Screen Drive Push Pull ideas

[tubelab.com]

While moving, it was funny to see the look on the guy's face who picked up the small box that I packed the transformer in. A bit of swearing followed.

Try stuffing a 35 pound transformer into one of the flat rate boxes from the post office and then take it to the post office. Place it on the scales for the clerk and watch their expression when they try to move it! The girl punched all of the buttons, the scale read 35 pounds, she printed my receipt, gave me the change, but didn't realize what 35 pounds really was until she tried to move it. After a few futile attempts, she had to fetch someone from the back to move the box. The transformer got to its destination in one piece.

I bought an even bigger one from Ebay. It came from California for $12. My delivery person refused to bring it to the door, I had to go out to his truck to get it.

[SpreadSpectrum]

One thing I'm concerned about is voltage sag.

That's not something to worry about. I think I am going to place an order for Maida boards some time in August, as soon as I run out of other parts to mount.

[whitelabrat]

I'm thinking the 6CG7/6FQ7 option may not be optimal? Rated to 300v. They are similar to the 6AS7, but I don't expect them to take the same beating. I'm thinking pair of 6S4 would be ideal. Maybe the 6N6Pi? Perhaps overkill, but I'm looking to build in a comfortable margin for reliability. Otherwise 7N7's are looking good too.

Any suggestions?

Good news on the Maida. I wish I could have been more help there.

[Bandersnatch]

Originally posted by tubelab.com


Try stuffing a 35 pound transformer into one of the flat rate boxes from the post office and then take it to the post office. Place it on the scales for the clerk and watch their expression when they try to move it! .

HA! I can beat that... I crammed a pair of UTC CG301W's into a $9 flat-rate box. Screwed them together on a frame and padded it with impreg'd cardboard. Standard USPS Priority shipping was calc'd at $71 by the clerk who was suprised by its mass.
cheers,
Douglas

[whitelabrat]

My understanding so far is that serious voltage swing is needed to effectively drive a screen drive output stage. I've been studying Jone's Crystal Palace, and he chooses a 6J5G (6SN7) where it can handle 450v. Ideally a tube with a low output impedance would be used to drive the output, so I would figure a 12B4 would be a good option with a lower plate resistance than a 6SN7, but a bit more input capacitance. Now with the published Universal Drive schematic we have Mosfets in the game so I would figure that low output resistance is less important?

It would also seem that you can swing twice the voltage if a transformer were used instead of a decoupling capacitor? I haven't found much detail on this yet, but is this real? Can I get away with a driver putting out 300Vpk-pk and get a theoretical 600Vpk-pk from a transformer? I would probably need four of them which may be costly?

My goal here is to get as close to 100w output as possible. One other concept is that more tubes in parallel should make for more output with less voltage. I got a bunch of the proposed output tubes (assuming they'll work) so I'm hoping to go PPP for that stage.

So here's the plan so far:
6J6's --> 6CG7 --> 2SK2700 --> PPP Sweeps tubes --> 2022 ohm OPT.

I dug up a quad of 12B4's from the box o' random tubes so they may be an option instead of the 6CG7. I may possibly go for a 7N7 as well, but I ain't got none of them, but that may change cause I need four more compactron sockets.

The power supply will be interesting. A lot of regulating will be going on.

[DougL]

It would also seem that you can swing twice the voltage if a transformer were used instead of a decoupling capacitor? I haven't found much detail on this yet, but is this real? Can I get away with a driver putting out 300Vpk-pk and get a theoretical 600Vpk-pk from a transformer? I would probably need four of them which may be costly?

A triode that has a resistor or Css plate load can swing about 2/3 B+. The same triode inductively loaded can swing about 4/3 B+.
There are advantages to using a center tapped choke rather than 2 chokes, because the current cancels in a center tap choke and gives higher inductance.
I am going to try an Edcor inter-stage transformer as a center tap choke.. Should be expensive.

HTH

Doug

[whitelabrat]

Doug,

What kind of ratio are you using? 1:1? The center tap makes sense if I'm going to split the signal into a balanced or push pull arrangement, but I don't see how it would work in an LTP or differential circuit. I can see replacing the decoupling caps with untapped transformers.

[ilimzn]

Originally posted by DougL

A triode that has a...Css plate load can swing about 2/3 B+.

A well designed semiconductor based CCS can work well with as little as a few V across it, so the actual swing into a resistive load of a triode with such CCS on top is bigger than one with a resistor on top. Depending on the tride in question, one can get the usual 2/3 B+ with far less distortion (or more output current) than with a resistor on top - in that cese it's more like 1/3 B+.
Chokes do approximate a CCS as far as load lines are concerned, except that there is bucking action so you can get swing over B+ - typically as much over as it goes under (more under clipping conditions, sometimes dangerously more, keyword flyback converter ).

[whitelabrat]

Originally posted by ilimzn


... it's more like 1/3 B+.

In any case the transformer option caught my attention. I think I need some serious B+ to get a lot of voltage swing to push the sweeps tubes.

I may have caught some foot-in-mouth disease and wrote too soon earlier. If I'm not mistaken the driver stage of the Universal puts out a signal that is already split for Push Pull, right? So perhaps a transformer can be used, but it would have to take that differential input and put out a differential output. That way I could possibly use a single push-pull interstage transformer rather than two SE transformers.

I thought this was interesting. The schematic kinda demonstrates what I'm thinking, but has an SE output instead of push-pull:

http://www.audiodesignguide.com/PF2007/PF2007a.html

This is interesting too: http://www.icl.co.jp/audio/english/RC20/RC20.htm

An Edcor XSM10K/10K for the job???

[SY]

I've been studying Jone's Crystal Palace, and he chooses a 6J5G (6SN7) where it can handle 450v.

The Beast might be a good place to start (unfortunately, it got pushed out of the 3rd edition, but I think the schematic has been posted here). A CT transformer is a neat idea, wish I'd had one when I built my screen drive amp. But my goodness, they're expensive!

Another alternative is to use a relatively low voltage tube on the bottom of a cascode and use a 1kV or more MOSFET on the top, although it still would require a pretty high B+.