Magnificient FDM-S3: the Millihertz Magnifier

1340kHz, 25Hz window, resolution bandwidth 0.0061Hz: more than 100 U.S. AM stations are discernable by their frequency offset.Antenna: vertical active dipole MD300DX, 2 x 5m. Visualized with SDRC V3 software by Simon Brown, G4ELI.

With Elad’s FDM-S3 SDR now hitting the market, we have a receiver at hand which is supported by an OCXO/ GNSS frequency reference. This combines short-time accuracy with long-time stability and allows for precise frequency measurement in the millihertz range (under 30MHz). Exploiting this feature is as exciting as it is innovative. With this new tool, also a new kind of DXing is evolving. One example is propagation analysis. See below the 24h spectrogram of Radio Gotel from Jabura/Nigeria on its exclusive channel of 917kHz:

Radio Gotel transmits from 04:00 UTC to 23:00 UTC on 917kHz. In this spectrogram you clearly see sign-on, sign-off; fade-out, fade-in, plus some other feature like two short power breaks in the evening as well as some instabilities.

What surprises, is both, the late fade out at around 07:30UTC and the early fade-in as early as 15:20UTC. It is important to note that you here see the carrier with a resolution bandwidth of 0.0009Hz, roughly just one millihertz. The gain, compared to a listening bandwidth of 10 kHz, is a whopping 70dB, allowing extreme DX. Audio starts to emerge only from around 18:00UTC. As DX Atlas shows, the whole path between my location and Radio Gotel is under daylight at the palpable fade-in at around 15:20UTC, see screenshot below.

At the first visible trace of Radio Gotel at DK8OK’s location on 19JAN2021, with the whole path still is in daylight. Illustration with the help of DX Atlas software.

As with all new things: “We’ve only just begun”, Carpenters, 1970. To be continued.


  • Karsten Niehaus

    Lieber Herr Schiffhauer,

    wie machen Sie das nur immer so am technischen Puls der Zeit zu sein und dann auch noch mit den Carpenters zu schließen? Hut ab!

    mein ganz herzlicher Dank für Ihre Bücher, von denen ich fast alle habe, und ihre „lockere Internetpräsenz“, die wirklich viel Freude bereitet. Ich werde sicher Ihr treuer Leser bleiben und weiter „Radioforschung“ betreiben!

    Mit den besten Wünschen,

    Karsten Niehaus

    Prof. Dr. Karsten Niehaus Proteome and Metabolome Res. – Bio27 Center for Biotechnology – CeBiTec & Faculty of Biology Bielefeld University


  • Chris Smolinski

    I’ve actually written a program, Carrier Sleuth, which can generate high resolution waterfalls from I/Q recording files from a variety of SDR programs:

  • Very interesting article and useful to watch offsets. In practice this can be done with RSP1 models and external clock and also with HF+ Discovery with a bit less accuracy (?). In live DX with maximum zoom in SDRConsole with the proper adjustments in weaterfall resolution and speed

    • Thanks, Zacharias – yes, you can try this technique with each and every receiver/if tuned properly. But if you approach the millihertz range, you must use an OCXO for short-time stability and accuracy, and GNSS for long-time. The live spectrum of SDRC is a welcome First Aid, but millihertz resolution is best achieved offline. A very pure oscillator helps … 73 Nils, DK8OK


  • Francesco Di Giovanni

    Dear Mr. Schiffhauer,
    how dou you obtained this interesting picture?
    and how do you set te recording parameters?

    Best regards
    Francesco Di Giovanni, IN3XZP

  • Hi, Francesco – the screenshots were made on the basis of a recording with SDRC V3 software / 1.5 MHz wide, 24h long, 16bit, WAV RF64. From this file, one channel was analyzed in the File Analyzer, a module of SDRC V3 software. There the bandwidth was set to e.g. 30Hz, the length to e.g. 1500 lines (for 24h), and resolution bandwidth x8. Result is a spectrogram which may be tweaked a bit with the “Color” function of the File Analyzer. Seven years ago, I wrote a short introduction, showing at least the concept, but at an old version of SDRC:
    Hope this helps a bit. 73 Nils, DK8OK

Leave a Reply to Chris Smolinski Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s