Hi Ap,

As long as you know that you are only interested in four well-defined frequencies, a set of four active filters would work well, at least so far as resolving the signal. Active filters are cheap to build. Of course, that still leaves you with the task of detecting those four discrete signals (the output from the four active filters).

Do you need to do this in real time? How long do you have to detect and do something with the signals? Another option you might be able to use is the FFT which was recently discussed in
this thread.

The advantage of using an FFT is that you aren't committed to a particular set of frequencies, but you can measure whatever frequencies are there instead. There is also the advantage that it won't require the additional hardware.

The drawback to using the FFT will be resolving the commencement and abatement of each frequency in succession. What you would have to actually do would be a series short FFTs, and you will have to evaluate your time constraints to see if it's even doable.

Good luck,

Joe