The signal consists of two separate tones, one at 2125 and 2295 Hz (designated "Mark" and "Space"). Notice that 2295 Hz - 2125 Hz= 170 Hz, which is where the designation "170 Hz" shift arises--it's the distance between the two tones. In a true FSK signal, the transmitter will produce a dead carrier on one of the frequencies, and the RTTY machine (or TNC) will then shift the output frequency of the carrier 170 Hz, thus producing the other tone. In an AFSK signal, the audio input of a transmitter is fed the two audio tones (in usually LSB mode) which then produces modulation.

In all respects, the AFSK signal and FSK are identical. But it is much easier to transmit a carrier on a frequency and then shift it a specified amount than it is to produce two audio tones independent of each other exactly on two frequencies. Factor in that most AFSK setups use voice inputs with their own characteristic filtering and frequency response and you can see why sometimes we run into trouble. Plus, take a look at the far left..see the harmonic? These are scattered out all over the audio spectrum when running AFSK with the 1278.

Ok, so at this point we know what frequency the tones are supposed to be..but how do we set them, and how do we know that where we set them is correct??