SS phase splitters?

[tinitus]

Whats the prob with fixed bias in an AB amp?

no problem I guess

 

[tinitus]

Whats the prob with fixed bias in an AB amp?

with your own words

But the problem: ....

I'm trying to understand what you are saying

you say we really need AB2 .... with fixed bias

but clearly still problematic

do you still insist on P devices ?
or would N devices make things easier(simpler) ?

 

[tinitus]

too difficult, I give up, sorry

cheers

 

[kenpeter]

Its clear asking "would N make things simpler" that
you and I are talking about totally different circuits.

If you ran the simulation before asking random stuff.
Only P makes this topology possible. I challenge you
to split and de-block with only 2 active devices of N.
Feel free to impress me by showing it can be done.

Ordinary N concertina with N followers is three parts.
And has no voltage gain (requiring one more stage),
so I'm already two devices ahead of the old game.

 

[tinitus]

ok, I shall not ask more stupid questions like 'how about pseudo triode copled`etc

don't let me stop you, please continue

A little feedback from collector Q2 cradles cap recharge much more gently.
Current in the Concertina is up to 45mA, and Q1 near smoking at 5W!!!
Current at plate is not exactly constant anymore, as base current draw
is becoming a significant portion of whats meant to leak through R4.

20Hz AB2 is now possible, not this result makes a whole lot of sense.
Still think the coupling caps should be bigger, but LTSpice won't even
start the simulation if I try. Something may be on the edge of no good.
I havn't checked for stability of the open loop and and all that junk yet...

 

[kenpeter]

As for sim freaking out on 22uF coupling caps, or any standard >10uF.
It calms down if I put a few ohms in series. Or move the existing grid
stoppers to the left, such that the deblocker is now on the tetrode side.

I still don't know exactly why. But I suspect the 1n914's capacitance is
series with the two coupling caps? Something gets weirded out by that...

Like to get the cap value down to something that ain't electrolytic.
Anything less than 10uF causes noticable distortions for 20Hz AB2...

 

[tinitus]

on the 'sidetopic' AB1 vs AB2 ... http://www.diyaudio.com/forums/tubes-valves/133034-6l6gc-ab2-amp.html

 

[kenpeter]

Thanks that thread. Was pointed out that triode might be
as good or better than UL, when you can drive G1 positive
and have a low driving impedance to overcome Miller.

I may plug that into the sim and see what happens...

 

[kenpeter]

Just tried that, and can't drive enough A2 to get the same power level.
Distortion is also much worse. The less A2 needed, the better. Having
seen the impact of Triode strapping, I'm inclined to try going the other
way and see what a Tetrode strapping has to offer.

 

[kenpeter]

That... Tried and clipped, would need a slightly different OPT impedance...
Screen currents at G2 at really high, and there's Volts there to make Watts.
10mA*20V = 0.2W peak is no threat to G1. 50mA*350V = 17.5W peak is
a huge threat to G2...

I'm not sure either of these experiments prove anything except that the
circuit would require a lot of re-tuning to discover if another strapping
has merit.

One other thing worth trying: might be to see if this new style de-blocker
could provide G2 grid currents in a screen drive configuration? Though I
sorta doubt we have enough voltage swing or current from a Concertina
to do so. Maybe simultaneous G1 + G2 drives with two de-blockers?

 

[kenpeter]

G1=G2 seems promising. But Concertina and coupling caps need to be much bigger.
Concertina is current clipping severely, driving only 27mA combined grids current.

The Concertina and caps have to pull up both grid currents without any help from
the de-blocker, which only serves to recharge caps during the cutoff half cycle.