1. Mike's simple circuit is OK but is rather subject to the beta of the transistor. In practice it may well meet your need. See at end for complexish ravings on gain of original circuit. 2. All transistors are not created equal as regards noise - if you have a choice choose a type specified as "low noise" (could be the start of a lovely argument here). 3. The circuit's distortion performance can be improved dramatically by placing a small resistor (approx 5 to 10% of collector resistor) between the emitter and ground and bypassing it with a 10uF electrolytic capacitor Ce. Chances are, with the Isd's ~ 8 bit accuracy, this won't matter. 4. Gain: Believe it or not, the gain of a one transistor circuit (simple or modified) is about (Vcc - Vcollector)*40 Where Vcollector is the DC voltage at the collector in the absence of input signal. As this is fairly fixed with the 2 resistor biasing system, regadless of beta, the gain does not vary with transistor characteristics. This expression applies to a circuit with emitter to ground OR with the emitter resistor bypassed with a capacitor. "Why" is probably covered in "The art of electronics" (which I have not got so it may not be).. 5. A more complex and (only possibly :-) ) better circuit follows: Rc == R3 etc More controlled gain can be had by using 2 resistors from the base instead of R2 - R2 now goes to Vcc and a new R4 to ground with each very roughly around 100k.(actual values to set Vbase and thus Ve and indirectly Vc - see below), plus an emitter resistor Re. In this circuit Re MUST be used . If this diagram looks a mess and you are a Windoze user, copy into Notepad. If it still looks a mess it may be :-) Vcc Vcc Vcc | | | | | R3 | | | | | *-------C2---- Out R1 | | | R2 | C | | |/ *---C1---*--| | | |\ MIC R4 > E___ | | RE | GND GND | Ce GND Gnd Re ~ Rc/10 say (eg Rc = 10K and Re = 1K - could be less) Decide drop across Rc to set desired gain - say 2.5v for gain of 100 (From above: Gain = Vcc * (5-2.5) = 100) Drop across Re will be 10% of this (as both share same current)(almost) So Vre = 0.25v Vb is 0.6v more than Ve so Vb = 0.25+0.6 = 0.85 volt. Choose R2, R4 to set Vb = 0.85 volt. VR4 = 0.85v VR2 = Vcc-0.85 = 5-0.85 = 4.15V R's are in approx proportion to voltage drops (ignoring here transistor base current). R2:R4 ~ 4.15:0.85 ~ 5:1 Say R2 = 100K, R4 = 22K Some playing will still be required but the gain will be variable to suit over a wide range and will be moderately stable once set. (For a 10K Rc The base current will be about 1uA for a beta of 300ish and will affect the result somewhat) Gain = (Vcc - Vcollector)*40 as above Where Vcollector is the DC voltage at the collector in the absence of input signal. As this is fairly fixed with the 2 resistor biasing system, regadless of beta, the gain does not vary with transistor characteristics. 6. GAIN OF ORIGINAL CIRCUIT Only for the pathologically enthused: B = beta Rc = collector resistor Rb = base resistor Gain ~ 40 * (Vcc - Vc) Vc = (VccRb + 0.6.B.Rc)/(Rc.B + Rb) The following table shows the approximate gain which you will get with the simple circuit for various values of Rb, the base resistor, and beta (transistor current gain). I have used the exact table values in the text so they can be compared to the table. Actual gains will be of this order. As may be seen, for low values of beta (say around 100) (such as an el- cheapo junk box transistor MAY provide, the gain may be varied quite widely (56 to 165) by varying Rb. For high betas (300 is entirely typical) the gain only varies from 147 to 194 as Rb is changed from 600,000 ohms to 50,000 ohms. Table was made in Excel Displaying in non-proportional font MAY help get columns aligned. Vcc = 5 Rc = 4700 Beta 50 100 150 200 250 300 350 400 450 500 Rb Gains 50000 165 181 187 190 192 193 194 195 195 196 100000 140 165 175 181 184 187 189 190 191 192 150000 122 152 165 172 177 181 183 185 187 188 200000 108 140 156 165 171 175 178 181 183 184 250000 97 131 148 158 165 170 174 177 179 181 300000 88 122 140 152 159 165 169 172 175 177 350000 80 115 134 146 154 160 165 169 172 174 400000 74 108 128 140 149 156 161 165 168 171 450000 69 102 122 135 145 152 157 161 165 168 500000 64 97 117 131 140 148 153 158 162 165 550000 60 92 112 126 136 144 150 155 159 162 600000 56 88 108 122 132 140 147 152 156 159 eg for Rb = 100K and beta = 200, gain = 181 (vvv approximately) -----Original Message----- From: Mike Keitz >On Wed, 3 Feb 1999 11:08:25 +1100 Steve Ridley >writes: >>So I now need a simple but >>low noise if possible, preamplifier for an electret microphone. >>Something using an op-amp or just a transistor would be fine. >> The input impedance of the ISD33000 analog input is 3k and it takes a maximum of 32 mV p-p >>when fed in single ended input mode. > >If you have an extra op-amp, may as well use it. Otherwise this >deceptively simple transistor circuit should meet your needs: > > Vcc Vcc > | | > | R3 > | | > | -----*-------C2---- Out > R1 | | > | R2 | C > | | |/ > *---C1---*--| > | |\ > MIC > E > | | > GND GND > > >The transistor is any old NPN you can find. To start with, try 4.7K for >R1 and R3 and 470K for R2. C1 and C2 should be at least 0.1 uF. Higher >capacitance will increase bass response but that may not be what you >want. If you have a 3-wire microphone, R1 and maybe C1 are built into >the microphone itself. > >If you have too much gain, make R2 smaller. This also changes the DC >bias on the transistor, but since only 32 mV of output is needed it is OK >to operate the transistor close to saturation. You can also add a >resistor in series with C1 to reduce the gain. If someone will be >speaking directly into the microphone, it may be able to drive the ISD >chip without an amplifier at all. > etc