Cathode Feedback?

[simonsez]

Can anyone explain to me or give me a link about cathode feedback in tube push pull amp? What the benefit? How to implementation?  I'm still very slow learning and everyday i have a lot of new question in my little brain. Many thanks.

 

[Kamel]

check this link for theory

http://frank.yueksel.org/other/RCA/Radiotron_Designers-Handbook_Fourth-Edition/07-Negative-Feedback.pdf

 

[simonsez]

Many thanks Kamel, that's a big one ! still downloading.... :

 

[simonsez]

Finished downloading, Yeah...don't fit in my little brain.. full of equation...Can you tell me the benefit of Cathode feedback connected to primary TI in push pull circuit like this? how to calculate the winding for CFB?

 

[PRR]

> new question in my little brain.

HOW "little" is your brain?

Do you understand NFB generally?

NFB stability?

Transformer phase shift?

If you have nearly-no understanding, it would take too long to teach it all.

It would be best if you say what you already understand. Teach it as far as you can take it. This allows lurkers with emptier brains to gain from your knowledge. And it allows people with more cluttered brains to take-up right where you fall off the rail, rather than go all the way back to baby-steps which you already mastered.

The idea is:

wide-band audio transformers have a top resonance. Leakage inductance and stray capacitance. Depending on details, this can give a significant resonance, say near 50KHz.

If the amplifier before it were "ideal", infinite gain and bandwidth, NFB could control the resonance.

Tubes have limited gain and bandwidth. Typically gain is falling near 50KHz. Phase shift of tube plus resonance makes the NFB unstable. You can increase gain with more stages, but this reduces bandwidth and increases phase-shift.

You often can not control 50KHz resonance with overall NFB.

However the Final stage in a power amp has good bandwidth, though not a lot of gain. (You can get Power or Gain, but you can't have both.) You can get a little local NFB, strictly between power tube and transformer, with Grid (G1), Screen (UL), or Cathode NFB. G1 NFB has bandwidth challenges. G2 NFB is non-obvious design and may have been variously Patented. Cathode NFB works the tube as Grounded Grid (great bandwidth), and can be designed by inspection.

> how to calculate the winding for CFB?

There's no right answer. Not even a good answer. But this will get you going.

Wind so that peak cathode voltage is similar to peak grid-cathode voltage. In the example you show, peak G-K voltage is probably 25V, so design cathode winding for 25V.

If the NFB signal to the cathode is any less, it hardly does anything. If any more, your driver can't make the increased voltage-swing with low distortion. The suggestion above requires 50V peak, which is a lot for a driver working at these B+ voltages.

Do that, and see how you like the compromise between top-ring control and excess drive requirement. You may be able to find a "no gross flaw" design. But finding a "really sweet" design gets into opinion, fashion, and wise choices, stuff which comes after many-many designs and builds.

 

[simonsez]

Thank you very much PRR. After about 17 years i'm not in touch with iron soldering (join in a band), now i'm falling in love again with DIY thing after accidentally visit this forum. My first project was when i was little kid build 100m band transmitter (transistor). Now i'm starting again with my childhood experiences. Yeah I know NFB generally, I read few article about it. But that CFB is really new to me. So I ask question in this forum, just like like i said....my brain is too little to understanding a lot of equation  . Well ..  I will read your expl. VERY SLOWLY. I hope can learn something from it. BTW I'm wondering, how can you be so smart, wise, and have a lot  sense of humor. Many thanks.

 

[PRR]

> that CFB is really new to me.

Surely you have seen a Cathode Follower?

You can put "load" anywhere in the plate-cathode loop.

Plate-loaded: all load is in plate. Drive voltage for typical audio pentodes is 6%-10% of supply voltage. Distortion is significant, damping is negligible.

Cathode Follower: with transformer coupling and for MAXimum output, drive voltage is 180% of supply voltage. Distortion is low, damping is high.

With a transformer, you can put part of the load in plate, part in cathode.

MacIntosh: 50% in plate, 50% in cathode. Drive voltage is 90% of supply voltage, distortion and damping are very good. OT performance can be so very good that a lot more overall NFB can be applied around the whole amp. But OH! still a real pain to drive. The Mac driver is a plumber's nightmare.

With 6%-10% of load voltage in cathode, drive voltage is 15%-20% of supply voltage, which is about the limit for simple drivers. Distortion and damping are not near as good as CF or Mac, but a better than a naked pentode plate.

Frankly, if you are just getting back into electronics, just note that this is an Advanced Trick. You can build superb amps without it. You can assume that most commercial amps using this trick know what they are doing. Don't assume that random plans found on the Web are "optimum": the old trick of using 0-4-16 ohm speaker winding as a center-tapped cathode winding tends to do little good and may strain the driver more than it is worth. Proper design considers the WHOLE amp holistically.

 

[simonsez]

Many thanks PRR. Yes I think it's too complicated to me, but i will experiment with it.That's what i like about tube thing, even i don't really know what i'm doing :-X.  I will consult with my local transformer winding service for the OT. Thanks.