Re: Developing an FSK receiver step-by-step

[papadeltasierra]

RFTM, thanks for the great article - very helpful. Here's something that I believe I can add. At one point, after the quadrature dmod block, you have to add a constant to bring the data signal down to 'centred on zero' - I think this is an error. I believe that this indicates that your frequency offset (-950kHz) is wrong.
If I understand quadrature dmod correctly, the output is 'how far am I from the centre frequency' which for 2-FSK should be between -X and +X i.e always centred. If you are seeing a signal that is not 'symmetric' then your assumed centre (offset) frequency is wrong.
So instead of 'hacking' the output by adding a constant, tweak the offset frequency. What seems to work for me is that if the 2-FSK offset is +/-X, then the centre frequency is a multiple of 2X.

[rumpeltux]

+1 great article, helped me a lot to understand various issues in demodulating FSK.

The adding a constant part is weird, esp. since it involves trial & error.
Instead, a simple high-pass filter to filter out anything below e.g. 1kHz should do fine to remove the DC component from your data.

Also it seems that the article ends abruptly? (section: Communicating with the receiver in real-time)