Noise from preamp

[vsmGuy]

Extremely insightful replies guys!

I will get down to it after supper and update this post.

As you can see, the circuit is the result ot theoretical learning and is a very basic practical implementation.

How many of you guys use EAGLE?

This one should work in the free version of EAGLE .. I will upload the .sch file so that we can all modify this .. will that be fun?

I want to know more about a few things:

1. Why do I need to bypass the lower resistor of the voltage divider?

2. The output impedance of the electret should be pretty high - how do I measure it?

I did all my calculations assuming it was 100k.

3. Someone told me I should scrap the opamp and use discrete BJTs instead?

What do you guys think?

 

[unclejed613]

the resistor generates thermal noise proportional to it's value. bypassing it shunts noise voltages to ground.

there are various ways of measuring impedances. all of them will show various degrees of frequency dependence, so for general purpose audio, the measurement is usually done at 1khz. an indirect method is to excite the microphone with 1khz with the output open circuited (not exactly simple with an electret), and measuring the voltage. then a load resistor of known value (usually equal to your "first guess" of what the output impedance is) and measure the output voltage again. if the loaded output voltage is half of what it was unloaded, then the output impedance is equal to the load impedance. if the output voltage is different than half, you can calculate the output impedance by solving the voltage divider formed by the output impedance in series with the load resistor.

going with discretes for a mic preamp means more components, more pc board real estate, and no common mode or power supply rejection (i.e. you will end up with more noise), and if you make a discrete op amp, you will have to contend with things like compensation caps (to keep it from becoming an oscillator), input bias currents that have to be nulled out, etc... a monolithic op amp is simple, and effective, and takes up very little real estate, and generally uses less current than a discrete preamp. also in a noninverting configuration you could have an input impedance of up to 10E+12 ohms (for an LF351).

 

[audioguru]

I also thought that the FET in an electret mic is a high impedance current source until I fed a signal to one and measured its impedance as 4.7k ohms. Then it is in parallel with the resistor (10k) that powers it.

My total impedance was 3.2k ohms so the preamp should have an input impedance of at least 32k ohms for low signal loss. Since the preamp has a high input impedance then the coupling capacitor can have a fairly low value.

A preamp made with an opamp can have much less distortion than a preamp made with one or two transistors.