I was playing around with P-SPICE a little and came to the circuit posted below.
I first wanted a fully complementary TWO-stage amplifier with MOSFET output stages and the results didn't look that bad but not too promising either.
One of the requirements were that the output devices have to be operating in common source.
So I went further and came to the "final" circuit after some additions.
The whole cicuit is inverting and has four inner feedback loops (2 x 2 to be exact) and one overall feedback loop. The feedback network might have to be rearranged in practice (i.e. so that smaller resistor values can be used) because it might be difficult to use that small a capacitor (C5) because the circuit's parasitic capacitances may be much higher than that.
There is a restriction to a maximum of 10 transistors with my trial version. That's why I used DC voltage sources instead of transistors, zeners etc for bias stabilising, DC-offset correction etc.
I also wanted to use cascodes instead of just Q7 and Q8 but with the aforementioned restriction .........
Also the devices used are not the most sexy ones, but I am always too lazy when it comes to import models .......
What I specially like is the driver/output stage topology. It have first seen this in the German magazine Elrad about 20 years ago used within a 300 watt BJT amp. I think this design was originally published in an Australian magazine before, but I'm not sure. With this topology driver tansistors with lower Uce rating can be used which are more easily obtained and sometimes have higher fT and higher beta.
The output devices have to be biased quite high (around 200 mA) to obtain good performance.
I know that simulations and "the real world" will never be the same. But the figures I got looked promising anyway:
THD figures were almost the same for 1 kHz and 20 kHz and are mainly consisting of k2 and k3.
For an output voltage-swing of +- 10 Volts (i.e. 20 Vpp) k2 was 0.005 % and k3 was 0.002 %.
For +- 1 Volt k2 was 0.001 % and k3 was 0.0005 %.
Frequency response reached from DC to > 800 kHz.
Regards
Charles
I first wanted a fully complementary TWO-stage amplifier with MOSFET output stages and the results didn't look that bad but not too promising either.
One of the requirements were that the output devices have to be operating in common source.
So I went further and came to the "final" circuit after some additions.
The whole cicuit is inverting and has four inner feedback loops (2 x 2 to be exact) and one overall feedback loop. The feedback network might have to be rearranged in practice (i.e. so that smaller resistor values can be used) because it might be difficult to use that small a capacitor (C5) because the circuit's parasitic capacitances may be much higher than that.
There is a restriction to a maximum of 10 transistors with my trial version. That's why I used DC voltage sources instead of transistors, zeners etc for bias stabilising, DC-offset correction etc.
I also wanted to use cascodes instead of just Q7 and Q8 but with the aforementioned restriction .........
Also the devices used are not the most sexy ones, but I am always too lazy when it comes to import models .......
What I specially like is the driver/output stage topology. It have first seen this in the German magazine Elrad about 20 years ago used within a 300 watt BJT amp. I think this design was originally published in an Australian magazine before, but I'm not sure. With this topology driver tansistors with lower Uce rating can be used which are more easily obtained and sometimes have higher fT and higher beta.
The output devices have to be biased quite high (around 200 mA) to obtain good performance.
I know that simulations and "the real world" will never be the same. But the figures I got looked promising anyway:
THD figures were almost the same for 1 kHz and 20 kHz and are mainly consisting of k2 and k3.
For an output voltage-swing of +- 10 Volts (i.e. 20 Vpp) k2 was 0.005 % and k3 was 0.002 %.
For +- 1 Volt k2 was 0.001 % and k3 was 0.0005 %.
Frequency response reached from DC to > 800 kHz.
Regards
Charles
Hi Kanwar
It is a bit different, a diamond buffer driven CFP. I bet it would make a good pre-amp circuit.
I wonder if the CFP stage with it's local feedback, could not provide enough voltage gain to drive the driver/outputs as followers. Source follower is more stable and easier on the driver stage wrt Cin, IMO. But then you would have to increase the front end VAS supply voltage to account for loss of Vgs....or use laterals.
It is a bit different, a diamond buffer driven CFP. I bet it would make a good pre-amp circuit.
I wonder if the CFP stage with it's local feedback, could not provide enough voltage gain to drive the driver/outputs as followers. Source follower is more stable and easier on the driver stage wrt Cin, IMO. But then you would have to increase the front end VAS supply voltage to account for loss of Vgs....or use laterals.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.