Filter for ultra low frequency

[Tinkerer]

I need to separate 2 signals.
Signal # 1 is 1 to 5 Hz
Signal # 2 is 2 to 15 Hz

signal # 2 is always 2 to 4 times the frequency of signal # 1

Any help is apreciated

Tinkerer

 

[Nigel Goodwin]

You need to be FAR more specific, that question's much too vague!.

 

[audioguru]

You are here too?
Both signals have frequencies between 2Hz and 5Hz.
The frequencies overlap so cannot be separated with filters.

Now you say the frequencies do not overlap.
If you use a 4th order Butterworth lowpass filter for -3dB down at 5Hz, and a 4th order Butterworth highpass filter for -3dB at 10Hz, then the lowpass output at 10Hz is -24dB and the highpass output at 5Hz is also -24dB. Only 24dB is not much separation.
All harmonics of the low frequencies will appear in the output of the highpass filter.

 

[Tinkerer]

The frequencies overlap, but the one is always 2 or 3 times higher than the other one. How about looking it from a different perspective: Frequency # 2 is 2 times higher that Frequency # 1. What if we have two bandpass filters?

 

[audioguru]

I just explained on the other site that a very complex and sharp lowpass filter probably isn't sharp enough because the frequencies are too close together.
Since the frequencies overlap then the filter won't work. It will filter out the wanted signal some of the time.

Over there you said the frequencies are half a sine-wave. Then it is full of 2nd harmonic that will not be filtered out.

 

[Hero999]

How about posting a link to the other site so other people here can see exactly what happened.

 

[audioguru]

It doesn't make sense. Noises from whales or vibrations from a train rumbling past.
This OP has a different login name than the other guy so maybe they are two school kids who are in the same class.

 

[philba]

heh heh, how about a DSP with an FFT? not exactly a couple of op amps...

 

[Analog]

Ultra low frequency to me is in the milli-hertz range. We use a Novocontrol unit that measures dielectric properties in the milli-hertz range that can take days.

 

[Sceadwian]

Where are these signals coming from? How fast does the frequncy change within the modulation range? Are they always perfectly in phase? You can't really filter frequencies of this range, not using typical filter topologies, a better understanding of the exact voltage and time domain you're talking about is really required to answer the question.