The spikey waveform happens when square waves go through a series capacitor. It performs the math function known as differentiation. I think your input is the sine wave? So how can it get converted to a square wave? Running it through an op amp with extremely high gain.
You have a lowpass filter at the input then feed almost no high frequencies to the highpass filter.
The highpass filter should be at the input and its output should feed the lowpass filter.
EDIT: Your highpass filter uses extremely high value 47uF capacitors causing its cutoff frequency to be extremely low at about 0.03Hz!
I think you want Butterworth highpass and lowpass filters.
Then instead of wrongly amplifying the highpass opamp, add a resistor to make the capacitor ratios 2:1.
Then instead of attenuating the input and wrongly amplifying the lowpass opamp, add a capacitor to make the capacitor ratios 2:1.
Is there any advantage in connecting up high pass filter first and then low pass filter to achieve a simple band pass filter. Or the other way around is better? Will there be any change based on the
Answer (1 of 3): Unique situations may have a high gain spec, so to prevent DC offset, a HPF 1st makes better sense. But your assumption is not a valid generalization, it may be done either way.
www.quora.com
LPF first can reduce noise overall, a HPF first can reduce DC offsets......
I checked both filter sequences and you were correct with the highpass filter at the input so that the lowpass filter at the output cut hiss. I was wrong and here is why:
If passive done noise sources, the Rs, would contribute symmetrical to topology, so
no change. Same for S&K done with unity G followers. But if done with VCVS, different
G's, then order would make a difference ? Or am I missing something here ?