The design requirements would atleast include amplifying ~100 mV signal to ~1V with a maximally flat response. ICs and discrete solutions should be considered. Biasing networks and component values can be simulated!
After about two years of owning a pair of electret microphones, I finally realized that I could learn to bias these by using a DC supply and a scope. I have seen reference biasing circuits before, so I knew that I would have to apply a few volts to each mic, and the transduction process would make the sensed pressure waves ride on top of this DC bias.
So I hooked a supply up (~1V, current limited to 0.5 A) to one of the two mics I had, but I couldn’t get it to detect a sound. I happened to touch the mic after some time, and it was hot…never a good sign…turns out that I had missed using a ~1k resistor between the supply and mic’s port, which is probably used to decrease the amount of current going through the supply to the mic. That burnt one of the two mics.
Next, I finally pulled up the reference circuit and confirmed that I had forgotten two key elements: a resistor between the supply and the mic, and an AC coupling capacitor(?)
[ by Tomi Engdahl ]
vcc O | / \ / \ EQUAL TO REQUIRED IMPEANCE / EX.= 1000 Ohm \ | | CAPACITOR MOST ANY VALUE 10uF 16v |---| | + | / | |---0----| |------- AUDIO OUT | | | \ | |----O-------------- |---| | | ----- --- -
So, I got the resistor and the cap in, but still couldn’t get anything (as I was still hooked up to the burnt out mic.) In the meanwhile, I realized that I was using a 10uF electrolytic capacitor with the wrong polarity. Replaced it with a ceramic capacitor:
The wired up circuit, with a supply and scope probe hooked up to a jack (thanks for having a jack lying around, GVU!)
Anyway, the other mic worked with this setup (I tried a 3V bias, and limited current to 5 mA – but the supply just read 0.000 A, so the current draw must have been <1 mA.) But I couldn’t get the first mic (burnt due to the missing 1k resistor) to work.
Furthermore, I noticed that a 50 mV / 4 ms scale gave a good looking output on the scope. But every time I’d power up the mic again, the output voltage would shoot up to supply, and then trickle down to a ~100 mV range. This behavior reversed when I tested the burnt mic again. In that case, the voltage would start at a negative value (wrt to the ground terminal), and shoot up. This gave me the silly idea for reverse biasing the mics, and that burnt the second one out.
(1) Use elements shown in reference circuits. They are probably there for a reason.
(2) Do not reverse bias a JFET. If a JFET / electret mic is drawing ~3 mA of current, its probably busted.
Had some more fun with the supplies next: