19–. Translated from the original Dutch by the author, using what ChatGPT had to offer as useful suggestions to speed things up. With the HTML kept intact, at my request.
Twitter can also lead one to good things – it’s not all arguments and bickering about politics. On 10 April 2023 I expressed as my opinion on Twitter (here translated to English) that radio has no future:
"Radio is really in decline now. The long, Long, medium and short wave are almost completely dead and empty now, FM is still there, but even that is practically outdated. I don’t think DAB+ will become a success. I think everything will become G5, that’s all we need."
(By "G5" I meant 5G, mobile communication of the fifth generation.)
Jan Willem Luiten (whom I hadn’t consciously met on Twitter before) saw things differently. I was needlessly sarcastic about FT8, which I now realize I had already forgotten was being discussed at the time. I thought I had only just discovered FT8 today, 15 April, and I now know that it will become a essential aspect of this article after many more words to come.
I just hope this falls under normal forgetfulness and lack of concentration, and is not an early symptom of something nasty.
Later that Monday afternoon, Jan Willem pointed me, or us, of course, since Twitter is a public medium, to WebSDR, set up by Pieter Tjerk de Boer (PA3FWM). The web address is: websdr.ewi.utwente.nl:8901. And only then dawned on me and the penny dropped.
I have never been a radio amateur myself, never been involved in ‘ham radio’ as they call it in English. However, I did read books about it between 1967 and 1970, which after some persuasion I was allowed to borrow (being too young, 13 to 15 years old) from a library in Arnhem, the address of which I can reconstruct as Hisveltplein 21, JO21xx, more precisely JO21xx46hh. These are Maidenhead Locators.
I cannot send or receive Morse code, although I am visually familiar with its encoding. But that’s no use if you hear it instead of seeing it. I knew about the 80-meter band and SSB (single sideband). But only now, not in 1968, but now that I am 68 years old, thanks to this largely software-based wideband receiver of the University of Twente, which has been made accessible to everyone via the internet, I could really hear what is being discussed on the amateur bands. I learned on Twitter that LSB (lower sideband) is used below 10 MHz, and USB (upper sideband) above it.
So, for example, I heard Germans assiduously exchanging call signs. Sometimes they spoke briefly in English with Americans and a Japanese, otherwise German. From a tweet that has since been deleted learned that a contest was going on that day, to make as many connections as possible.
When I was about 10 years old, I got my first transistor radio, made in Japan – they were ahead in this field at the time. It only had medium wave. I also opened it up, cause I wanted to see what was inside. A ferrite antenna and a tuning capacitor. And little thingies with adjusting screws, which only later I understood were transformers of the tuned intermediate frequency stages of the superheterodyne receiver.
I also had a circuit diagram, which unfortunately has not been preserved. I don’t have the radio itself anymore either. On the schematic, I think the resistors were not shown as rectangles, as was customary, later, or then already in Europe, but as wiggly things with diagonal lines, in American style (I think).
There were eight transistors. I used to stare at that circuit diagram and try to figure out how it all worked, how the transistors amplified, how that very weak antenna signal ultimately became audible sound. I only learned that much later, during my technical vocational training. Books I read told me that detection, the rectification of the radio signal to extract the modulated signal, was done with a diode. But there was no diode in the schematic of this radio. Only those 8 transistors. Presumably one of them was biased in such a way that it could also rectify. There were 4 transistors for the mixer and intermediate frequency amplifier, and 4 for the audio section, 2 for the bridge and 2 as preamplifiers. I only understood that afterwards.
My parents thought my interest in radio technology I might have been inherited from my grandfather. I never met him because he died in 1939 and I was born in 1955. He had built a radio receiver himself. No transmitter, I never heard anything about that. My mother was asked to help with the soldering because she had thin fingers. She held the wires while my grandfather soldered.
A few years ago, I used to let my car radio scan the medium wave and long wave bands. Even in the evenings, when medium wave signals have greater reach, there was very little to hear. That was certainly different in the past. Many stations on those bands have stopped broadcasting because it became too expensive to keep them operational, and there are better alternatives: FM for shorter distances, and the internet worldwide.
AM stands for Amplitude Modulation. This usually refers to the simplest form, double-sideband with non-suppressed carrier. Generally used for broadcast stations in the long, medium, and shortwave bands. Single-sideband modulation is more efficient in terms of bandwidth and transmitter power, especially when using suppressed carrier (SSBSC), but it is technically more complex and therefore back then wasn’t suitable for affordable consumer products.
The web receiver made available by the University of Twente covers the entire frequency range from 0 to 30 MHz, including the broadcast bands LW and MW. One of the few long wave stations that still come through loud and clear is BBC Radio 4 on 198 kHz. It used to be 200 kHz and exactly 1500 meters, and as far as I remember, the program was then BBC Radio 2. The transmitter is located in Droitwich, a name that used to be on the tuning dials of radios, when modern Dutch radio stations NPO 1 and 2 (NPO = Nederlandse Publieke Omroep, Dutch Public Broadcasting Service) were still called Hilversum 1 and 2, named after the location of the transmitters in part of that era.
That evening, April 10, 2023, I noticed a strong radio station on the medium wave at 648 kHz. Six four eight, very familiar to me, because there was a time when, still in bed for a while in the morning, I always followed the world news via the BBC World Service. I mentioned that before here (in Dutch). I believe the Dutch World Service also rented that transmitter for a while. Radio Oranje was not broadcast on that frequency, as I read here. But it was broadcast on 1500 meters. And in fact also by the Orfordness transmitter, because one of the frequencies it uses, or used to use, is 1296 kHz, which corresponds to 231 meters, one of the wavelengths mentioned by in the Dutch Wikipedia article on Radio Oranje, which was a daily 15-minutes News bulletin in Dutch, broadcast from England to German-occupied wartime Holland.
Strangely enough, what can be heard nowadays on 648 kHz is Radio Caroline, once a pirate station broadcasting from the ship Mi Amigo off the English coast. Apparently it now has a legal status. They use the Orfordness station, i.e. probably the antenna, but with a different, far less powerful transmitter. The English-language Wikipedia is quite detailed about Radio Caroline: 1, 2.
Mi Amigo, Radio Veronica, Radio Noordzee, there are all kinds of complicated historical links between them. I won’t go into that any further.
Again using the Twente WebSDR (SDR = Software-Defined Radio) on the evening of 14 April 2023, I heard a strong signal coming through (and I saw it too, in the waterfall display of the radio spectrum), which was in Arabic. The web interface provides labels from a database for the broadcasting bands. One of them at the corresponding frequency, 1467 kHz, is “ARS SBC Radio Riyadh”. Riyadh (الرياض), that is the capital of Saudi Arabia. Hence the Arabic. Could that broadcast really come from that far? With that signal strength? That can’t be true, right? The furthest transmitters that I heard on medium wave in the past (only at night, of course, because then a reflection layer is active, which is often also present for shortwave bands at other times) were Vatican Radio (from a transmitter near Rome, about 1250 km straight-line distance from the Netherlands), and maybe (but my memories are vague) Radio Tirana, at 1630 km.
But Riyadh is over 4600 km away. That can’t be bridged on medium wave unless it’s with a monstrous amount of power, but even then it wouldn’t work. The mystery was quickly solved: one of the transmitter using that frequency, 1467 kHz, is in Roumoules, I read here. It mentions TWR, which was also on one of the WebSDR labels. TWR = Trans World Radio, a Christian station. And Roumoules, that certainly sounds familiar! Yes, that is the transmitting station in southern France that used to belong to Radio Monte Carlo. I have often heard that in the Netherlands in the evening, at 205 meters, close to Radio Luxembourg at 208 meters. Stations were identified in meters back then, not kilohertz. 205 meters is 1462.4 kHz, not equal to 1467, but close enough. International frequency conventions were different in those days.
So Trans World Radio rents that transmitter, to bring their Christian message in Arabic to North Africa and the Middle East. But it also reaches further north. That confirmed what I thought I had heard in Arabic, a language I don’t understand, although I sometimes recognise a few words. Corinthians, something like that. That’s a Christian concept, right? But Saudi Arabia is a strictly Muslim country, they’re not going to talk about the Epistles of Paul on the radio there, are they? Indeed, but then that broadcast didn’t come from Saudi Arabia.
Meanwhile, as I write this on 19 April 2023, around 1 o’clock in the afternoon, on that same frequency 1467 kHz I now hear a station called Radio Twente Gold, with non-stop music that I associate with the years just after World War II, even though I didn’t experience that time myself. A kind of cheerful dixieland-like jazz, but not the original, by black people, but a white imitation. Sounds pleasant and optimistic to me. The Netherlands were recovering from war time damage. The Dakota DC3. But maybe that music is even older. 1920s, 1930s? The station has a website.
Something completely different, also about leased transmitters: years ago, for a time, a broadcast in Chinese could be heard from the Chinese government on the former transmitter of Radio Luxembourg on 208 meters.
A former transmission tower for Dutch Radio 3, previously also used by Radio 1 and 2 (or was that only as a backup transmitter?), not far from my hometown, broadcast a Catholic program from Radio Maria on 675 kHz for a couple of years. That agreement was eventually terminated, and the transmitter tower was blown up. But the concrete foundations of the guy wires are still visible in the landscape.
On the evening of 14 April and morning of 15 April 2023, I heard peculiar sounding signals in the 80 and 40 meter bands. It sounded like a bunch of audio frequencies, kind of whiny and out of tune, with breaks of a few seconds in between. A choir of organ tones, which did not always start exactly at the same time. Here is an example.
I immediately suspected that it was a form of digital data encoding, and after some searching on Wikipedia, I thought it might be Olivia. But that sounds rather different.
My informant on Twitter thought it might be FT8. That could be right. The individual tones are not a single signal, but each come from a different transmitter. The protocol works with 4 blocks of 15 seconds per minute. The last 2.36 seconds of each block is a pause. A synchronous computer clock is required for synchronised transmission. Because the clock of some transmitters does not run perfectly, you sometimes hear the uneven start of the ‘organ voices’.
Each ‘voice’ has 8 frequency variants within 50 Hz, 6.25 Hz apart. Switching the frequency encodes raw data bits. That’s how FSK, Frequency Shift Keying works. To the unprepared musical listener, it creates the impression of a messy vibrato.
See also the Signal Identification Guide. FT8 was designed by Joe Taylor (K1JT) and Steve Franke (K9AN). Joe Taylor wrote the program WSJT-X, now maintained by him and others as FOSS (Free and Open Source Software). This program can decode multiple FT8 signals within a certain bandwidth simultaneously.
I found it hard to figure out how exactly this FT8. The deeper parts of the software were written in Fortran 90, which doesn’t really increase accessibility for me. This description does shed some light. There is also an FT8 library in C instead of Fortran, written by Karlis Goba (YL3JG).
FT8 apparently compresses common elements from a dialogue, adds a CRC (Cyclic Redundancy Check), and applies FEC (Forward Error Correction) in the form of LDPC (Low-density parity-check code). Also, 7x7 Costas arrays are applied at the beginning, middle, and end for synchronisation, as explained here.
LDPC, devised by Robert G. Gallager in 1960, is similar in purpose to Turbo codes. They are error detecting and error correcting codes. Walsh matrices (1923) and Hadamard matrices play a role in it. Bits from the coding are spread out to reduce the impact of interference.
By applying these techniques, FT8, like Olivia, also works well in the presence of fading, noise, and interference, and in case of poor propagation, low power operation, and inefficient antennas in restricted urban spaces. Radio amateurs can communicate over much greater distances than with voice or older telex-like digital transmission modes.
See also these FT8 and FT4 specs by Steve Franke (K9AN), Bill Somerville (G4WJS) and Joe Taylor (K1JT).