Active narrow band-pass filter design

[Teik]

I'm trying to limit audible bandwidth outside of 5kHz to 9kHz range. Just to get started I designed a passive band-pass filter successfully. Because I would like to get more aggressive attenuation and use filter with audio amplifier I found that multiple feedback narrow band-pass filter design can be used. I used online tool to calculate values for necessary components I built this circuit by using LM318N op-amp. I used Analog Discovery 2 to generate sine wave from 0Hz to several hundred kHz and used scope/spectrum tool to observer output. Problem is that this circuit did not manage to attenuate output signal. I noticed that I get almost the same result even when op-am is not powered. Might be it is because of those feedback components? If that is the case how can this design even work? I'm very much a beginner and guidance would be much appreciated!

 

[duke37]

The circuit is the same as shown in Electronic Filter Design Handbook by Arthur B. Williams so should work. I use LTspice to simulate circuits.
The LM318N is a high speed op-amp and needs a positive and negative supply which is by-passed. People hate the slow speed 741 but you could try one.
I would think the power supply is the problem.

 

[Audioguru]

Most opamps and your signal source cannot drive those very low resistor values.
A multiple feedback bandpass filter has very gradual cutoff slopes (6dB per octave) about the same as the bass and treble controls on your stereo.

My simulation shows a peak at 7kHz and the response is -3dB at 5.2kHz and 9.6kHz. The response is -12dB at 2.5kHz and 20kHz and -12dB is plenty of output.

If you want 5kHz to 9kHz then you should use a lowpass filter and a highpass filter for a flat response from 5kHz to 9kHz. The filters should have enough orders (maybe 3rd or 4th orders?) to have sharp cutoffs.

 

[AnalogKid]

Agree with AG, separate highpass and lowpass circuits in series will get you much better performance. However, going beyond two poles for each one will require some very precise components, particularly capacitors, and these are expensive. As a next step, look into two-pole highpass and lowpass circuits, such as Sallen-Key and equal-value configurations. Here is an equal-value circuit. If you set the gain to 2, it is very simple in terms of component values.

You can prototype these with standard parts, but you won't get the calculated performance without 1% tolerance resistors and capacitors. Xicon makes 1% mylar or polyester film capacitors at reasonable prices.

ak

 

[Teik]

I added symmetrical power supply mentioned by @duke37 and now it works as it "should". Yes, you guys are right the output is not really what I'm looking for. I'll take those suggestions into consideration. Thank you!