Hello, i have an issue designing a band pass filter that allows frequencies between 300-3400Hz to bypass with a total gain of 1.
My calculations using the Sallen-Key model are correct for the resistances and capacitors, but when using a FFT in LTSpice I do not get my desired result.
Any help on what is wrong would be greatly appreciated!
Attached there is the circuit and the FFT for the exit signal, above R12.

 

Your 15V supply is floating. It has no reference to ground.
Re-do it as two 7.5V voltage sources, each with one terminal connected to ground.
I don’t know where V(n014) is, but if it’s anywhere near Q1, you won’t get a good output as Q1 base has no DC bias. What is Q1 supposed to do?
And finally, you have the op-amp inputs reversed on U2, and V1 output is shorted to ground.

 

If you want to use a single 15V supply, you will need to ground its negative terminal, and generate a pseudo-ground voltage at 1/2 the supply voltage to bias the op amp inputs.
This can be done with two equal value resistors, and an op amp buffer circuit (example below):
You can use the 741 as the op amp with a 15V supply.


Label the nodes you are plotting with the Label Net command, so we know what is being plotted.
V(n014) tells us nothing about where that is.

A nit, but AC analysis is not an FFT, which is used in the Transient Analysis.
It gives a Bode plot of the output, using linear models for the devices.
For an AC source value of 1V (as you have), a gain of 1 will give a 0dB output.

 

Just expanding on what @crutschow said about the FFT.
If you take the FFT of the transient analysis, the result will depend on your input signal.
If your input signal is 5kHz, then your FFT will show a big peak at 5kHz, and small peaks at 10kHz and 15kHz corresponding to the distortion that your circuit introduces. What it won't do is give you the frequency response of the circuit.
I suppose, that in theory, if you use a Dirac delta-function as the input signal then the FFT of the output would be the frequency response of the circuit but I don't know if SPICE can do that!

 

A small addition: for the bipolar transistor to work correctly, you need to set the DC offset for the base of the transistor. It is necessary to use at least one resistor for this!

 

Thank you all for the replies, I will try and modify it according to your specifications.

 

Thank you all for the replies, I will try and modify it according to your specifications.

Good luck - but what is the reason for coupling capacitors (C5, C6) having only 1nF (lower frequency edge of the filter at 300 Hz !)

 

A suggestion;
Always label your nodes, specifically the ones you intend to probe.
Many times in this forum we are shown plots that look sketchy, only to learn that the TS placed the probe on a node different from what people expected.

 

Out of curiosity, are you trying to simulate an old fashioned analogue telephone service frequency response?