Looks like you are having fun 🙂
Q1 and Q2 are a current mirror and force equal current through each side of the differential pair. Always remember in simulation that transistors of the same type have 100% perfect matching. You might try adding a little emitter degeneration (low value resistors) to the emitter of all four of those transistors.
here i did all kinds of final changes and it still clips a bit but i got to go for an hour or two
Nope 😉
Clue... run the sim and look how much current is flowing in Q603 and Q605 (which is the same current flowing in R1 and R2)
Clue... run the sim and look how much current is flowing in Q603 and Q605 (which is the same current flowing in R1 and R2)
OK, Q3 and Q4 is NOT a CCS because you have NPN transistors where you need PNP, and you appear to have tried to make it work by fudging the resistors.
But you have created another problem: Turn-off resistors are not optional for audio amplifier outputs. And the emitter resistors on the drivers is not a great idea.
Also, both CCS may not be stable when both transistors are the same part/speed. They should have a capacitor on the sense transistor or a base resistor.
And the temperature compensation from the VAS CCS may not be enough. R617 needs to be replaced with at least a diode+resistor, if not a VBE multiplier.
I suggest you stick with a classic "blameless" circuit.
But you have created another problem: Turn-off resistors are not optional for audio amplifier outputs. And the emitter resistors on the drivers is not a great idea.
Also, both CCS may not be stable when both transistors are the same part/speed. They should have a capacitor on the sense transistor or a base resistor.
And the temperature compensation from the VAS CCS may not be enough. R617 needs to be replaced with at least a diode+resistor, if not a VBE multiplier.
I suggest you stick with a classic "blameless" circuit.
the main issue is that the stability of the amplifier depends on one really low value capacitor. and the output of it is really rough
You might playing with these (I think they will run from the folder but extract the folder first). There is a bjt version of the ACA Class A amp and some other doodles...
This is what's in the folder:
And if you haven't already done so I would recommend ticking and saving this option in LTspice:
This is what's in the folder:
And if you haven't already done so I would recommend ticking and saving this option in LTspice:
you can very easily see that the bottom line is quite rough and that without the 100pF capacitor everything goes into oscilations and what not, 100pF is low value enough that diffrent ait humidities and other factots may influence the capacitance and so on.