Decoupling...

[Nick]

Its a good question isn't it

[IslandPink]

Well first of all, there's no DC energy dissipated in the speaker (!) , but a fair bit in the pair of output valves.
It's about what happens in AC. If I start with a 300B class A output stage, above about 20Hz, the output transformer primary looks like a 6k resistor with a centre tap where the B+ comes in.
If there's a big signal swing going into the 300B grids, the current divides differently in the two halves of the transformer priamry, but the total coming in from the power supply is about the same unless you get close to the limit.

[Nick]

Well first of all, there's no DC energy dissipated in the speaker (!) , but a fair bit in the pair of output valves.

Yes, of course its all about AC, The point is work is done by the speaker motor to move air, energy is contained in the sound waves and some heat is generated. So, that energy has to come from somewhere. Now you may argue that its not very much energy, and I would counter, no its not that much, but its the bit of the system that we are interested in (the only bit) as its the bit we hear.

the output transformer primary looks like a 6k resistor with a center tap where the B+ comes in.

I would argue not even close to a 6k resistor, remember the two half's are not just two separate chokes, they share a common core. But if the total current was constant, why would a current be induced into the secondary?

If there's a big signal swing going into the 300B grids, the current divides differently in the two halves of the transformer primary, but the total coming in from the power supply is about the same unless you get close to the limit.

Well, if the valves were perfectly linear then yes, but they are not, some more so than others, You can see the effect of the lack of linearity in the various tricks WE found to try and cancel the differences out on the cathode.

BTW, I am not trying to be clever here, I am asking the questions from the perspective of realizing how little I understand about whats actually happening.

[Nick]

But if the total current was constant, why would a current be induced into the secondary?

Because one is out of phase, ignore that stupidity.

[IslandPink]

BTW, I am not trying to be clever here, I am asking the questions from the perspective of realizing how little I understand about whats actually happening.

Absolutely.
So one thing we do know is that the only way power gets into the speaker is through magnetic potential energy in the core being converted into current in the secondary.
I'm less clear about how it gets into the core from the primary. At reasonable frequencies, the inductance of the primary & core seems like a big resistance. But is that just saying the 'delta I' in the primary is very efficiently converted into magnetisation ?

[ed]

love this guys drawings

http://spiff.rit.edu/classes/phys213/le ... sform.html

[Mike H]

[IslandPink]

Oh, Mike - I thought you were going to solve the puzzle for us ?

[pre65]

Oh, Mike - I thought you were going to solve the puzzle for us ?

He's probably giving it some deep thought first.

[IslandPink]

But if the total current was constant, why would a current be induced into the secondary?

Well, if the valves were perfectly linear then yes, but they are not, some more so than others, You can see the effect of the lack of linearity in the various tricks WE found to try and cancel the differences out on the cathode.

I would have thought that with a good DHT, the current change would be very near equal and opposite in the two valves at least up to 25% or 30% of full power - but that's thinking in DC , maybe once your are in AC does the V vs I relationship become complex on each side, and some energy goes into the core ?

[Nick]

I guess so, but also remember the primaries have the reflected load of the loudspeaker in parallel with them, so certainly complex in both senses,

[IslandPink]

Indeed, but not a necessary part of an argument about how the power gets into a loudspeaker - because a class A PP amp is quite capable of driving audio power into an 8 ohm resistor !

[Nick]

Certainly, I was not trying to suggest some magic only in complex loads situation.

Though if you have a source of 8R resistors with zero inductance and capacitance, I would be impressed

[izzy wizzy]

Foot in mouth time. I think I see what is being said. Say using my amp as an example, at idle the HT is 1000V @ 180mA = 180W. At say a power out of 30W the HT delivers the 180W plus 30W so the HT bounces around to the dynamics of the music not so much to the frequency changes but a little as the load is non-linear as are the valves and all other things coz nothing is perfect. Class A amps do increase the HT current draw depending on power out. Prolly explains why low Z rectifiers have an effect esp on dynamics and bass, IME anyways.

As the HT bounces around to the dynamics not so much frequency variation, then maybe caps in the HT are not so "critical" as say a SE amp where usually the AC signal is coupled to the OPT via the HT cap.

And so the idea is a seperate front end power supply is still advantagous coz HT still bouncy.

I'll be interested to try this but I just stuck my 10H chokes in the bias supply ... hmmm ...

Cheers,
Stephen

[Nick]

Yes, that is sort of my feelings, would be easy to measure with a 10R resistor in the 0v line and a scope (and a push pull valve amp). I only have two of those three. It suggests that the supply would see an average of the power output. What would be interesting to see what the time constant on that average was and what the effect of a no signal to 30W signal transition would look like from the power supplies POV. A hall effect current sensor would be a good way of measuring it as well.