Do you like Class A?

[Stee]

this is an intresting solution @ low power
need a big dissipator

[Ipanema]

What is the purpose of including Q2, Q3 and its twin?

[Stee]

increase the current to drive final device

[Stee]

two resistors can make this job

[Circlotron]

Not entirely sure how to explain this but seeing the output transistors have the collectors to the load they are effectively current sinks so when the output swings though half rail I expect there would be less chance of crossover distortion even if it were class B. As class A, so much the better.

[Mooly]

Well the topic is "Do you like Class A?, and my answer is that for power amplifiers I do not

Reasons (just a few) include excess heat generation and excess cost and bulk. I believe the class of operation is but a small part in how a final design of an amp actually sounds... and that is the most important aspect.

[Stee]

viceversa common collectors on output does
at the same time
for accuracy
it's important common emitters on input

[lineup]

I tested in SPICE.
It works, with rather low distortion.
Not very low. But low. THD 0.012% at 1 Watt

Interesting idea, Stee. A good one

[Stee]

probably need a reduction of output resistors to 0.1 ohm
can you post the printscreen of SPICE values?

the original schematic comes from Musical Fidelity A-1 (enhanced VAS)
common emitters input schematic from LITE A-100 (common base)

[Stee]

separate VAS from OS as usual
but common source input stage
and again CFP with bias control

(mix comes from Musical Fidelity AX-50 and Mark Levinson M9)