Also known as: AFSK
AFSK (audio frequency-shift keying) represents bits as audio tones that then modulate a radio (usually FM).1 The classic case is the Bell 202 standard — 1200 Hz and 2200 Hz tones — carrying 1200 bps APRS packet over AX.25. Because the modulation lives entirely in the audio band, AFSK turns any ordinary voice transceiver into a data radio without touching its RF section.
How it works
Instead of shifting the RF carrier between frequencies the way plain FSK does, AFSK shifts an audio-frequency tone between a mark and a space pitch and feeds that audio into an ordinary FM (or SSB) transmitter’s microphone input. The RF end is oblivious — it just transmits whatever audio it is given — so a stock handheld or mobile becomes a modem. In the Bell 202 scheme a mark is 1200 Hz and a space is 2200 Hz, sent at 1200 baud. The bits are usually NRZI-encoded (a change of tone denotes a zero, no change a one), which keeps the receiver’s clock recovery fed with transitions during long runs of the same bit and makes the link insensitive to which tone is “high.”
A receiver FM-demodulates the RF to recover the audio, then decides which tone is present in each bit period — historically with a pair of bandpass filters and an envelope comparator, or a phase-locked loop, and in software with a correlation or Goertzel detector at the two tone frequencies. The recovered bits are framed as AX.25 packets by a terminal node controller (TNC) or its software equivalent.
Variants
The Bell 202 1200 bps form dominates VHF packet radio and APRS, but AFSK spans a family. Bell 103 (300 baud, used on HF) shifts by only ±100 Hz. FFSK is the coherent cousin, where the tone frequencies are exact integer multiples of the bit rate so each bit holds a whole number of cycles — cleaner to detect, used for MDC1200 and DSC. Higher-rate packet abandons audio tones for direct GFSK at 9600 baud (the G3RUH modem), which keys the transmitter’s modulator directly rather than through the audio stage.
Relevance to SDR
AFSK is a natural fit for software decoding: an SDR FM-demodulates the channel, then a tone detector recovers mark/space and the AX.25 framer extracts packets — exactly what Dire Wolf does as a soundcard TNC. GopherTrunk decodes AFSK as part of its APRS pipeline, detecting the Bell 202 mark/space tones after FM demodulation and passing the framed AX.25 to its position/telemetry parser.
Sources
-
Frequency-shift keying — Wikipedia, for audio FSK and the mark/space tone concept. ↩