long tailed pair - need more gain

[hidnplayr]

I'm building a push-pull amplifier with balanced inputs.
Right now the only gain stage is a long tailed pair (with current source)

It works, but I need more gain (and i'm already using high-mu triodes).
I could simply add another pair of triodes and a current source, but something tells me there is a better way. (involving less current sources and coupling capacitors)
I have been searching for such a schematic/topology/.. but havent found one.

Anybody knows such a thing?

[SY]

Originally Posted by hidnplayr

I'm building a push-pull amplifier with balanced inputs.
Right now the only gain stage is a long tailed pair (with current source)

It works, but I need more gain (and i'm already using high-mu triodes).
I could simply add another pair of triodes and a current source, but something tells me there is a better way. (involving less current sources and coupling capacitors)
I have been searching for such a schematic/topology/.. but havent found one.

Anybody knows such a thing?

Cascode LTP. Use high gm FETs on the bottom.

[jcx]

like he said; cascode - say isn't there some tube thingy with extra pins just so you can do that inside the bottle?

[Yvesm]

Yes, you may also try pentodes.

[Arnulf]

Originally Posted by hidnplayr

Anybody knows such a thing?

Paraphase inverter ?

[hidnplayr]

I need a lot of extra gain, pentodes wont cut it.
Cascode seems the way to go. I'll try to use some more triodes, as I dont have any jfets laying around..

Thanks guys!

[jcx]

if you are gm limited rather than mu then cascode won't help - very much the the same as using pentodes

a ss current mirror load could do the trick if you're not too doctrinaire about tubes - large degeneration, beta canceling, cascoding output Q should give way better than -100 dB distortion of the mirror stage itself so arguments about "adding transistor character" are pretty weak

bjt ccs load are likewize very linear for push-pull

[tomchr]

A cascode using a high-voltage MOS or BJT would work. If you have the headroom, you can rig a tube to do the same job. You should end up with a differential gain of 2*mu.

[smoking-amp]

Umm.. But the current mirror has low input Z, so the tube stage itself will be distorting without it's internal voltage feedback. Same deal for cascodes, but cathode degen. can fix mostly.

By the way, what is actually needed here, more Mu or more voltage swing capability?

[jcx]

very much the same as with cascodes or pentode - both "interrupt" the triode internal negative voltage feedback

current mirrors could also multiply output current, refer output to gnd

some circuit "complications" can improve stage linearity – even in no global feedback circuits - the "fewest devices" heuristic is naive