Monitor the ionosphere with an AM radio and tape recorder

D-layer ionization is a daytime phenomenon around the world. An easy way to measure the absorption in a qualitative fashion is by monitoring skywave AM MW signals. Assuming one has a clear-channel station distant enough such that is inaudible or weak by day, and very loud at night, the time-based changes can be explored using the AGC voltage of a typical AM receiver (vacuum tube or solid state). An analog-to-digital converter, or even a hand-built comparator bank could be enough. But we prefer a very simple system to start, where a cassette tape can serve as a basic recording medium, scalable up to computer recording.

Such systems could be mailed to solar eclipse areas, such as a massive eclipse that happened May 1994 across the USA. The totality of a solar eclipse lasts less than 10 minutes. The time constant of the D-region is estimated on the order of 1 second.

D-layer absorption remote sensing system

Using a simple AM radio (even vacuum tube All-American Five) or transistorized with this circuit and a tape recorder, I log D-layer absorption. The key principle is monitoring the AGC (called AVC in vacuum-tube days) inherent to end-user AM radios. We do need a radio old enough to not have an all-in-one AM receiver IC.

Medium- to high-quality AM radios use a tuned circuit and one or more IF stages. At each IF stage, fixed-tune filters reject unwanted signals on adjacent channels (10 kHz spacing in USA, 9 kHz abroad). For multi-IF radios, a cascade of filters and amplification at once increases sensitivity and selectivity.

In words, the signal flow in an medium-quality AM radio goes

  1. antenna (ferrite core with Litz wire wound ‘round)
  2. tracking filter (comprised of tapped coil on ferrite core and multi-gang tuning capacitor)
  3. Mixer (LO tuned by same multi-gang capacitor and separate coil)
  4. Transformer & 455 kHz bandpass filter
  5. IF amp(s)
  6. detector
  7. audio amp
  8. speaker

There is a feedback path from (6) to (5) in the form of an AGC voltage. Assuming (5) uses an NPN transistor, then the AGC voltage will go toward zero as the tuned RF signal gets stronger. This voltage can be monitored via a buffer op amp and then into a voltage-to-frequency converter. Such chips are available off the shelf.

This circuit will monitor signal strength within about a 10 kHz bandwidth centered on the tuned frequency. This bandwidth is the result of the cascaded RF & IF filtering before the detector, and is chosen to be just enough to pass the desired channel while rejecting all other channels strongly.

The added circuit for this project starts at the AGC voltage tap:

  1. buffer op amp (isolates, shifts, flips AGC voltage)
  2. frequency-to-voltage IC
  3. buffer op amp (amplifies, isolates)
  4. tape recorder (or computer storage via sound card)

Elapsed time from start of tape can be recorded in the other stereo channel for an absolute reference from CHU on 40 meters or WWV.

Re: sound card, the storage can be heavily compressed, say via 4-bit ADPCM since amplitude information is unimportant. This yields a data rate of 14.4 MB/hour assuming 8 kS/s sample rate. This is problematic as a 100 MB hard drive would be filled in less than a night.

Tone is still desirable as just a simple cassette recorder, possibly modified for slower speed would be easy and simple to mail.

output spectrum

The voltage-to-frequency converter output is scaled to 500 Hz to 3.5 kHz, to fit within an 8 kS/s sample rate while allowing for cassette machines of varying fidelity.


  • human presses record on a C-90 tape, yielding 45 minutes per side unmodified. This would occur during known transition times (dusk/dawn).
  • Computer records to disk, for transfer to ZIP drive or in the near future CD-R. CD-R would allow several days to be recorded and mailed.

Perhaps we could provide a reduction program that would reduce the by definition single tone down to a single-precision or half-precision floating point number. Given typical AGC time constants, 50 ms sampling period is probably adequate. This yields 7 MB/day, a much more tractable number. A 14.4kb/s modem can handle up to 6.5 MB/hour, so assuming the data is 50% compressible, a user can transfer a day’s data in half-an-hour.

Of course, much of the data will not be interesting, so we can discard on-site at least another half of the data. If we assume a human reviews before sending, or a simple program is employed perhaps considering time derivatives, further reduction, even dramatic reduction occurs in data bandwidth.

Physical package

The circuitry is less than Walkman size including batteries. The largest part of the system is the typical AM radio and cassette recorder. The circuit can tolerate a wide range of voltages, and it should be straightforward for ham radio operators to modify an old radio and set a potentiometer to get the frequency within cassette tape and circuit dynamic range. So we can simply mail the circuit board or let each person construct their own, since many people have an old AM radio and cassette recorder.