Power Supply - Ability to change the sound

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
G

Guest

Guest
After reading this months Tape Op with the interview with the bloke from Requisite Audio
He was talking abotu offering 2 Power Supplies for his products a valve based one and a solid state based one.
With noticeable difference on the final audio if either were used.

Now I am a simple man who does need educated people like yourselves to guide me through this.
How on earth does DC differ between "good "regulated DC from a valve PSU and "bad" regulated DC from a solid state PSU?
 
Two obvious ways:

1) Different output impedances at different frequencies.

2) Different response to current transients (i.e., well-damped vs. poorly-damped).

Other possibilities:

Different grounding.

Different amounts of RFI radiated by solid-state vs tubed rectifiers.

Peace,
Paul
 
[quote author="pstamler"]1) Different output impedances at different frequencies.
[/quote]
What frequencies? (DC...) Pardon my ignorance if this is a dumb question.

2) Different response to current transients (i.e., well-damped vs. poorly-damped).
Are you talking about transients in the AC line? If they are both equally well-designed I doubt there would be much difference.

My first idea was the amount and speed of instantaneous current that could be delivered resulting from transients in the signal. I know this is a big deal with power amps, maybe also true for preamps, etc? I could be way off here...
 
Dont know if this is what he´s refering to but...

A typical (read cheap guitar amp) tube rectified PSU wont provide as low output impedance to the circuit as a typical SS-psu.
Therefore under sustained load it will "sag".
Wich means that the voltage will drop as the circuit tries to draw more current. This creates kind of a compression effect.

There are exeptions to this of course.
 
:thumb: :thumb: :thumb: :thumb: :thumb: :thumb: :thumb: :thumb: :thumb: :thumb: :thumb:

mystery_man.jpg
 
[quote author="unheardof"][quote author="pstamler"]1) Different output impedances at different frequencies.
[/quote]
What frequencies? (DC...) Pardon my ignorance if this is a dumb question.[/quote]

Not a dumb question, but I was talking in general terms. One power supply might have a higher impedance at 100Hz, another might have a higher impedance at 10kHz, etc.. You were asking how supplies might sound different -- that's one way.

2) Different response to current transients (i.e., well-damped vs. poorly-damped).
Are you talking about transients in the AC line? If they are both equally well-designed I doubt there would be much difference.

My first idea was the amount and speed of instantaneous current that could be delivered resulting from transients in the signal. I know this is a big deal with power amps, maybe also true for preamps, etc? I could be way off here...

You're on-target. I was thinking about transients in the signal that draw large jolts of current. Some power supplies will do a small dip at that moment, then recover smoothly, while others will swing up and down a couple of times before settling down. Again, that could cause audible differences. Or not, depending on the audio circuit.

Peace,
Paul
 
[quote author="Kit"]Dont know if this is what he´s refering to but...

.[/quote]
It's a LA2A inspiered limiter
 
I'm not particularly technical, but I tried two different rerservoir caps in the power supply of a Hi-Fi amp and the tone and character of the overall sound dramatically changed. It was night and day, and verified by others who tried the same caps in other power supplies.
 
the power supply is a part of the circuit and in the path.

something to remember the devices modulate the power supply

then you have the series and/or shunt reg

Series supplies tend to be one way most only supply current

Now in the 70s I read about a power amp that addressed balancing the current on a +- supply I believe it was a Cotter

What was done is one channel was sent an inverted signal to do a kind of load balance on the + and -

Think about it.

Think about current paths in board circuit layout
 
okay - I think I am starting to get this
Did I get this right - the PSU doesn't deliver the current - the circuit asks for the current.
Therefore the current, even via a regulator, could not be what was asked for straightaway - but may come slightly later and or in bits..
Sorry - just brain dumping here.. am I making sense?
 
In a nutshell the PSU does deliver current , but only when the amp circuit ask for it. So you see, how well the PSU meets the amp circuit?s current demands is a very critical issue as far as audio is concerned. It's a two-fold quest - we need an amp circuit that asks for current at just the right times (depending on our particular sonic needs) and just as importantly, a PSU that can deliver it in the same manner. Of course, we could go a step further and talk about designing the PSU to compensate for the amp circuit doing something we don't like. Using tube rectifiers for tube amp PSU sag (mentioned above) is a good example of that.
 
I had a tube mic and two power supplies for it, one had a solid state regulator the other had a tube regulator. The mic sounded like a different mic entirely when plugged into one or the other. I had only previously seen such a difference in tube amps and then only when pushed to the top %30 of the power, so I was pretty suprised to see that happen with a mic. As mentioned before, the caps in the supply you use will absolutely change the sound of the amp, I would think that if you went with say, an unregulated tube supply vs. a regulated solid state supply, in a limiter, youd have a circuit that would probably behave diffferently, let alone whatever tonal changes you get. Thats just my experience, sorry nothing too technical, but its not just marketing, theres a real noticable difference.

dave
 
is it possible to have a parallel psu having both valves and solidstate to get a better psu?

I assume it is possible updating a "insert midquality brand console" with a beefier psu, if I read more or less correct? Also seen a lot of SSL console's with a double psu, that's not only done so in case one fails the other takes over I guess? Never knew this makes so much difference in sound :shock:
 
So, it seems like what we're saying is that the kind of power supplies that affect sound are the ones that were designed around specific limitations in performance (like the tube rectifier).

It seems to me an ideal PSU would have instantaneous current and voltage would remain rock solid regardless of the load. However, the designer can take advantage of weaknesses in certain PSU designs that actually sound more pleasing, even if they are not as accurate. Same thing with colored tube/transformer preamps compared with the infamous "wire with gain" :roll: (I'm generalizing here). They might not be as accurate or as faithful to the original signal, but they color the sound in a pleasing way. I guess I never really thought about that relating to power supplies.
 
Never knew this makes so much difference in sound

Its all about delivering current to a load, the problem is that the load (audio amplifier) is a variable one. Now what happens when the PSU cant supply the current needed for a given power output and frequency?

If the PSU is "stiff" (low impedance) it will simply blow a fuse or worse, burn.

If the PSU is a high impedance source it will kinda colapse with the timeconstant of the filtercaps and varying circuit resistance as current increases. Voltage will fall.
This doesnt mean that you shouldnt fuse a soft PSU, you absolutely should.

The importance of the PSU feeding an audio amp should not be neglected. Fortunitly in this day and age of IC-regulators and cheap large value electrolytics, its not to hard to make a stiff PSU.
But i can imagine that it would have been a major pain in the early days of tube amp design....
 
[quote author="Kit"]
Never knew this makes so much difference in sound

Its all about delivering current to a load, the problem is that the load (audio amplifier) is a variable one. Now what happens when the PSU cant supply the current needed for a given power output and frequency?
[/quote]

I understand completely, so this also means that the same theory goes for any preamp, channelstrip or entire console, hence my previous question about doubling the psu on SSL-consoles. Therefore I assume, and would like confirmation or opposition, that you can make a better console out of a 'medium one' by making the psu more performant. I also think the speed at witch the psu can deliver (size of caps?) will support transients much better, litterally translate more transparant.
Could developping better, stronger, faster psu's upgrade soundquality more then say, changing opamps? Hypothetically speaking of course... I know it will depend on the design etc... but let's suppose for the sake of letting me trying to understand this matter, that all of that is in comparable range.
 
nowthink about what happens when you don't need anymore power shutting something off..... but this happening very fast and the load is inductive and the power supply is a series type.

if regulated the error amp speed and setting time counts ............

its more than caps to big a cap can hurt. It more about balance and understanding the interplay of parts. A bigger cap at the output need more surge current and then you need to almost like a PID controller.

Basic stuff you can find in them things called books
 
It's all about the stability of the rails.

Yes your amps can modulate the supply rails and throw junk onto them which couples to your audio via the active bits tugging the output high or low.

Lets take a DC potential and put some mid freq hash on it. Now everytime your output transistor turns "on" current flows through the transistor, which includes the hash. This is over simplified of course. Now take a very small opposite polarity signal and modulate it on your rail voltage. Now every part that has the normal polarity signal will also see the reverse polarity. Add together and your signal has just degraded slightly. Repeat, rinse, repeat and you get mud.
 
Two different areas, tube supplies and solid state.

Much written about both supply types.
Power supply includes reactive elements, so variation with frequency is a possibillity.

Solid state supplies can in many cases feature a 78/79 series regulator, which have been shown to have impedance vs current draw variations.

Jung did a ton of work in this area, as many here know.

Layout can be just as important as the circuit.
For example, errors can be introduced by improper lead lenghts on filter caps.
 
Here are a few tidbits on supply impedance:

http://vacuumbrain.com/The_Lab/Power_Supplies/supply_impedance_a.jpg
http://vacuumbrain.com/The_Lab/Power_Supplies/supply_impedance_b.jpg
http://vacuumbrain.com/The_Lab/Power_Supplies/supply_impedance_c.jpg
http://vacuumbrain.com/The_Lab/Power_Supplies/supply_impedance_d.jpg

and layout...

http://vacuumbrain.com/The_Lab/Power_Supplies/layout_a.jpg
http://vacuumbrain.com/The_Lab/Power_Supplies/layout_b.jpg
http://vacuumbrain.com/The_Lab/Power_Supplies/layout_c.jpg
http://vacuumbrain.com/The_Lab/Power_Supplies/layout_d.jpg

working on the focus problem. :oops:
 
Back
Top