descriptions and the timestamps, for the purpose of interference resolution. You may consider that inconsistent with everything above. I would call it prudent: the Board resolves reported emissions whether or not it believes in them, and it cannot do that from a picture.
Second, you asked whether this office made its own recording on the pass of 24 June, and the answer is yes, from a facility I am not going to name, on equipment considerably better than a dish in a garden. I am not going to tell you what it shows. It is not a public product, it was not made for you, and I would be obliged if you did not characterise my declining as anything other than what it is.
Published unedited under our right-of-reply guarantee.
How was this story?
We publish the result, whatever it is. Reader verdicts appear on the front page and in our newsroom metrics.
Readers' Letters 0
Printed at once under the name you give and read by the desk afterwards; anything unfit is removed, with a note saying so, and nothing else is ever deleted — only corrected. Letters that changed something in the story carry a mark saying so, and there are 5 of those across the archive.