4 transistor 40W class A 0.0004% THD amplifier

[Kokoriantz]

The output stage is a voltage follower that I was using with an input step up transformer. I didn't tried it out yet with this configuration but only on Tina simulator.

The temperature dependent CCS LM334 is mounted upon the pmos.

[T3sl4co1l]

Looks like something Nelson Pass would cook up (or maybe did?).  Good for keeping warm on these chilly winter months.  Well, in Vietnam maybe not so much.  Maybe in the mountains? Idunno.

Tim

[Geoff-AU]

Class A is fun and all but it's needless environmental terrorism especially at speaker level power outputs.

I made a lovely valve headphone amplifier with 0.006% efficiency (1mW audio out, 15 watts of power in).  Kinda stupid really.

I'm going to re-use the valves from that amplifier in a Class AB amp for my turntable.  The tube in a "starved anode" mode is a low voltage circuit and produces some of the fun distortions at higher volumes.  There's no reason why you can't follow that with an accurate AB push-pull or even Class-D (sacrilege!) power amp.

[Kokoriantz]

The above circuit can have the output replaced by a lateral mosfet complimentary pair of EXC10p/n16 biased 150ma and provide 40w with 0.002%THD at 7.5W. The problem with discrete component amplifiers bellow 50w are replaced by chips, as LM3886, only 3$ cost. when it comes to class D, only recently we got fast drivers/ GaN transistors that can achieve down to 4ns dead time resulting open loop THD of 0.007% at 1W or 100W, see https://www.eevblog.com/forum/projects/20bit-pwm-class-d-amplifier-idea/msg3713314/#msg3713314     

[Kokoriantz]

The output, Nelsons or mine is SRPP, shunt regulated push pull circuit invented in tube era.
 
Nelson uses two weighing resistors and a transistor making it much slower.
By using lm334, I get much simpler and very fast shunt regulator.

[BrianHG]

An LM334 has more than 4 transistors in it let alone the rest in your design.
Though, it is still simple...

[Kokoriantz]

The LM334 serves as temperature compensator, it doesn't make part of the signal path. The zener diodes and their resistors can be replaced by voltage regulators for better PSRR.