Hello everyone

I'm looking to implement an FSK demodulator, I did some research regarding FSK modulation and demodulation and I understood that essentially the concept of working is quite simple.
Where a digital 1 translates to a high frequency, and a 0 translates to a low frequency.
Taking a look at the available demodulation circuits, they all seem quite complex and not really explained furthermore, not many variations of the demodulator exist...

What I wanted to ask was whether a simpler approach would work, using a filter followed by a comparator...
since we know the frequencies which represent a digital 0 and 1, we can expect the filter's voltage output (response) to each frequency. Now we compare these voltages to a reference.. if it's higher than the reference we output a 1, else if it's lower we output a 0.

The issue which comes to my mind is... in the comparator what would be the reference? Furthermore, our input is an AC signal and in a comparator we would need to compare it to an AC signal as well right?

That's all for now
Thank you

 

The problem with a filter is its slow response time to the change in frequency.
That will significantly limit the data transmission rate below the maximum that can be obtained with other demodulation schemes.
Is demodulation/signal speed a concern for your application?

 

http://www.ti.com/lit/an/slaa618/slaa618.pdf

http://www.ti.com/lit/an/snoa651/snoa651.pdf

Above might help.

Regards, Dana.

 

http://www.ti.com/lit/an/slaa618/slaa618.pdf

http://www.ti.com/lit/an/snoa651/snoa651.pdf

Above might help.

Regards, Dana.

Thank you

 

The problem with a filter is its slow response time to the change in frequency.
That will significantly limit the data transmission rate below the maximum that can be obtained with other demodulation schemes.
Is demodulation/signal speed a concern for your application?

I can't really be certain at this point, regarding the speed. I'll check.
What would be some other possible schemes of demodulation? and isn't demodulation using an algorithm (software) through a microcontroller easier, given that the algorithm already exists.

 

isn't demodulation using an algorithm (software) through a microcontroller easier, given that the algorithm already exists.

Likely true, if the micro is fast enough to do the demodulation.
What is the FSK data rate?

 

A possibility -

http://www.cypress.com/documentatio...and-psoc-5lp-low-frequency-fsk-modulation-and

This is Xmit side, sine generation, Recv side (correlator) in separate project workspace, also in ap note.
See attached. Single chip solution, both recv/xmit.



Regards, Dana.

 

Back "in the day", my first micro had a Kansas City Demodulator for the cassette interface.
It used a number of gates and timers and stuff.
https://en.wikipedia.org/wiki/Kansas_City_standard
http://www.swtpc.com/mholley/AC30/AC30_AssemblyInstructions.pdf

Here is another one...
https://hackaday.io/project/25877/logs
It may help if you want to go the lots of bits way