Audio Spectrum Analyzer V2
It works! If you have already seen the V1 from last year, you will know the difficulties I ran into: switching noise from the power supply and LED driver got into the mic circuit, got amplified, and drowned out the audio signal.
Although the noise is bad to start with, I did make some mistakes the first time that allowed it to be such an issue. The main one was the biasing on the electret microphone, but amplifier biasing was also problematic:
Biasing the mic with a simple resistor is the classic, textbook way to go. Also... it sucks! Noise on the supply line gets right into the audio signal through the resistor, and that's it. This time around, I wised up both regarding the mic and the amplifier chain:
I am still biasing the mic with 20k of resistance from 5V, but here I distribute it and include low-pass filters in the biasing chain. By distributing half of the biasing resistance on either side of the mic, and splitting each of the 10k resistances into 2 with a LPF in the middle, almost all of the noise present on the supply lines is rejected.
To further ensure low-noise-ness, I then pass the signal from top and bottom of the microphone through a difference amplifier to reject any common mode noise that has made it through. Second stage is just a normal non-inverting amplifier, albeit with a voltage reference and LPF setting the bias point to - once again - reject noise.
