Amplifiers
Common Emitter with Shunt Feedback
The above circuit is about as basic as the voltage amplifier gets. The emitter resistor can be replaced by a straight wire to ground if desired. An internal emitter resistance will still be present. The input resistor may also be eliminated for minimalism. The collector has two resistors. One supplies the positive voltage to the collector and serves as a load to generate a useable output voltage. The other supplies the base with the bias voltage to turn the transistor "on." In addition to supplying a voltage to the base, this collector-base connection also provides negative feedback for both ac and dc conditions. There is often another resistor from base to ground that improves circuit stability and can be used to fine tune the operating point. I left the base-to-ground resistor out in the above diagram.
The Astro Amps Astrotone is a good example of a basic version of this circuit. The Zvex Super Hard On is a nice variation that uses a discrete MOSFET and a variable source resistor as a gain control. The Electro Harmonix Big Muff Pi uses this configuration in 3 of its 4 transistor stages.
The shunt feedback resistor closes a negative feedback loop and complicates the gain calculation. The gain of the stage is first determined by the "open loop" gain, which can be imagined in the above diagram. Open loop gain can be roughly estimated to a bit less than the ratio R1/R2. For example, if R1 is 10k and R2 is 1k, then gain will be a bit less than 10. (In testing, I got a gain of 8.5 for such a circuit.) We can try to make R1 larger or R2 smaller to push the gain further. R2 can even be eliminated as an external resistor, but an intrinsic resistance in the emitter will still be present, so "R2" as shown above can never truly reach 0Ω and thus allow for infinite gain. Neither can R1 ever reach ∞Ω in a simple single transistor circuit (using transistors as active loads, especially in IC fabrication, near ∞Ω loads are possible).
If R2 is an external resistor, it does create negative feedback because the voltage developed across R2 works in opposition to the voltage developing at the collector. If a large bypass capacitor (large enough to have a low reactance to the lowest signal frequency of interest) is installed across R2, then R2 is effectively shorted at signal frequencies, which increases the signal gain significantly.
Open loop gain can quickly get beyond what is needed to simply get a small signal amplified to line level, and even a tiny signal will clip the output. At this point, negative feedback can be introduced to tame the amp.
Returning to our original shunt feedback amplifier at the top, we now have an input resistance and the feedback resistor. If the open loop gain is much higher than the closed loop gain, then the shunt feedback gain can be approximated to the ratio of Rfeedback/Rinput.
Common Emitter with Series Feedback
This circuit will usually have all four resistors present. The emitter resistor is nearly always included, and it is where we get our series feedback from. Remember that the common emitter works such that a larger base voltage produces a smaller collector voltage and vice versa. The base and emitter voltages have an in-phase relationship. We can count on the emitter to follow the base voltage less one diode drop while the transistor is forward biased. The voltage developed across the emitter resistor works in opposition to the collector current, and thus the collector voltage as well. Just as in the shunt feedback amplifier, the emitter resistor may be bypassed to eliminate the negative feedback at signal frequencies.
For maximum stability, the base resistors must have a low impedance compared the the transistor's input impedance. This ensures that the base voltage remains constant regardless of the specific transistor's hfe parameter or temperature at any given moment. This is turn guarantees what the emitter voltage will be (about 600mV less than the base voltage), which in turn guarantees quiescent collector voltage so long as the transistor is active (neither cut-off nor saturated). Because of these "guarantees," this design is widely used and highly regarded due to its freedom from beta and temperature issues. The primary drawback is that low value base resistors are not always practical (they load the input) so that must be addressed in some sort of compromise or complication, examples being: less stability with high value resistors, high beta transistor for higher transistor input impedance, stick a buffer stage before this, add bootstrapping, etc.
This circuit can be found as a stand alone effect, usually called a "booster." The Electro Harmonix LPB-1 is a good example of this. The Big Muff Pi and the Univox Super Fuzz use this circuit for the make up gain amplifier following the passive filter stage.
More on the Common Emitter with Series Feedback (Design Example)
2 Common Emitters with Shunt Feedback
Here is a dc coupled two transistor topology famously used by the Fuzz Face, Tone Bender, and other fuzz pedals. The use in fuzz designs seems to have stemmed primarily from English companies, namely Vox and Sola Sound, so I have a separate article regarding what I call the "2 Transistor English Fuzz" or 2TEF for short. Outside of guitar pedals, this topology is common in old '50/'60s designs. The more I look for it, the more I see it. It is not usually drawn in the above manner when found in an old hi-fi schematic though.
Note that if you remove Q1 and the resistor connector to its base, the Q2 circuit looks an awful lot like the previous series feedback example. And Q1 looks quite similar to the previous shunt feedback example. That is exactly how you should approach this circuit: it is the combination of the previous two topologies. Q1 is effectively replacing the lower base resistor in the simple series example.
Broken apart, we have two common emitter voltage amplifiers cascaded together. The voltage developed across Q2's emitter resistor gives us series feedback. The same voltage applied to the base of Q1 gives us shunt feedback. Both instances of feedback are negative feedback, so this amplifier can function as a nice linear amplifier if setup for that kind of operation. The cascaded transistors also provide for plenty of gain if the negative feedback is reduced. The topology is used in a number of English fuzzes where the emitter resistor is completely bypassed by a capacitor when the FUZZ control is at maximum clockwise. This causes the our negative feedback to disappear at signal frequencies and the maximum gain of the circuit is achieved causing square wave distortion at the output.
When shunt negative feedback is applied to the input, input impedance lowers. As Q1 will often have no emitter resistor in fuzz designs, like the Fuzz Face, Q1's input impedance itself is naturally very low, and the shunt feedback resistor lowers this even further. If there is a large bypass capacitor across the Q2 emitter resistor, then input impedance is going to be the value of the shunt feedback resistor in parallel with the input impedance of Q1. Because of the lack of an external feedback resistor, Q1's input impedance is going to depend on the beta, or hfe, parameter, and this in turn is affected by temperature changes. If the transistors are germanium, the instability gets worse yet. Practical arrangements for non-fuzz applications generally include an emitter resistor for Q1, as seen in the next diagram.
Q2 can be a common collector by taking the output from the emitter circuit and eliminating Q2's collector resistor. Adding the emitter resistor at Q1 increases input impedance and circuit stability. These changes result in less overall voltage gain than the previous circuit. An example of a nice x10 amplifier utilizing this topology is seen in the first stage of the Univox Superfuzz.
Adding an additional feedback path (Q2 collector to Q1 emitter), which must be high impedance, as seen in the Univox Square Wave. The Colorsound Overdriver takes only ac feedback from the far side of the output capacitor.
3 Transistor Shunt-Series Example: Altec 1592A. Buffering the output allows for the feedback resistor at Q1's emitter to be lower in value. This is used as a mixer amplifier in the 1592A, and the shunt input is a bus that has five input channels connected to it.
2 Common Emitters with Series Feedback
If we allow ourselves to use 1 NPN and 1 PNP, we can make a neat 2 transistor circuit with series feedback at the input. The feedback loop reminds me of an ouroboros, which is a snake that eats its own tail. Now Q1 resembles our simple series feedback amplifier, and Q2 resembles the simple shunt feedback amplifier. Instead of Q1's emitter going straight to the power rail, it takes a detour to get feedback from the output of Q2. This topology appears to be a favorite of Roger Mayer. Series feedback at the input increases input impedance.
This is the arrangement seen in the old Octavia preamp traces I've seen floating around. Q1 and Q3 are the same as the first version, with Q2 now buffering Q1's relatively high impedance output.
Another variation where the buffer is now at Q3. Output impedance is lower which allows the feedback resistor connected to Q1's emitter to be lower.
The NPN and PNP devices are reversed, and Q1's emitter is now between 2 resistors instead of having one in series. This is the Complementary Feedback Pair, or Sziklai pair. The Heathkit TA-28 Fuzz is an interesting application of this circuit.
Differential Amplifier or Long Tailed Pair
This arrangement of 2 transistors allows for the amplification of the difference between the two inputs. Ideally, there should be no output when the inputs are equal, and full output when the inputs are very different. The basic form shown above has 2 inputs and 2 outputs. This is used in balanced and splitting circuits. Often only one output is used, and the other collector resistor can be eliminated.
Above is a differential amplifier which samples the output from the dc coupled common emitter stage as negative feedback. The output is phase coherent with the input due to 2 phase inversions from both the input and the voltage amplifier stages. The in-phase output is attenuated by the a resistor divider network and then sent to the other differential amplifier input. Good examples of this amplifier can be found in the schematics for the Boss OD-2, OD-2R, and PW-2. Adding an additional buffered output stage in the form of an emitter follower would be the next evolution of this circuit.
Pages related to discrete transistor amplifiers:
Troubleshooting examples: