Radically different behavior between ceramic and electrolytic capacitors in electret microphone circuit

Question

I've been trying to set up a microphone circuit for use in an AVR, essentially just to detect levels of noise. I observed some strange behavior in the amplified circuit I set up, so i elected to strip the circuit down to its base functional unit to see if I could figure out what the problem was. Schematic follows.

^{simulate this circuit – Schematic created using CircuitLab}

The resources that I've used to build this circuit all seem to indicate that the type of cap used doesn't matter, so I opted for the higher-capacitance electrolytic. This did not behave as expected.

The following is a waveform captured from the circuit using an electrolytic cap.

Now, the ceramic cap:

To be clear, these waveforms are captured while the room is quiet. blowing into the microphone produces a response.

The ceramic capacitors I have do not dampen the 60Hz noise to my liking, and it seems silly to wire a bunch of them in parallel when I have a few much larger electrolytic caps.

I've tried a few different sizes of electrolytic caps, with the same result.

I've tried reversing the polarity, with the same result.

Why am I seeing this discrepancy? How can I remedy the situation?

Additionally and secondarily, is there an established best practice for using a microphone with AVR?

EDIT turns out i'm seeing this effect because my cheap chinese power supply has a poor rectifier :^)

Answer 1

You should use a 100 kohm load resistor between +sense and -sense. Maybe even a 10 kohm would be better.

With just a basic open circuit, the electrolytic capacitor's leakage currents would place a DC voltage at the output and the signal would hit the end stops of the AVR ADC circuits.

Answer 2

Your problems appear on scope to be CM hum not distortion. But the Electrolytic has distortion from probably even more pickup on the signal.

Solution: use CM choke SM, THT or large ferrite torroid. Wires should be shielded twisted pair.

Load R affects signal gain and DC operating point which allows for AC swing near V+. A larger V+ allows a bigger R with more gain and more DC drop from V+.