[emtor]

For a low power single ended tube amp, would a choke input PSU be beneficial compared with a capacitor input PSU?

[FullRangeMan]

For a low power single ended tube amp, would a choke input PSU be beneficial compared with a capacitor input PSU?

I have never see a choke as input at a PSU, its more expensive than a cap. It will do the soft start task at the mains line ?

EDIT: Its not clear what are you referring to, if you refer to a inductive PSU as used by Aloia the inductive PSU is much more efficient than a capacitive PSU, the amp feels like a much more powereful amp at the speakers.

[Docere]

The original poster was referring, I think, to the two broad types of B+ power supplies that use chokes:

• Capacitor-input, aka pi filter aka CLC...
• Choke input aka critical inductance input aka LC...

There are variations on these themes: flywheel (lCLC), small input capacitor (cLC...) and so on, but they are beyond the scope of this post.

To answer the question, there is no absolute answer. Each approach has its advantages and disadvantages. For SET amplifiers, I like to design for low-enough ripple, fast clean (no ringing) recovery, and low impedance (relative to the tube's rp). That can be a challenge even before I add objectives for specific situations. I'm starting to lean towards capacitor input supplies with SiC Schottky diode rectification - let's see how my next build performs before I claim any real benefit. Critical inductance choke input supplies have their place too and can probably work very well.

Design matters more than the broad PS topology. All power supplies are a compromise and some are more compromised than others. 

[emtor]

I should clearify . . . what I mean is a L-C --- etc. filter after the rectifier instead of a C-L --- filter.
-In other words a choke input filter.
What you mention regarding a PSU with fast recovery and less ringing is interesting. -I would guess that L and C values (resonance) play a role here?

[Docere]

I should clearify . . . what I mean is a L-C --- etc. filter after the rectifier instead of a C-L --- filter.
-In other words a choke input filter.
What you mention regarding a PSU with fast recovery and less ringing is interesting. -I would guess that L and C values (resonance) play a role here?

Yes, the L and C values and DCR of the power supply elements etc.

It is also worth noting that the need for this is contentious. Some designers believe that SET amps have constant current draw, hence recovery and even low impedance is irrelevant (or at least not very important). Others believe that although the average current draw is constant, there are very short term dynamic current demands. Yet others believe that current draw is reasonably constant until the amp begins to generate a certain level of distortion. I'd like to be in a position to experiment and compare different amplifier power supplies.

[sfox7076]

Though more expensive, I tend to stick with LCLC power supplies.  Ripple is the enemy. 

Shawn

[emtor]

Though more expensive, I tend to stick with LCLC power supplies.  Ripple is the enemy. 

Shawn

It seems like people design amps and then the hook a power supply on to the amp as if the PSU is an external system.
But the fact is that the PSU is just as important as the amp itself. The amp is only modulating what is supplied by the PSU, so being sloppy with the PSU is a bad idea.
And yes,-ripple is the enemy and probably causes more sonic trouble than just being an irritating 100 or 120 Hz hum in the background.
I will build a tube amp sooner or later and I think I'll go for a LCLC filter.

[Docere]

Simplifications and heuristics are handy; they can also be misleading.

I have not seen anything convincing that aiming for lowest ripple at the significant cost of other objectives provides the best sounding (or performing) power supply. I also don't believe LCLC supplies always provide the lowest ripple or are inherently the best solution; lower than CLCLC? Thankfully SET amps can sound nice even with average supplies, though greatness is something else.

My advice is to just suck it and see. Learn - experiment - learn some more... just check your expectations.

[emtor]

Simplifications and heuristics are handy; they can also be misleading.

I have not seen anything convincing that aiming for lowest ripple at the significant cost of other objectives provides the best sounding (or performing) power supply. I also don't believe LCLC supplies always provide the lowest ripple or are inherently the best solution; lower than CLCLC? Thankfully SET amps can sound nice even with average supplies, though greatness is something else.

My advice is to just suck it and see. Learn - experiment - learn some more... just check your expectations.

In the art of engineering focusing on only one parameter is not a good thing.
When they discovered negative feedback in amps they used tons of it which introduced more problems than it solved.
Freedom from ripple? -Yes, but not at all costs.
I have never buildt a choke input PSU, but I do know that it will be kinder to a tube rectifier. Also chokes with low DC-resistance reduces ripple without throwing voltage regulation out the window. But,-the danger with LCLC filters is ringing and resonances if the values of L's and C's is not chosen correctly.
I have wound a few mains transformers and simple output transformers so the extra cost of chokes in the PSU is not a problem.
But, as you say, experimenting is always the way to go.