4 tube push pull question

[J M Fahey]

This is more out of curiosity for down the road......

On amps that use more than two tubes for a push-pull output, it looks like you just put the tubes and associated parts in parallel with each other. Do you have to change the supply voltage to the sets of tubes, as opposed to just two tubes?

No.
You halve the load impedance to reflect the doubled current capacity, leave everything else exactly as before.

You will get doubled power, but of course PT and rectifiers must be able to supply double current and OT must handle doubled power.

As of *sound*, amp will sound and behave exactly same as before.

[Helmholtz]

As Juan said, two tubes wired in parallel and operating from the same plate/screen voltages can deliver twice the load current as a single one. In other words they will produce the same RMS voltage across half the load resistance, which means doubled power. Consequently two paralled pairs can produce twice the power of a single pair, if the voltages stay the same.
But this works only if a suitable PT and OT are used and power losses do not more than double.

So the answer to the question is no. Doubling the output power does not require an increase of B+. But amp voltages are mostly measured/specified at idle. The comparison needs to be made at full output instead.
If the B+ of the 100W version sags by 100V at full output and the 50W version only by 50V, the 100W version will show a 50V higher B+ at idle, even though there is no difference at full power.

I think the reason why some old amps actually use higher voltages in the 100W versions (often ignoring the tube limits) might be that the designer could not achieve the 100W goal otherwise, using the 100W transformers available/affordable at the time.

[pdf64]

...I think the reason why some old amps actually use higher voltages in the 100W versions (often ignoring the tube limits) might be that the designer could not achieve the 100W goal otherwise, using the 100W transformers available/affordable at the time.

I suspect it may be often overlooked that the stiffness / sagginess at the g2 node and actual g2s is somewhat more significant to plate current than is plate voltage, eg as average/instantaneous Vdc at g2s sags with increasing signal, plate current will fall.
And for pentodes, in drawing a greater ratio of g2/plate current, this is even more pertinent.

[Helmholtz]

I suspect it may be often overlooked that the stiffness / sagginess at the g2 node and actual g2s is somewhat more significant to plate current than is plate voltage, eg as average/instantaneous Vdc at g2s sags with increasing signal, plate current will fall.
And for pentodes, in drawing a greater ratio of g2/plate current, this is even more pertinent.

Looking at 50W and 100W Marshalls both using 1k screen resistors (I am aware that early 50W version in a fly-by-night manner did not have any, while the the 100W versions definitely required them to avoid HF oscillation) it boils down to the DC resistance of the choke. Is it this what you are aiming at?

[pdf64]

Yes, that’s it, but also it’s worth drawing attention to this key difference (ie that from the perspective of the cathode, g1 and hence plate current, Vg2 is more important Vp) between the power tubes and preamp triodes that are used is regular guitar amps.
I acknowledge that the choke resistance of the 100 watt Marshalls Is often a lot lower than the 50 watters, so that doesn’t explain the high HT of the super lead.

[loudthud]

Some of those 100W Marshalls use the same uF for the reservoir cap as the 50W models. That means more sag at the screen node due to higher ripple at the OT node.

[Helmholtz]

Some of those 100W Marshalls use the same uF for the reservoir cap as the 50W models. That means more sag at the screen node due to higher ripple at the OT node.

Good point.
Early 100W Marshall used a total of 32µ of reservoir capacitance and 16µ at the screen node. And these were typically the ones with the higher voltages (around 500V@EL34s to over 560V@KT66s, at idle).
Often these caps get replaced with higher C value types even further increasing the voltages and the stress on the tubes. Plus it changes the sound/responsiveness of the amps.