active filtering circuit design

Your specifications incomplete -



Also what latency can you tolerate, delay from a step input in frequency to
a filtered output ? Additionally accuracy of filter response. Do you want
linear phase response or is that unimportant ? Any noise considerations ?
Lastly can you comment on what are the filter outputs being used for ?


Regards, Dana.

Active filter implies an op amp, as well as capacitors and resistors. (Usually 2 caps and 2-3 resistors.)

You specify several passbands, many of which are narrow. I suppose your project has an op amp for each passband? Look up articles with circuits that contain an op amp, and narrow passbands, adjusted for high Q.

An alternative to analog is use simple reciprocal frequency counter techniques and software.
Effectively you can get very brick wall kinds of response to detecting a signal is in passband.
Like tone detection. Primary downsides is you do not have a processed signal out the backend,
just a decision freq is in passband. And latency to decision. All done with very high accuracy and
no isues drift, tuning, etc.. So the questions I posed to you in #2 will answer if this is an alternative.

Comb filters another approach.


Regards, Dana.

danadakk said:

Your specifications incomplete -

View attachment 170385

Also what latency can you tolerate, delay from a step input in frequency to
a filtered output ? Additionally accuracy of filter response. Do you want
linear phase response or is that unimportant ? Any noise considerations ?
Lastly can you comment on what are the filter outputs being used for ?


Regards, Dana.

Click to expand...

Thank you for your answer. It is for homework. Latency and accuracy are not really important. Only important things are the defined output frequencies. How can I make a circuit to get these outputs, can you help me?

BradtheRad said:

Active filter implies an op amp, as well as capacitors and resistors. (Usually 2 caps and 2-3 resistors.)

You specify several passbands, many of which are narrow. I suppose your project has an op amp for each passband? Look up articles with circuits that contain an op amp, and narrow passbands, adjusted for high Q.

Click to expand...

Thank you for your answer. How can I use choose which opamp should I use and how can I define the values I need for capacitors and resistors? Also to create the circuit on breadboard, what should I use to give signal and measure output signal. I don't have lab environment. Can I use phone or something else?

circuitt said:

Thank you for your answer. It is for homework. Latency and accuracy are not really important. Only important things are the defined output frequencies. How can I make a circuit to get these outputs, can you help me?

--- Updated Jun 25, 2021 ---

Thank you for your answer. How can I use choose which opamp should I use and how can I define the values I need for capacitors and resistors? Also to create the circuit on breadboard, what should I use to give signal and measure output signal. I don't have lab environment. Can I use phone or something else?

Click to expand...

Start by goggling band pass active filter, there are a number of topologies using OpAmps,
R & C, to create BPFs. Along with tools/wizards to set the parameters and the wizard will
generate R & C values.

One such tool - https://www.ti.com/design-resources/design-tools-simulation/filter-designer.html

Another https://tools.analog.com/en/filterwizard/

Regards, Dana.

circuitt said:

how can I define the values I need for capacitors and resistors? Also to create the circuit on breadboard, what should I use to give signal and measure output signal.

Click to expand...

Try configurations at the website below. There is the Wideband type, and the narrowband type. They have a slightly differing placement of a capacitor. The narrowband is a multiple feedback active filter. It has high selectivity (high Q), which makes it suitable for your specs. Its passband can be moved up or down by adjusting a resistor value.


By running simulations you get an idea what capacitor values work best for your range of frequencies.

To generate an input signal, a sinewave oscillator is what you need. Your specs are for tones below the audio range, therefore consider building it in hardware. It's doubtful a computer (or smartphone) can produce such low tones.

A frequency counter and oscilloscope are practically a requirement for this task if you're building in hardware. As an alternative you can construct a 1 or 2 digit counter from a few IC's. You can make an led bargraph to indicate amplitude.

PC Oscilloscope and Function generator -


If you do this make sure you protect your sound card inputs and outputs



To protect sound card output a unity gain OpAmp buffer with s simple
series R of say 1K ohm or more to OpAmp input should do.

Note this has a spectrum analyzer with peak hold in it, so you can generate white
noise with its generator, feed to filter, and using spectrum analyzer with peak hold
see its response curve.


Regards, Dana.

I'm searching for studies about band pass filtering, but couldn't find an application including multiple separate filtered outputs as in my homework. How can I apply it in a circuit, should I connect opamp circuits sequentially to each other's outputs?

You'll use a separate LPF/HPF combination for each band. You didn't yet specify the required filter steepness. I would expect 12 or 18 dB/octave, corresponding to 2nd or 3rd order filters.

Each filter you need is essentially a stand alone unit, with an input and its
own output, so you would not connect them in series unless you make
cascaded sections for a specific frequency and want steep slopes which
implies higher order filters. Normally cascades are done with second order
sections, and in this design case you do connect in series. Note the tool
generated this solution, but its not practicable due to order way to high.

Here is an example for your 100 - 1000 Hz filter with steep slopes -




If the order of this filter was only 2 or 3 then a cascade would not be necessary, but you would not
get the steep slopes as shown above. Also the order of the above filter way to high for analog
solution, as cumulative errors of passive values and drift. "Normally" 6 - 8 second order sections
limit of cascaded opamp filters.

Digital filters can do very high order, here is one with much steeper cutoff characteristics -



Another approach is comb filter design which are simple but images have to be removed.



Regards, Dana.

Band filters with a corner frequency ratio of two and more are better implementes as HP/LP combination.