Hello,
I do not have one of the specified chips, but another TI chip in combination with a third party chip. I am here because I hope that someone here can help me understand the concepts a little better.
The third party chip receives a Bell 103 BFSK signal (110 / 300 bps), and forwards the samples on a I2S / PCM bus. The sampling rate is 8 kHz.
What is the best way to determine 1's and 0's from this PCM signal?
I have looked at the Goertzel algorithm. This seems very promising, but I am starting to doubt that now.
With 8 kHz sampling, so 1 sample per 0.125 ms, this would come down to around 26 samples per 3.33 ms (the time per bit in the 300 bps FSK signal).
If I understood it correctly, the Goertzel algorithm can only achieve a frequency bin width / resolution of 8000 / 26 = ~300 Hz.
This would not be sufficient, as the BFSK mark and space frequencies used in this system are 1070 and 1270 Hz.
I was hoping perhaps someone here could give me hints to tackle this.
