CF PP design

[jane]

This is a PPCF design or more precisely a 50% cathode follower push-pull design.



The "50%" refers to the way the driver signal is connected to the cathode followers, between the grids and a 50% tapping at the primary side of the OPT. It is actually a variation over a unity couplet output stage.



Gain in the "50%CF" is ~2 and output resistance from this cathode follower is about 2/gm. A gain of 2 reduces the demand of voltage swing from the driver circuit to the half compared to a plain cathode follower.


There are used pentodes as output tubes and instead of using separate chokes and a capacitor connected to the cathode to make the screen grids track the cathodes there are an extra winding in the OPT to keep the screens follow the cathodes. The capacitor is still included but now just to help the transformer with the leakage inductance between the primary and screen windings.




If you do read Norwegian the whole circuit design is explained in this thread:
http://www.hifisentralen.no/forum/in...?topic=31206.0



Here is a summary of the performance:
All tests measured with KT88 biased at 65mA, Vak=460B, Vgb=310V:

o Power output 40W at onset of clipping.
o THD measured at 1dB below max power: 2HD: 0.4%, 3HD 0.5%, higher harmonics < 0.1%
o THD at 5Vrms@8R: 2HD: 0.15%, 3HD: 0.05%, higher harmonics: < 0.05%
o Bandwidth: 20Hz - 50kHz +/-1dB
o Output impedance: 0.89 ohm
o No global feedback

One of two mono blocs on the test bench:



Jan E Veiset

[Geek]

Neat little setup.

How does it sound compared to grounded cathode?

Cheers!

[jane]

I haven't had the time for serious listening yet, just listen to some small bookshelf speakers at the test bench. I have to make the other monoblock finished before I can sit back with a cold beer and enjoy the music..

Jan E Veiset

[SpreadSpectrum]

Jan,

Have you had a chance for some serious listening yet?

I'm considering building a zero GNF Unity-Coupled amp and was curious how this amp stacks up against other amps you have built.

[Jeb-D.]

That's a really cool concept.

I'm curious how it sounds too.

[jane]

I think it’s a good sounding push pull amplifier. The output impedance is low enough to give a tight and defined bass with normal speakers and the mid range and treble are smooth and warm.
The weakest part is probably the driver stage with ~10dB local feedback which may cause the amp to start sounding unnecessary harsh when pushed over the limit (We all do that sometimes?). The class A driver doesn’t really need this feedback and in my opinion the amp sounds better without this local feedback..

Jan E Veiset

[Wavebourn]

Nice amp, but what about cathodes to filaments? Looks dangerous.

[jane]

It has separate filament supply at cathode potential for each output tube.
You can see the two double (2x 6.3V+6.3V) filament transformers at the bottom of this photo:


Jan E.

[smoking-amp]

A roughly similar scheme is attached. 40% UL taps on the OT xfmr provide 28.5% CFB. Some advantage here in using more than half the primary by each tube for better coupling in class AB during switchoff. The nominally 2K Ohm P-P Hammond 1650T provides about 4K P-P "equivalent" in this configuration.

Could put caps between the cathode and screen of each tube also, less stress on T2 then.

The cathode windings on both sides of ground are both driven by each tube independently (nothing has to pass thru the xfmr to get both phases), so they can be used for high bandwidth feedback taps for a driver stage if wanted.

Don

[Jeb-D.]

Just think, If you did a CF with %60 g2 feedback, you would have an inverted ultra-linear.

Common cathode w/ %40 NFB ~ Cathode follower w/ %60 PFB