Power Supplies

[rollo]

As far as I am concerned, it is all about the sound. Different designs require different power supplies.

Cheers.

I find this comment represents my thoughts, too. For example, I have a pre that is battery powered, and the lack of noise plus product performance does it in spades for me.

The sound of what exactly. The circuit or power supply. What my original question was what would make a state of the art power supply. The design lets say is there. Would the trannie, chokes, regulation of voltage be any more important than the other to achieve ones goal.
     As an example would substituting a finer trannie be an example ? Adding  sophisticated parts to the design ?
     The reason for the post was that IMO we listen to the quality of the power supply more than the circuit design. every time a power supply was upgraded there was a big overall improvement.



charles

[_Scotty_]

The circuit design sets a hard upper limit on what performance is possible from the circuit. Reaching the point of diminishing returns power supply wise simply allows the circuit design to perform to its maximum potential.
 I make the distinction of separating the basic power supply from what I will term circuit support mechanisms. A couple of examples are things like sophisticated regulators and active current sources.
 My opinion here is based on personal experience,YMMV.
Scotty

[steve]

As far as I am concerned, it is all about the sound. Different designs require different power supplies.

Cheers.

I find this comment represents my thoughts, too. For example, I have a pre that is battery powered, and the lack of noise plus product performance does it in spades for me.

The sound of what exactly. The circuit or power supply. What my original question was what would make a state of the art power supply. The design lets say is there. Would the trannie, chokes, regulation of voltage be any more important than the other to achieve ones goal.
     As an example would substituting a finer trannie be an example ? Adding  sophisticated parts to the design ?
     The reason for the post was that IMO we listen to the quality of the power supply more than the circuit design. every time a power supply was upgraded there was a big overall improvement.

charles

My comment was somewhat vague, because the circuit design does dictate the supply design. Class A is different than Class AB1, AB2, B.

Thanks for adding more detail Rollo. I misunderstood.

I believe the power supply is much if not most of what we hear. Of course we hear parts in the circuit design, and even HD, IMD as well.

First, I would think that one should minimize the reactance of the reference supply. This would depend upon the value of RL and tube used, so no hard numbers.

Regulation is important. However, the design of such is very important for proper operation. Sometimes probably better not to even use one.

In a preamplifier, operation is class A, so average current is constant. We need to distinguish between one stage and multiple stages.

With one stage, the transformer is virtually a non player in a properly designed power supply. I know of one nearly retired gent on another forum who uses a 1kw power tranny in each channel, and 150,000uf of capacitance. From decades of experience what a waste
and inefficient, and obviously expensive.

A two stage is much more problematic. I would suggest two totally separate power supplies, including separate  power transformers, to eliminate frequency dependent feedback through the power supply itself.

The worst mistake I see are the same schematics/designs being used for all the different tubes. Doesn't work that way.


Cheers.

[BobM]

Hopefully the engineers will clarify what I'm going to say below, and correct my bad assumptions and mistakes. But as I understand it an amplifier of any type really is misnamed. It does not amplify a signal, it effectively recreates the signal from the power entering the amplifier.

The "amplifier" expects a perfect sine wave coming in and modifies it to match the incoming signal then adds gain to that recreated signal. So if you have several devices that amplify a signal, like a preamp and a power amp then you are recreating the signal twice, or maybe three times depending on the circuits, so any deviation from a perfect sine wave in your incoming power is picked up several times and those deviations are multiplied as the signal is recreated through multiple amplifiers.

Did I get the bones of this right guys or am I totally incorrect here?

[richidoo]

Sounds good to me Bob!

At first I was going to say that the input signal is not copied before amplified, but is used directly to control the active device which then modulates DC from the PS to create a new signal with voltage or current gain.

But then I got to thinkin... 

When a signal goes through any active device, like a transistor or tube, or a capacitor or a transformer (including magnetic cartridges) the electrical signal is actually destroyed and recreated again. These are all transducers, converting the electrical signal into another form of energy and then back into electrical again. They are all "analog" conversions with infinite resolution, but each has a flavor of distortion. Resistors aren't transducers, so they have the lowest distortion.

Back on topic... Most PS distortion is non-linear (unrelated to signal,) like noise, ripple, and sag. Non-linear distortion is more noticeable and more damaging to signal than linear distortion which is signal-related colors added as signal passes through analog transducer parts. As steve has pointed out in previous posts, power supply can also contribute linear distortion when there is stereo leakage through a shared power supply.

One other thing, Dave mentioned his preamp above, it is a passive optical resistor type which is particularly insensitive to power supply problems because technically the power supply is not in the signal path as it is with all active amps. But the LEDs do flicker with PS noise so quiet power is also important to optical resistors. In his case optima battery is a perfect power supply for that preamp. Low noise, low impedance, fast response.

[tmazz]

I had a professor in engineering school that explained it this way (He was actually talking about a single transistor, but the analogy holds for a whole amplifier as well).

Think of a transistor as water spigot. The power supply is the equivalent of the  city water supply and the input signal applied to the base is the equivalent of turning the knob on and off which simply lets more or less of the city water flow out the spout at a given moment in time  such that the water pressure from the city mains is made to vary in a way that mirrors the pattern that the knob is being turned. Looking at it this way it is easy to see that the output pressure will only mirror the input pattern to the extent that the city water pressure is constant since the variations in the output will be equal to the sum of  the change in restriction o the valve being opened and closed, plus any fluctuation in the input pressure. Likewise in our audio equipment any variations in the power supply voltages (or unfiltered noise on the line which will simply sit on top of the PS output) will simply get blended into the output signal as an added distortion.

While there are many complex reasons why a PS voltage can vary  the basic results of that variation are pretty much this simple.

PS design is an art and science onto itself, especially when the outputs need to be stable to the micro levels required for high end audio. Keep in mind that for most application voltage variations that would be deadly to a high end audio signal are tiny and well within the acceptable range of voltages that the device can operate under.

Designing a power supply for use in a high end audio system is no trivial matter and as a engineer my hat goes of to those who do it well.  

[Lizard_king]

I can convey based on my own experiences that superior power supplies do have that musical quality that bring you closer to the music.

In my DAC, I had the Analog Tube power supply switched given what I call a nice upgrade and the SQ has taken on a more musical presentation.   The Bass did firm up and the pace was faster, more liquid.


I had a custom shunt regulated power supply built for my M2tech EVO and that brought the SQ to a whole new level of musical enjoyment.

[steve]

I suppose one can think of it almost like a motor. Power should be controlled in perfect proportion to the modulation of the foot feed. 

X input relates to X rpm. 2X input relates to twice rpm, no distortion. The goal is perfect output in relation to the input control.

Hopefully that makes some sense.

Cheers.

 

[tmazz]

I suppose one can think of it almost like a motor. Power should be controlled in perfect proportion to the modulation of the foot feed. 

X input relates to X rpm. 2X input relates to twice rpm, no distortion. The goal is perfect output in relation to the input control.

Hopefully that makes some sense.

Cheers.

 

Just like computers - Garbage in, Garbage out

[Triode Pete]

I suppose one can think of it almost like a motor. Power should be controlled in perfect proportion to the modulation of the foot feed. 

X input relates to X rpm. 2X input relates to twice rpm, no distortion. The goal is perfect output in relation to the input control.

Hopefully that makes some sense.

Cheers.

 

Just like computers - Garbage in, Garbage out

or when voicing horns, "Garbage in, Garbage out"

[pmkap]

The circuit design sets a hard upper limit on what performance is possible from the circuit. Reaching the point of diminishing returns power supply wise simply allows the circuit design to perform to its maximum potential.
......

I'm unsure of what you actually mean. I'd restate it as the circuit determines the slope of the cost/benefit regarding improvements in the power supply, specifically the circuit's PSRR and CMRR.

One can implement a circuit with very poor PSRR, but if I use herculean efforts on the PS, the result can be quite excellent.
Alternatively, a base circuit with very high (throughout the audio range) PSRR can show substantial benefit with minimal efforts to improve the power supply -
Dead Quiet
http://www.tubecad.com/2011/12/blog0221.htm

FWIW,
Paul

[steve]

I did not know tube guys had to deal with popping and sizzling these days?

For those who are bass conscious, 30uf output cap is
-6db at 20hz with 300 ohm load.
If one wishes more bass, a large increase in output capacitance is in order. Otherwise no big deal.

Cheers.