Separating R and L channel in input stage of a power supply

Hi folks. I would appreciate you to check my logic before I buy the parts.

I have a CLCRCRC power supply for my Luxkit A3600 amplifier. The first amplification stage (the last RC in the power supply) has a 100Kohm plate resistor and needs about 15uf capacitor to provide the low ripple I'm after. To achieve it, I have a 3x 15uf multi-section capacitor. I will employ one section for the driver stage, which leaves me two sections for the input stage - an opportunity to split the supply for the right and left channels.

I have some questions about how to do this as I would need two:
  1. dropping resistors - I go from one 100Kohm plate resistor to two 200Kohm resistors to keep the plate voltage the same?
  2. capacitors - I presume going from 1x 15uf to 2x 15uf in parallel means the ripple rejection remains the same?
  3. bleeding resistors - I go from one 330Kohm resistor to two 660Kohm resistors to keep the voltage divider the same?
 
Yes, provided low frequency stability is properly considered and is not degraded by subsonic response peaks.
In the circuit model, for identical L and R inputs, the L and R supply nodes are the same at both DC and AC.

So no difference should exist if the two nodes are connected or not, until the L and R input signals differ.
Which is the benefit of separating the supply nodes for stereo.
 
Thanks, R! So, I'm not really sure if low frequency stability will become a problem if I split the supply nodes. Might it mean changes to the global NFB circuit?

I just had a thought - the first two stages (input/preamp and phase inverter) are directly coupled - this wouldn't negate the benefit of splitting only the first stage, would it?
 
No, but it may lessen the chance of subsonic instability that can be caused by high decoupling resistor values.
The input stage is seldom run from the same supply node as the phase splitter, to reduce hum by extra filtering.
 
Great - thanks! That's a bonus! Yes, I haven't seen many circuits like it. It creates another level of complication that I've enjoyed learning about. On top of that, I've had to get to grips with the asymmetric plate resistors and voltages for the two sides of each phase splitter tube (6GU7 or 6CG7) as they are configured as a long-tail pair without a separate tail resistor. I've found that having balanced tubes in both stages is very important.
 
Maybe the schematic should be posted here. There is a common tail resistor.

When you drive only one side of a long tail pair, the outputs are not balanced.
The higher plate resistor in the undriven tube is an attempt to balance the outputs.

The only way to get balanced outputs with single ended drive in this type of circuit,
but with matched plate resistors, is with a very high value tail resistor (and a large negative supply),
or else a high impedance current source in place of that resistor and supply.
 

Attachments

Thanks, R. Yes, I'm familiar with the problem. As you say, from what I've seen of other LTP circuits, to keep the plate resistors the same, there is normally a voltage divider between the cathode resistor and a large tail resistor with a pair of grid leak resistors at their junction.

In my case, the two plate resistors and shared cathode resistor are 27Kohm, 33Kohm, and 8.2Kohm, respectively, designed around a niche and long since out of production 6240G tube. I've experimented with the value of the cathode resistor since changing the tube to the 6CG7 and the 6GU7 tubes I'm using now.