More power in this

[felixx]

The power output -about 20W but could be increased without problems but what is the best compromise to achieve...let's say 35-40W?
I want some opinions in this:
http://www.audiodesignguide.com/my/pp3.html
What do you think?

[SY]

That will be complicated, and the complication will get you only a very small increase in perceived volume. The problem is that paralleling tubes (which you'd need to do to double the power) doubles the Miller capacitance. The transformer won't like that. It can be circumvented by interposing another stage to buffer the output stage grids, but that is not trivial.

My advice: don't sweat the extra 2-3dB if you're heppy with that amp. If you want more power, start with a fresh design.

[Svein_B]

Higher B+ ?

320V seems much lower than most EL34 / KT88 I have seen.

SveinB

[SY]

That's pretty normal for class A or rich AB operation. If you raise the B+ and readjust the cathode resistor, you might squeeze another watt or two out at the expense of linearity. The difference between 18 and 20 watts is zip.

Basically, 20W is about the most to expect from trioded EL34, and even that is pushing it. KT88 might give you 30, again, barely noticeable power increase but you'd have to change the output transformer.

[felixx]

What do you think about that?
4 x KT88 per channel
480v 80ma per tubes
LL1627/PP (2300ohm/8ohm)
same input circuit
Rk(ohm) = 150
Ck(uF) = 1000uF

or better

6 x EL34 per channel
470v 50ma per tubes
LL1627/PP (2300ohm / 8ohm)
same input circuit
Rk(ohm) = 150
Ck(uF) = 1000

What about Miller effect?

[SY]

What about Miller effect?

There will be plenty!

[jane]

Originally posted by SY
There will be plenty!

How much is plenty? About 100pF per tube?
The E182CC driver has a rp of ~2k at 20mA so I suppose it’s able to feed some Miller capacitance, even through a (step up) transformer.

Reflected primary capacitance at 1:2 step up and 4 x 100pF tubes:
Cprim = (Csec)*(2/1)^2 = (4*100pF)*4 = 1600pF
F-3dB = 1/(2pi C rp) = 1/(2pi * 1600p * 2k) = ~50kHz.

Jan E Veiset

[Shoog]

You are tending to forget that the interstage transformer is likely to have its own capacitance to bring to the mix which is likely to make your 50Khz estimate very optimistic. Interstagers are hard beasts to drive at the best of times and a significant element of the driver stages capacity will be taken up simple with dealing with that.

Also if you try to get things to operate a bit more into AB, the driver is going to start to crumble big time.

It strikes me that this design is optimised for the components and trying to extract more power from it will likely end in failure.

Shoog

[jane]

Of course the 50kHz is optimistic, and it does not reflect the real circuit (the LL1660 transformer has a bandwidth of ~35kHz etc). It just tells something about the contribution of the transformed Miller capacitance, seen from the driver’s point of view.

Jan E Veiset