I'm going to design some bandpass filters that will select out harmonic frequencies. I have filter design down except for one part- the initial parameters. How do I go about choosing the two bandwidth frequencies and the stop frequencies? I mean, if the signal is a 300Hz signal and I select out 600 Hz does it matter as long as they're far enough away? So I could just choose -50dB 450Hz and 750 Hz or something for the stop frequencies? How should I choose the frequencies and the level of attenuation? How do professionals go about this? The circuit will be experimental, because if you couldn't already tell I'm still learning so I guess I could experiment but I'd rather not you know, reinvent the wheel or anything. Any help is really appreciated!

 

The center frequency is determined by the harmonics that you want to see. The bandwidth and needed attenuation of the filter is determined by how much you need to suppress the harmonics and/or fundamental frequency from the frequency of interest. The bandpass is determined by how much the frequency can vary from the nominal. The order of the filter determines the steepness of the rolloff at the band edges to get the desired attenuation of the other harmonics.

So you need to determine the frequency of interest, the distance between harmonics, the variation of the signal frequency, and how much you need to attenuate the signals outside the frequency of interest.

 

The last part is what I need to figure out. How much suppression is good enough? I want to effectively have no other frequencies so what voltage is acceptable for other frequencies? My example is 347Hz taking the 2nd harmonic (694Hz) I originally chose 50dB by 347Hz which was a voltage of 0.007V, which would be great. But that would need a 20 element high pass filter which is crazy. So I'll redesign- what's really the highest voltage I can get away with for the 347Hz and the next harmonic (1041Hz)? Assume that I just want to see the sine wave on my oscilloscope by itself, as in good enough to pass for it's own signal. Myself, I'll try a few different things since I am experimenting after all, but time is money, and so is you know... money. I'd like to save it where I can.

 

The voltage of the signal is 2.21VRMS by the way

 

Ah I miscalculated before, it was a 10 element filter for 50dB of attentuation.. Anyway I'll solve the problem by myself I suppose I think I can handle it

 

If you want to get rid of ALL harmonics, you could just use a low-pass filter with the cutoff frequency above your desired fundamental tone and below the second harmonic.

How are you planning to implement the filter? Analog passive, analog active (e.g. op amps), digital?

Is the desired fundamental frequency fixed or will the filter need to re-tune on the fly?

 

The last part is what I need to figure out. How much suppression is good enough?

Only you can answer that, it depends on what you're doing with the filter output.