Discrete Class AB Amplifier
Specification:
|
GAIN |
27 dB |
| INPUT IMPEDANCE | 24kOhm |
| FREQUENCY RESPONSE | 10Hz-30KHz |
| THD | 0.04% by 1W |
| POWER | 60W by 35V, 8Ohm |
| POWER | 100W by 35V, 4Ohm |
| DC OFFSET | minor of 0.1V |
Description:
The schematic show many common charapters to many '80 amplifiers like Pioneer, Sony, Mashita.... The circuits present also a tipic auxiliary frontend designed by Elliott [1], SIM ( Sound Impartment Monitor ) that is very useful in benchmark-phase. Another feature the BYPASS input.
The circuit maked up by four blocks: input stage, main amplifier, drivers, output stage.
The input stage is a differential pair and therefore the bias of this stage is a simple current pump (Q3 in schematic). The green-led in current-pump is a simple way to obtain the voltage required for biasing the base of bjt (this element can be replaced by a couple of silicon diodes or, better, a zener of opportune value).
The main amplifier is a PNP-BJT (Q4) in CE (grounded-emitter) configuration, the resistors serie permit the polarizzation of the BJT of driver-stage. This amplifier operate in class A as you can easly verify.
The drivers are a couple of BJT (Q5 & Q6) with purpose to iniect the base current in the output-stage. This work is simplified by a Vbe's multiplier another BJT (Q9) that operate in this sense too.
Finally the output-stage that is maked up by classic push-pull configuration therefore by a couple of complementary BJT (Q7 & Q8).
The feedback from out to the input is implemented by the resistors R4 & R5 with C3 (AC-bypass).
We can note as the schematic tend to simmetric topology but not in main amplifier. This is a peculiarity of more of '80 amplifier.
I like show the original Elliott's schematic in which the stage-bias-voltage is the same for all stages. In one of variants that I tested I tried to lower the bias of the first two stages by a resistor with two capacitors ( pi-greco configuration) this is possible but it's necessary to decouple the base-bias of drivers. This can prevent damage of main amplifier for high-power amplifiers (more of 100W). In my actual realease I lowered bias at +-25V and changed resistor's values of main amplifier. Another test that I will try (in the future!) is to change the main amplifier to try to make a full simmetric topology.
Realization:
And my homebrew loudspeakers..........
