In the course of my research into ferrorresonant systems I found the need to analyze the small hum coming from my transformer. The most sensitive and easily obtainable microphone element around is the electret. In this article I begin by discussing what the electret is and then how to amplify its weak signal by 400 times using an LM358 op-amp. This page may be useful for hobbyists working on audio electronic projects.
A condenser microphone is essentially a capacitor formed from a diaphragm and a backplate. As the diaphragm vibrates it varies the capacitance and modulates its electric field. Earlier microphones required a relatively high DC voltage bias but many modern versions contain an electret
for the dielectric that automatically maintains this voltage bias. With a constant electrical field in the capacitor it only needs to amplify the weak signal coming out.
The electret is an interesting phenomenon that has so far found limited practical uses. It is known that some materials have a ferroelectric effect which is an electrical analog to the to the ferromagnetic effect. The effect is similar to magnetic hysteresis
except that instead of maintaining a residual magnetic field an electrical field is retained.
In the case of the electret the material is heated while an electrical field is applied. As it cools the electrical field becomes frozen in the atomic structure just like a magnet has a frozen magnetic field. This effect was discovered back in 1775 by an inventor named Johan Carl Wilcke
who experimented with a device he called the electrophorus. He was able to freeze an electrostatic field into a combination of resins and waxes and he could use it to repeatedly create a charge on a metal plate. I have heard rumors that you can actually use this principle to create an “electret diode” providing DC current but that is for another post.
Soldering The Leads
Before you get started testing the component you will need to solder positive and negative leads. For this example I am using a 270-090 condenser element obtained from Radio Shack for about $3. If you look at the photo below you will see that the negative lead is connected to the cylindrical body by 3 small contacts. The leads are 22 AWG solid wire.
Be careful not to heat up the element because you could melt the electret. First heat up the solder on the copper wire and then quickly bond it to the element lead.
I was hoping that I could simply use my oscilloscope to measure the very weak audio level I am working with but it is too noisy and requires amplification. I found an LM358 opamp in my component collection. The signal I am trying to amplify 400 times is about 5mV which requires sensitivity that pushes the limits of the op amp. If you are dealing with louder audio then you probably won’t need such sensitivity.
Update: This circuit uses a cascade configuration to increase the gain and bandwidth of the signal. We are amplifying by 400x and one amp in inverting configuration is not enough. It was pointed out in a comment that the noise can be reduced if this is changed to a non-inverting configuration.
In this diagram you can see the 270-090 element represented as the pickup and an NFET amplifier. The NFET is actually inside the body of the electret and is needed to amplify its extremely weak signal for input in to the preamp. This is a 2 stage A/C amplifier. Because of the relatively low slew rate I needed to cascade 2 amplifiers to get descent bandwidth.
Gain = R3*R7/R4 = 400
R1: This channels power into the NFET
C1: This is a ceramic capacitor. I tried tantalum but they don’t deal well with the weak voltage.
R2 and R5: These setup a voltage divider that provides a 4.5V reference offset for the signal.
R6 and R8: These are necessary to pull current from the op-amps so they operate correctly.
The completed prototype
The image below shows an input test signal of 5mV being amplified to 2V. The square wave reveals the frequency response of the amplifier. There is some distortion but the noise is a bigger problem. The ML358 will not deliver much better performance then this but for signals below 10 Khz performance is not bad.
Square wave test
Here you can see the results of putting a weak 6 Khz audio signal into the mic element. At this level of sensitivity even the halogen lamp on my desk adds interference. The blue trace below is so noisy you can hardly read it. The yellow trace is amplified 400x and can be properly analyzed by the scope.
6 Khz audio test