So would this work?

[lochness]

Bought a box (BIG BOX!) of assorted tubes at the flea market. Mostly random single (and some quite useless unless I'm building a tv...unless I build a 30,000-50,000v amp...) tubes. There were 6X 6AS7G, 20X 955, and 10X 957(most of them have their pins broken off... ) as well.
I've played around with the 6AS7G(s) a bit in cathode coupled output stages being driven by various tubes. Considering I was using the cheapest radioshack components I could find I was quite pleased.
Recently I read online that the Cathode Coupled amplifier can accept balanced input and still have an unbalanced output. Well seeing as I've never really done anything with phase splitting and push-pull (design wise at least).
What I was thinking of doing is using a Cathodyne (tube folks like putting -dyne at the end of everything for some reason) as the phase splitter and a differential amplifier for the gain stage (since as I understand it the Cathodyne has no gain). Having it all end with the cathode coupled 6AS7G.
The thing is that I really don't know much about phase splitters and push-pull amps and such. Would I be able to use any of my 955(s) or 957(s) for this (they're begging to be used for something)? If not what tubes would you suggest for each stage?

[opcom]

The 955 triode as a class A amp can do 135 miliwatts and is decently linear. It is also good beyond 50Mhz. It can also be used as a detector and oscillator.

Sockets are a little scarce but can be found cheaply with those who diddle with old military or ham gear. Usually always are ceramic. The 955 and similar tubes are know as "acorn" tubes.

955: indirectly heated triode with 6.3v heater
http://www.bunkerofdoom.com/tubes/tt3/index0067.html

957: a directly heated triode with a 1.25V filament.
see TDSL at duncanamps.com for more data on these tubes.

These are quite good tubes and were made to military specifications. there were no cheapie consumer versions made AFAIK, so you got something of good quality there.

[lochness]

Already knew that about the acorn valves. What I'm wondering is if they are suitable for an amp of this design.
For instance would the 955 serve as a good cathodyne? What makes a tube good for cathodyne service?

Also after reading some more stuff would it be better for me to put the gain stage before the phase splitter or after?

[Miles Prower]

Originally posted by lochness
What makes a tube good for cathodyne service?

Usually, a highish g(m) and current sourcing. That usually means medium-u types (6J5, 6SN7, 12BH7, 6C4, med-u half of 12DW7 and so forth).

Also after reading some more stuff would it be better for me to put the gain stage before the phase splitter or after?

Depends. What phase splitter topology are you considering? The original Williamson placed the phase splitter (a cathodyne) between two gain stages. This is probably the best since the biggest limitation with the cathodyne is output swing. If directly driving the finals, you also run into problems when a transient causes the grids of the finals to pull grid current. That causes a cathodyne to roll over and die.

If using an LTP splitter, you may be able to use it as a combination splitter and gain stage since it has voltage gain, and the cathodyne does not. For one design that used the 6SL7 as an LTP splitter, it needed a bit of up front help to improve the input sensitivity (a 6J5 gain stage with cathode degeneration to get a reasonable Av= 10, and to provide a gNFB summing node.)

In another design, cascoded 6BQ7s were used in an LTP. Cascoding the 6BQ7s made for huge gain. Even losing half that gain when making the LTP left sufficient gain margin so that another gain stage wasn't necessary.

[SY]

If directly driving the finals, you also run into problems when a transient causes the grids of the finals to pull grid current. That causes a cathodyne to roll over and die.

Indeed. Even the output stage transitioning from A to B will unbalance it. That's why an intermediate stage is a good idea, or for pentode output stages, the use of very large grid-stoppers to minimize the change in load when the Miller capacitance goes away from the tube in cut-off.

[lochness]

So I should have the gain stage before the splitter and use a high gm tube for the splitter. This helped out a lot. Time for me to go play!
...
...
...on my next day off...