1
\$\begingroup\$

schematic

simulate this circuit – Schematic created using CircuitLab

I designed this as it would pass frequencies between 400-600 Hz. But it passes every single frequency anyway. What is the problem?

schematic

graph of Vin vs Vout

\$\endgroup\$
  • 2
    \$\begingroup\$ Where is your in- and where is your output? How do you measure? What is your expected attenuation? What is the measured one? \$\endgroup\$ – Marcus Müller Jan 18 '17 at 21:10
  • \$\begingroup\$ Each stage has an input and output impedance (f). Be mindful in choose RC impedance to prevent loading on previous stage. the source impedance affects the next stage must be much lower ... and visa versa. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jan 18 '17 at 21:16
  • \$\begingroup\$ I edited my entry. \$\endgroup\$ – eleelo Jan 18 '17 at 21:16
  • 5
    \$\begingroup\$ Do an AC analysis and you will see it works fine. Bandpass filter doesn't mean all others are reduced to 0, they are attenuated by a certain amount, and the AC analysis will show you nicely how much. \$\endgroup\$ – PlasmaHH Jan 18 '17 at 21:42
  • 1
    \$\begingroup\$ Filter design is a huge art. You may want to read about "notch filters" there's a style called Twin-T notch filters that can reject a single frequency quite well - you may be able to use one of these for each frequency you want to reject. \$\endgroup\$ – Brian Drummond Jan 18 '17 at 22:08
1
\$\begingroup\$

The problem is you are designing your poles in the wrong place, if you want a pass band, your going to have to space them further apart, or add gain. It looks like you have a passband, but you will see attenuation on all frequencies. When you simulated your using a 10V signal but that is probably because a 1V signal showed much more attenuation so it was increased.

The cutoff frequency is for the -3dB point, not the frequencies where attenuation starts.

from wikipedia

I split the two filters out to illustrate, The lowpass is blue and highpass green. (sorry LT spice wouldn't give me bigger lines) Both filters by themselves never reach more than -2.2dB together they never reach 4.4dB. Even at 1kHz there is attenuation.

I would suggest going back to the drawing board and spacing the filter poles farther apart, make sure you know what your pass band looks like. It may be necessary to increase the roll off with more poles (-40dB with two poles or -60dB with three) to achieve the desired results (and impedance buffers may be necessary).

Or you may just want to add some gain with an op amp at the end of the filter to bring the -4.4dB point up to your desired passband.

Use the .AC command in lt spice to see the frequency results of the simulated filter.

enter image description here

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ I actually changed the frequencies, because spacing the poles further apart did not actually work, so I will work with my new frequencies to guarantee the best results. Thank you very much for great explanations :) \$\endgroup\$ – eleelo Jan 18 '17 at 23:49
0
\$\begingroup\$

Impedance ratios are MOST important and 1st order bandpass filters are not very steep. (6dB per octave) Note the changes in LPF I made Rx10 and C/10

schematic

simulate this circuit – Schematic created using CircuitLab enter image description here

Although you have yet to give specs for BW , Q, gain, dB vs f, this is not optimized but shows two cascaded BPF which could be isolated in parallel with 37dB gain and a Q=10 enter image description here

This trace from 10Hz to 30kHz shows a range of 80dB from +37dB gain in the pass band. Minor Pole shift is easy with the input shunt R to gnd.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ I tried the same things, I need exact filtering for my project, rejecting the signals I don't want completely and letting the ones I want pass. Should I use RLC circuits and if so which one? \$\endgroup\$ – eleelo Jan 18 '17 at 21:50
  • 1
    \$\begingroup\$ Not RLC, but you must define pass band thresholds (f,dB) and stopband thresholds (f,dB) then order and type of filter can then be found., including passband ripple and phase if important. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jan 18 '17 at 22:39
  • \$\begingroup\$ wow thanks so much \$\endgroup\$ – eleelo Jan 18 '17 at 23:47
  • 1
    \$\begingroup\$ this website literally saved my life, I cannot thank you enough \$\endgroup\$ – eleelo Jan 19 '17 at 0:30
  • \$\begingroup\$ that's worth something, although hard to measure. Your thanks is more than most here offer. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jan 19 '17 at 0:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.