There have been many hits on this web page recently so I thought I should add some notes and let viewers know that this circuit is far from perfect. It was designed to use parts that I had on hand and to use as few of them as possible. This circuit is to be used with voice only in an intercom system. This circuit will cause a lot of distortion when attenuating a strong signal at the threshold of signal reduction due to the input to the FET source turning the FET on and off. Feeding some of the signal at the source or drain to the gate will eliminate some of this distortion. See the schematic at:http://www.electroschematics.com/248/automatic-gain-control-agc/
Note the 100K resistor and the .1uf cap going to the gate. The dynamic range of the circuit on this page is somewhat limited because the FET is shunting a fairly low impedance and feeding a low impedance following stage. See the schematic at:http://electronicdesign.com/analog/effective-agc-amplifier-can-be-built-nominal-cost
In this circuit the FET is shunting a high source impedance of 120K and looking into a high impedance load of an op-amp. The circuit on this page meets my needs for an intercom mic. preamp. The electret mic. does not produce enough output to cause much distortion and the dynamic range is sufficient for an intercom. The ear is surprisingly tolerant of distortion on voice as long as it is not crossover distortion. If one is looking for an AGC element to control a signal and distortion is of concern, consider a Cds photo cell instead of a FET. There are several integrated circuits on the market which are designed to operate as AGC preamps and may be a better choice than using discrete parts.
M1 ELECTRET CONDENSER MICROPHONE
R8 2.2meg (see text)
C1,C2,C3,C4,C5,C6 .33uf 50V
C7 10uf 16V
D1,D2 1N914 or 1N4148
This pre-amp will be used as the front end of a home intercom system
but could be used ahead of the modulator of a transmitter to provide speech
compression. The supply voltage should be held to within +or- 1 volt
after the compression level has been set. The value of of R8 may have to be
selected if transistor Q3 is other than a 2N5089.
The 2N5089 has a very high beta usually 800 to 1000. A value should
be selected to give a voltage of one half the supply voltage at the
collector of Q3. The drain and source of Q2 may be interchanged.
If the compression level of circuit Figure 1 cannot be set low enough
use the circuit Figure 2. R11 is used to set the compression level in
Detailed circuit description
M1 is an electret condenser microphone, R1 provides power to
the microphones internal fet. C1 isolates the D.C. levels of the microphones
output and the bias divider R2, R3. The divider R2, R3 sets the operating point
for transistor Q1. Q1 has a large un bypassed emitter resistor R5. The unloaded
voltage gain of this stage would be close to 10. The gain is determined by R4
divided by R5. Any load placed on this stage would appear to be in parallel with
R4 and would lower the gain. Thus the stage has a high output impedance. The fet
Q2 acts as a voltage controlled load on the Q1 stage. Capacitors C2,C3
isolate the fets drain and source from the D.C. output of Q1 and the bias on the
base of Q3. Resistor R6 maintains the drain and source of Q2 at +12 volts. Gate
resistor R7 may not be necessary but it lowers the high frequency gain of the
fet and makes me feel better. The drain and source of Q2 may be interchanged
with out changing the operation of this stage. Resistor R12 returns the fets
gate to ground. With the fets drain and source at +12 volts and the gate at
ground the fet will be cut off and no load will be placed on the Q1 stage
allowing maximum gain. Q3 is a high gain stage to provide enough output to be
useful. R9 is the load resistor for Q3. R8 provides bias for Q3 and should be
selected to set the voltage at Q3s collector at one half of the supply voltage.
If a scope is available R8 may be selected by setting the bias to show equal
clipping of the positive and negative swings of the output signal. Resistor R10
is not necessary but I included it to get similar gains out of all 4 units I
will be building. If maximum gain is desired from this circuit leave out R10 and
just ground the emitter of Q3. C7 bypasses the power supply at the circuit board
and keeps the amplifier stable, without this capacitor the amplifier will
oscillate. C4 keeps any D.C. off of the output. D1,D2,C5,and C6 develop a
positive voltage to control the fet and the overall gain of this amplifier. The
positive voltage is proportional to the output of the
Q3 stage. D1 clamps any negative going wave form to ground so the only output to D2 is positive. D2 passes the ground to positive wave form on to C6 and R12 and on to the gate of the fet Q2. The decay of the AGC voltage at the fet gate is controlled by the time constant of R11 and C6.
The positive AGC voltage is proportional to the peak to peak output of
Q3 minus the diode drop of D1,D2 of approximately 1.2 volts. Circuit Figure 1 generates the AGC voltage only from the output of Q3. This does not provide enough gain control with some fets so circuit Figure2 was developed. The circuit in Figure 2 references diode D1 to a variable plus D.C. voltage set by R11. The AGC voltage is then that set by R11 plus that generated by the output of Q3. If the pre-amp is used alone I would suggest using the circuit in Figure 2. I will be driving a power amp with this pre-amp and its output will be 30 volts peak to peak so I will use circuit Figure 1. The hot side of R11 will be connected to the output of the power amp. There is nothing magic about all of the .33uf capacitors used in this design, I just have a bag of them I bought at a flea market. The .33uf caps just happen to roll off the frequency response around 300cps so this is good for speech. If this pre-amp is to be used for music, all of the .33uf capacitors should be replaced with 4.7uf capacitors. Any hi gain npn silicon could be used for Q1 and Q3 the higher gain the better in the Q3 position. Transistors with a beta from 39 to 900 were tried in the Q1 position and worked fine with no circuit changes. I used 2N5089s in the Q1 and Q3 positions because they are hi gain, low noise and I had a bag of them from an earlier project. The fet in the Q2 position could probably be any N channel device. The 2N4856 used has a saturation resistance of only 25 ohms and probably gives more gain reduction than some others. There was an article years ago in a ham magazine of a circuit like this using an op amp instead of transistors. The article was called "The Amazing Audio Elixir" It was in the September 1979 issue of 73 magazine page 116.
For another audio AGC scheme check out the method used by "RADIO ANIMAL" in his Grenade pirate transmitter. The resistive optocoupler as of 8-31-08 is available from Electronics Goldmine and Allelectronics.
Return to index page