How to correctly scale down a 100W tube amp to 50W

[frusciante89]

Hi everyone I have a question on how to correctly scale down the power an amp. I've built myself a clone of a Friedman BE100, which sounds great.
I want to create a smaller version, in combo style, but at 50W. It not necessarily for scaling down the volume of this beast (we all now 50W is not half volume of 100W, but more like 70-75%), it's more a cost and weight issue.
What's the correct way to scale down the power tube section to go from 100W to 50W? right now the transformer impedance is 1.7k, will this scale down as well?
What about the screen and grid resistors?
I'm attaching a schematic of the 100W version to give you an idea.

[Helmholtz]

the transformer impedance is 1.7k, will this scale down as well?
What about the screen and grid resistors?

The 50W OT will need to have a primary impedance of at least 3.4k. You can use up to 4.5k ( at least 4k is recommended if plate voltage is > 450V).
The 100W amp uses 2 paralled EL34 pairs, tube paralleling halves their output impedance.

Screen and grid resistors can stay the same.

[GainFreak]

I don't like how FB changes with every speaker impedance change. It's either some kind of "feature"or they just made a mistake. Friedman is a nice guy and makes good amps but he's known to ship some amps with wiring mistakes in the past.

[g1]

Aside from different PT and OT, 50 vs 100 watt versions are usually very similar.

[Chuck H]

Keeping the caveats mentioned in mind, differences in tone are noted between two bottle and four bottle amps.The only real difference that comes to mind might be that miller capacitance is greater with a four bottle amp. So, to make a two bottle amp more similar should be as simple as adding capacitance from plate to grid for the two bottle amp. I know this is considered heresy because it's the addition of a capacitor instead of the "natural tube tone". But that's the difference, other than watts.

[GainFreak]

Keeping the caveats mentioned in mind, differences in tone are noted between two bottle and four bottle amps. The only real difference that comes to mind might be that miller capacitance is greater with a four bottle amp.

I also prefer the sound of a 4 bottle amp. To me it sounds more "open", with better attack and dynamics than a 2 bottle one. This of course is subjective but maybe there's also some science behind that.

[Helmholtz]

The only real difference that comes to mind might be that miller capacitance is greater with a four bottle amp.

The Miller capacitance of power pentodes is so low that it can be ignored. E.g. 6L6, EL34 : 15pF typical.

[mhuss]

You could also use 4 x 6V6GT or 4 x EL84 to get 30-ish watts with four bottles.

[Helmholtz]

Some possible reasons why the 2 tube version of an amp might sound different from the 4 tube version (assuming that speakers and voltage levels at idle are the same):

1) Different OT. OTs influence sound by primary inductance (bass response), leakage inductance and primary self-capacitance (both influence HF response) and saturation headroom.
2) Different PT. Influences sag and sag induced dynamics.
3) Power supply filter caps. The 2 tube version should have lower value reservoir and screen node filter caps for the same sag compression response times at full output.as the 4 tube version.

When just pulling 2 tubes of a 4 tube amp (and correcting output impedance), the amp might sound too stiff, because of the the oversized PT and filter caps. Also operating voltages will be higher.

[Chuck H]

Some possible reasons why the 2 tube version of an amp might sound different from the 4 tube version (assuming that speakers and voltage levels at idle are the same):

1) Different OT. OTs influence sound by primary inductance (bass response), leakage inductance and primary self-capacitance (both influence HF response) and saturation headroom.
2) Different PT. Influences sag and sag induced dynamics.
3) Power supply filter caps. The 2 tube version should have lower value reservoir and screen node filter caps for the same sag compression response times at full output.as the 4 tube version.

When just pulling 2 tubes of a 4 tube amp (and correcting output impedance), the amp might sound too stiff, because of the the oversized PT and filter caps. Also operating voltages will be higher.

Yes. All that. Plus... Ever notice the grid load/bias feed resistor values are never adjusted between two tube and four tube amps (well, rarely anyway)? So this could be another difference too.

[Helmholtz]

[QUOTE=Chuck H;n907011]
Plus... Ever notice the grid load/bias feed resistor values are never adjusted between two tube and four tube amps (well, rarely anyway)? QUOTE]

Yes, but that's no difference, isn't it? It means that the PI load stays the same.

[mhuss]

Another reason is that even "matched sets" aren't really matched, particularly over a range of amplitude and overload conditions. The more tubes you have, the more complex this sublte distortion stew becomes.

[Chuck H]

[QUOTE=Helmholtz;n907057]

Plus... Ever notice the grid load/bias feed resistor values are never adjusted between two tube and four tube amps (well, rarely anyway)? QUOTE]

Yes, but that's no difference, isn't it? It means that the PI load stays the same.

Well I can't say why for sure. It seems to me that the tubes input impedance should be high enough that using the same load for the PI would be the better scenario. But I'm certain that I read here (can't remember where though) that IDEALLY you would half the grid load for parallel tubes.

[Helmholtz]

[QUOTE=Chuck H;n907140]

Well I can't say why for sure. It seems to me that the tubes input impedance should be high enough that using the same load for the PI would be the better scenario. But I'm certain that I read here (can't remember where though) that IDEALLY you would half the grid load for parallel tubes.

That comment probably referred to a different context: Power tubes have limiting (max) values for grid to cathode resistance to avoid grid leak problems with somewhat gassy tubes. As these values hold per tube, the max for 2 paralleled tubes should be half the specified value. For EL34s the limit is 500k, so 220k+bias circuit resistance should be low enough for 2 and 4 tube amps.

OTOH, these bias feed/grid leak resistors are the external loads for the PI. And changing them means different PI loadlines, different gain, different max PI output voltage and different PI distortion characteristics.

[Helmholtz]

However you describe it, altering that PI loading is going to affect the sound, right?
So the 50W vs 100W Marshall has some difference in sound because they do not change the grid leak/bias feed resistors.

I listed some possible reasons for sound differences between 4 tube and 2 tube versions above. Not changing the grid circuit/PI load could hardly be one of them. The power tubes don't load the PI as long as there is no grid current.

I don't understand your logic.