converting BAL to UNBAL using tubes and without transformers

[kenpeter]

You gotta be kiddin me.... This thing works fine on 48V?
Don't seem to tolerate much other simultaneous abuse,
but 48V supply definitely works. With no circuit changes
to accomodate the decrease.

Now I can't seem to load the ouput any lower than 4K7,
or it distorts pretty bad. And below 2K2 the sim hangs...
I can't verify the actual load impedance of half amplitude,
in this state of starvation, but its sure lower than 2K2.
It works very nice with loads above 4K7.

I had the DAC outputs swing 2V apiece, and no common
mode or power supply noise in this run.

--------------------------------------------------------

Works even better (at48V), replacing plate resistor with
a choke. I tried 4H 100ohms, driving 1K loads no probs.

[Michael Koster]

Very nice, Ken,

Adding the signals together in the totem pole is brilliant!
That plus balanced ripple gets you some kind of award...
Plate feedback and that 47K bridge makes it all work.

I keep getting smoke out my ears when I try to work out
the gain and balance equation. Iterate with spice?

The triple play would be adding balanced input impedance

Plate feedback is da stuff for circuits using high-gm low-rp tube
circuits.

Michael

[kenpeter]

The triple play is there. Both ends see 47K.

The difference node is near zero impedance.
Difference input sees 47K to a virtual ground.

Driven in true diffferential, the sum node is
very high impedance. Sum input driving into
a voltage node the same as itself. I added
a 47K appendix to ground on the left edge,
specifically to make the input Z's balance.
It also completes a circuit to bleed the grid.

DC loading might be a little different, but
as long as the DACs aren't bothered by it,
the tube stage don't mind.

[sunrise]

oh man ... i just wanted an idea for a nice little tube output stage ............

thnx for the effort........

[revintage]

Three caps in the signalpath, feedback, SRPP. Those ingredients says NoNo to me .

Kenpeterinstead of that overworked "wako2", I would stick to the KISS Diff-CF.

[sunrise]

revintage - tha is exactly what i had in my mind.... to use exactly that circuit in your previous post...........

[revintage]

Don´t use the BD140 though. Go for somtheing that can take a little more voltage.

[sunrise]

mje350

[revintage]

Good, classic choice!

[ErikdeBest]

Hi Revintage

Is that 330pF capacitor doing the same thing as the cap in your first schematic (the one with DN2540 and negative rail)?

The filter between the cathodes had to be added to adjust a 2dB lift from 100kHz that I can not explain.

And, if the input is referenced to ground, would it work better if connecting the cathode resistor to a negative supply, say 12V?