Maker Pro
Maker Pro

How low does a notch filter go

chagz

Jan 4, 2010
7
Joined
Jan 4, 2010
Messages
7
I am in the process of testing out a wien bridge notch filter that i designed. The circuit simulation promised me a gain of -100dB at my chosen frequency. Which means that is almost zero output considering that I am to use about 2V at my input.
The problem is that I have never before, until now, seen a practical active filter circuit. For an input of 2V, I have an output of a noisy 20mV signal, which is absolutely OK for my purposes. I am just wondering, having taken all noise aspects into consideration,common grounding etc, how low an output can one get from a practical filter.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
The circuit simulation promised you an attenuation of 100dB (do you mean dBV?),

Q: What attenuation are you seeing?

Q: What could alter the depth of the notch?

Hint - there are several factors. (and don't say Q!)

Can any of them be made "perfect"?
 

chagz

Jan 4, 2010
7
Joined
Jan 4, 2010
Messages
7
Is not an attenuation of 100dB the same as a gain of -100dB. For an input of 2V, a gain of -100dB would produce an output of 10uV. Instead, I get a 20mV noisy signal at my output. That is a gain of -40dB.
The depth of the notch can be increased by closely matching resistors R1,R2 and R5 in the attached file. For maximum attenuation R1,R2 and R5 are made PERFECTLY equal. In my practical circuit I tred as best as I can to make them equal(at 22.1kohm) using a variable 4kohm resistor in series with a fixed 18kohm resistor.

I am completely satisfied with my gain of -40dB. My main concern is whether I could have done this better, whether it is possible for better results to be achieved(how better?).
 

Attachments

  • notch madness babe!!!.gif
    notch madness babe!!!.gif
    5.1 KB · Views: 379

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
OK, clearly you have identified that component tolerance (and matching) is one thing that affects the actual vs. theoretical performance of a filter.

It is almost certain that for maximum performance, the filter would require an op-amp with infinite gain and input impedance, and zero output impedance.

Compare the filter design you've chosen with a twin T filter.

This, and this may be of interest.

Please note. Filters are not "my thing", and aside from a few basic rules of thumb, I know very little :)
 
Top