Ayre style X-Preamp?

[aparatusonitus]

Hi guys,

I've been hanging around here for same time, trying to figure out (as a non EE and almost a newbee) how to implement a x-feedback to folded cascode topology (like in Nelson patent), as it seems that folded cascode sounds superior to other design if done correctly. N. Pass, C. Hensen and J. Carr (to name a few) use it with great succes, Analog Devices uses it in their opamps which sound and measure excelent. So I first try to convert a A75 input stage/ folded cascode (as preamp) to use x-feedback, but I have no luck/knowledge to succed there. Then, I try with Ayre schematic, and it seems that I finally have some luck .
Q: is it a x-feedback or not?

[aparatusonitus]

Better picture...

[Nelson Pass]

Very nice.

[aparatusonitus]

Well, what can I say, thank you very much Nelson and others heroes around here. Without your help I'll be nowhere.

During my regular work today, I have some spare time to figure out how to implement x-feedback around A75 folded cascode driver. As is explained in service manual, one may choose how to wrap around negative feedback: from folded cascode drains only, from +/- output only and from both as in new schematic. Your choise, works great in sim whatever you deside to do. Or one my choose folded cascode or regular cascode operation as writen in service manual.

I have noticed that output impendace goes up for about 100% (from 0,75 to 1,5 Ohm) when mix feedback is used. I know there is nothing to worry about, but I would like to know why is that, 'couse I would expect to go down?

Best, a

[Nelson Pass]

If you mean feedback from both the "VAS" and the output,
I should imagine that by applying only 1/2 as much output
feedback, it would roughly double the output impedance,
assuming that the open loop output impedance is much
higher to begin with, say 10 ohms or more.

[lumanauw]

Hi, aparatusonitus

R50 (1meg) and C18 (1u) are forming low pass filter?

What is the advantage of having M9,10,11,12 referenced to the output stage, compared if they are referenced to solid reference (like zener or R divider between ground and rail, not towards output node)?

[aparatusonitus]

Originally posted by lumanauw
R50 (1meg) and C18 (1u) are forming low pass filter?

I think they are forming a high pass filter, at least thats what my sim shows.

Originally posted by lumanauw
What is the advantage of having M9,10,11,12 referenced to the output stage, compared if they are referenced to solid reference (like zener or R divider between ground and rail, not towards output node)?

For one thing, you get full control over absolute and relative output DC offset thru potentiometers U1-4 (in sim I can get zero offset easily), not shure about others benefits or shortcomings.

Best,a

[aparatusonitus]

To honor Nelson, a pure second harmonic A75 X-Preamp driving 600 Ohm load

[aparatusonitus]

...and new schematic with complementary Vgs multiplier (to please my eyes), some protection and extra output mosfet pair.

[aparatusonitus]

...and finnaly A75 X-PowerAmp