Lowering noise in small signal pentodes?

[SY]

This circuit, presented at ETF2009, is claimed to overcome partition noise. I haven't measured it myself yet...

[vega65]

Sorry for the hand drawing.
I use this circuit in the g2 supply of a 13E1 Se.I was recommended this circuit by 7N7, who used it in the g2 supply of a pentode cathodefollower.
Initially I had S130P (pre ww2 vr tube) but it was not performing as well as the other vr tubes in the amplifier( driver tube is pentode with 0c3 in g2 supply, unbypassed cathoderesistor).
The amplifier delivers 32-34 w .
THD 1,2 % at onset of clipping.noise on output with 1 k termination on input, is 400µv.with the 400 hz filter connected around 100µV.
As a newbie many things is unclear.can anyone explain if the cap from collector to ground is needed when b+ is regulated ?

[leadbelly]

Originally Posted by vega65

Sorry for the hand drawing.
I use this circuit in the g2 supply of a 13E1 Se.

A 13E1 is about as far away from a small signal pentode as you can get; I think they are the size of glass Coke bottles

The resistor to ground is called the cathode bypass cap. It has nothing to do with PS smoothing, as in regulated vs unregulated B+.

Thanks all for the thoughtful and informative replies and keeping the OT somewhat down.

[leadbelly]

EDIT: Oops, you said collector to ground. Sorry.

[smoking-amp]

Miles:
"If the variation is greater, then a single series dropping resistor can introduce an undesirable "variable-u" feature that can increase distortion ..."

Moonbird:
"Miles - can you talk more about this variable-mu and how/why it causes distortion?? 2nd harmonic?? We geetah players like distortion after all "

Here are some curves for screen resistor effects:

Suppresor Grid used for Feedback?

[moonbird]

Originally Posted by SY

This circuit, presented at ETF2009, is claimed to overcome partition noise. I haven't measured it myself yet...

Very interesting SY. Could you possibly provide more info on the selection of the other component values for a particular pentode?? Thanks for the post.

[moonbird]

Originally Posted by smoking-amp

Miles:
Here are some curves for screen resistor effects:

Uhhh ... WOW whadda thread .... I think you should be awarded a research grant or something. THANKS for this link!!

[Wavebourn]

I once significantly reduced noise powering 6J32P filament from 5V voltage regulator.

[plovati]

Originally Posted by SY

This circuit, presented at ETF2009, is claimed to overcome partition noise. I haven't measured it myself yet...

What explanation had been provided for that claim?

[smoking-amp]

"Why should it do that?"

The idea is to return the split off current from the cathode stream (via the screen) back to the plate current. Hence no partition, no noise (ideally). Its more obvious in the Mosfet version I mentioned some years ago, since no current is lost thru the gate, versus some current loss occurs with the bipolar base connection. Don't forget a gate stopper on the Mosfet as well.
The bipolar device may have some advantage as far as lower Z at the screen compared to the Mosfet.

Also, the circuit will need further extensions for a power pentode, where the plate voltage can dip below the screen voltage (and then cutoff collector or drain operating voltage for the screen current bypass device).
This is readily remedied by using a cap from the collector/drain to the tube plate, with a resistor pull-up, or CCS, on the coll./drain (to B+) to keep the DC operating level above the plate.

Another clever approach is to provide a tap on the OT xfmr winding (or a resistor tap, if just resistor loaded) at around 80% of the winding from B+ center, to attach the drain or collector to. This keeps the collector/drain above the plate voltage when it dips below screen voltage.