Very different hybrid... opinions?

[DrG]

Thanks traderbam.

They have an interesting approach and sensible reasoning. The topology is not exactly like mine though. Mine has symmetrical emitter/collector connections for the pnp/npn devices to the power rails and output via small resistors.

But it's very, very similar.

 

[darkfenriz]

Hi Darkfenriz

My inexperience on the finer points of circuit design is clear. By my reasoning R16/R11 and R6/R12 should set the (inverted) gain for the 'solid state' part of the circuit. I take your point about the high-impedange mirror output however. As I recall, the simulation worked, for what it's worth.

In another iteration I took non-inverting feedback to the gate/grid of the 'other' input mosfet/tube pair, which also simulated ok.

I hadn't got as far as considering frequency compensation in the feedback loop yet. In fact, I would welcome a few suggestions around this.

Been toying with B2 spice v5 this weekend, hoping to have another go at simulation but I don't quite understand how to add new part libraries to the program.

Hi
Your front end is a voltage to current converter, while the feedback part is I/V converter. The gain is determined by triode transcondutance and 22k resistors.
Concerning compensation you may find a nice way

here

Regards,
Adam

 

[DrG]

Holy whack, Adam...

It's gonna take me awhile to figure out your circuit. Does it come with a manual? Seriously, could you briefly explain the operation of the input stage and VAS?

And by the way, which spice package are you using?

 

[darkfenriz]

The input stage works like a modulated wilson current mirror which compares output to input. Input transistors work as followers with one diode drop and the difference in input-ouput voltage modulates the reference current form I sources. VAS is common base stage.
I stongly advice the compensation scheme I've used, which is 67pF capacitor from Vbe multiplier to ground, 68k resistor in parallel in not mandatory.

Regards
Adam