Are you sure you need an ADC? back in the old days, we did this with simple analog (RC) filtering and an input port. One simply waits for a transition and times the high frequency half-cycle +10% and then concludes, if there wasn't a transition, that it's the lower frequency, but waits for and times the next transition anyway, then restarts the timer. The whole thing fits in a single software loop, but it's more efficient if you have a timer available. Keep in mind, too, that you could differentiate the incoming waveform with LSTTL, with a series cap and a pulldown resistor. I used to use about ~220 pf and a <1K pulldown. That generates a positive pulse of somewhat under 100 ns on a rising waveform and, a negative one for a falling waveform if a pullup is provided. Take your pick, you only need one of them, i.e. only rising OR falling.


We drove our cassette drives, the popular storage medium of the time, with an output port bit, filtered (rounded) somewhat, and the receiver generated a pulse on the rising edge of the waveform. 4800 bps was what we used, though it was quite possible to get to 8k bps, though tape and drive became a factor by that point. Your code has to be in a precisely timed loop, of course, so you end up with a byte and can put it away before the next edge. If you use a timer to measure the interval between edges, you can probably use interrupts to handle other processes in your system, so long as the ISR's are short.

RE