Possible improvements for Aleph?

[Nelson Pass]

I just know there's going to be a point to this...

[kenpeter]

I thought the point was merely to keep you entertained?

I was actually about to make a point regarding the mirrors.
But I'm afraid my mind has gone and wandered off again in
another direction. Something cool enough it might just be
worth the temporary distraction. We'll get back to it soon.

Born of several nutty ideas tossed about this tube thread:
Anti-Triode SEPP, how to do best?

Aleph with DC servo!

At DC, its a voltage source.

At subsonic, its a Gyrator
of increasing inductance.

Above 20Hz, its an Aleph.

Go figure...

[Nelson Pass]

You're not done yet. I still have some you haven't listed.

[kenpeter]

Allrighty then, a Challenge!
Does your magic bag of tricks contain this one?

[kenpeter]

Obvious compromise made in the upper right corner.
Wringing what I can out of LTSpice's default libraries.
Better matched compliments would certainly help.

[Nelson Pass]

Originally posted by kenpeter
Allrighty then, a Challenge!
Does your magic bag of tricks contain this one?

Yes it does.

[kenpeter]

Outside the context of abusing 200mV offset chip amps?

I can't find the link now. Not 100% certain that was you.
Maybe just something you had commented on. I don't recall.

Been driving myself crazy trying to follow long ago moved
or otherwise not functioning Pass links posted in forums.
I'm largely having to re-invent the old wheels over again.

Whatever, its entertaining...

[revintage]

So when you are into complemantary solutions:

Why don´t do it all MOSFET in Jung style.

[kenpeter]

Cause I can't predict the VGS-ON of any specified MOSFET.
Much less that any random pair are going to match well.
VBE-ON of Transistor (or Aleph) is much more repeatable.

My design above is already obsolete. I'll soon show you
exactly what I have changed, and more importantly: why.

But I got 150 Class-D amps and a few trays of DSP eval
modules waiting on me to push em thru the test bench.
And thats assuming none of em decide to fight back.

I don't know if I'll get enough of a break at work today
to revisit the Aleph. But maybe sometime this week.

[kenpeter]

This drawing isn't clean as I wanted to show it.
But will have to do for now... Time is short.

The emitter coupled Aleph is a base follower,
not a source follower. And the voltage gap
between bases can be partially spanned by
other voltage drops besides Iq set resistors.
Schottkys here tighten the damping factor.

In the old drawing: currents into the gates
look like dErAnGeD squiggles, not sines...
The collectors were overshooting the mark,
and the local feedback was not helped by
bigger gate or base stoppers, nor extra
miller cap.

R3 and R10 are new. They put the bipolars
into a zenlike state without the miller cap's
phase shift. And the gate currents now look
like clean sines through the full audio range.
High order artifacts in the output FFT have
been quieted down considerably.

This added control also makes it possible to
smoothly shut off either gate without the
transistor overreacting. Before, they would
latch into deep saturation. Overdraining the
gates, making them harder to recharge later.

To show what I'm talking about:
I've deliberately biased this one into Class AB.
Sim this one, how smooth the crossover is!
Would have made a total mess before...

The portion of Zen voltage drop on the base
resistors also contributes to the 1.3V base to
base voltage gap. With some further tweaking,
the Schottky drops might not be necessary.

This is not optimized, I don't even know if the
MOSFETs specified in the sim are power types...