Jump to content
IGNORED

Is it possible for a power supply to affect sound quality?


Recommended Posts

I see a lot of talk, and a lot of expensive options on how this power supply is magical and this one even more so and that confuses me.

 

I'm not talking about snake oil, about taking a picture of your stereo and putting it in the freezer, or buying a nice little clock that will make everything sound better.  We all know how well those things work.

 

I'm looking at it more from a technical perspective - a basic I/O equation - money in, better sound out. 

 

When I grew up, a linear power supply was a full diode bridge with a zener for voltage regulation and maybe a choke/capacitor to smooth things out a bit.  It didn't really matter what the supply voltage was doing, it all came out as regulated DC anyways.    

 

Switching supplies?  Sure, the potential for a lot of noise that isn't easily mitigated but gives you a small package size.  Let's not get into that and stick with the linear side of things.

 

Am I missing something?  For those of you here far better informed than I will ever be, I would appreciate your input.  

 

Can a linear supply actually make any difference?  And if it did, then just use a gel-cell battery and charge it with a DC power supply set to the value specified by the battery manufacturer?

 

Confused,

 

Stephen

 

 

Link to comment

Yes, the power supply is very, very, very important ... why? Well, the theory is that the PS is perfect, and that the circuit it feeds perfectly rejects any electrical anomalies that just might pass through the parts of that supply. And of course both of those theories are a very long way from the real world ...

 

Designers like to think of designs being the "real" circuit, which just hang off some ideal power supplies - but voltages and currents are not so forgiving ... Murphy's Law does apply, and the behaviour of the PS in every likelihood will impact how well the circuitry being fed performs ... as far as wires and electrical parts are concerned, everything is just one, spread out, circuit, going right back into the wall socket that you plug the gear into.

Link to comment
23 minutes ago, SJK said:

Hold the phone, dude.  I'm talking about this from a technical perspective, not how you feel about it. 

 

And so am I. If one uses simulation software, typically Spice, to model how real world PSes work, and how the circuits they connect to then behave, as I have done, then one discovers the far more complex interactions that occur. To give an example, a reasonable audio amplifier design generates a THD figure 100 times worse when working from real world supplies, as compared to ideal voltages being present. Most thinking in the audio world makes lots of assumptions that things work better than they actually do - it takes some effort to better analyse what's really going on - which is why it's usually not done; it happens to be an interest of mine to delve further.

 

Quote

 

Murphy's Law?  What does that have to do with the price of tea and the output regulation of linear power supplies?

 

That one can't just assume, "anything" - getting some brilliant numbers of output regulation for some PS design is highly likely to be only part of the story - in the real world, other factors may figure; pretending that they can't be relevant is not helpful for best understanding.

Link to comment
1 hour ago, SJK said:

Can a linear supply actually make any difference?  And if it did, then just use a gel-cell battery and charge it with a DC power supply set to the value specified by the battery manufacturer?

 Yes it can, however it is also possible to clean up SMPS sourced power as in PC's with additional LOW noise voltage regulation.

The main thing is to get very low noise combined with a FLAT output impedance from almost DC to 1MHZ.Unfortunately, many low noise voltage regulators such as the LT3045 when used as per their Data sheets have a considerably lower output impedance at >100kHz, which can exaggerate HF detail, and may result in a cold and clinical sound.

 You could use a Gel or 12V Lithium Ion battery, but don't connect it to the charger when playing, as this will lead to additional capacitance to A.C. Mains Earth and a resulting higher noise output..

 

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 13-11-2020

Link to comment
9 hours ago, One and a half said:

An SMPS conducts noise 100kHz upwards, which capacitively couples to the power supply and the network within an amplifier or DAC, doesn't matter. Floating heatsinks 'conduct' noise to ground planes, depending on the atmosphere, can change and is very difficult to control. SMPS only advantage is its small physical size, for audio SMPS is a nuisance, don't use them.

 

Battery. Well, you can use them, their internal resistance changes, so their source impedance will also change, which may affect noise paths. Batteries also need to be regulated and charged, so I find 

 

So a linear power supply is quiet electrically, so why not use them. Rectifiers can cause noise, mitigated by Schottky diodes, or just plain caps across them. Regulation is nice but not essential to be within 0.0001%, typically +10-15% like most mains rated equipment. Unless probably in some DACs the clock circuit is critical, but that's a DAC designer's issue, not ours and cannot depend on a customer's supplied unit (see Lumin thread on 3rd party power supplies).

 

What's important for a linear supply is the recovery caused by rectifier pulses and sudden load changes, The voltage shouldn't sag and swell over milliseconds, this leads to instability and load malfunctions. Some linear power supplies, like the Uptone JS-2  and the Acopian have sensing terminals that monitor and adjust the output voltage due to long lead length, I use the Acopian Gold Box at 9V sensing cable to an ifi Micro USB3, works well over a 3m cable length.

 

I've included a paper from Acopian which discusses SMPS and Linear supplies as well as how noise is created not only from SMPS but circuit elements, transistors, CPU, MOSFETS, gates,  where the noise comes from in the first place. 

 

 

An Overview of Chip Level EMC Problems - IEEE Santa Clara Valley ...032010Radu.pdf 4.54 MB · 2 downloads Trouble_with_wallwarts.pdf 27.8 kB · 6 downloads Wurth Electronic Product_Training_CMC_100728.pdf 717.37 kB · 2 downloads Linear-Vs-Switching-Power-Supplies-Whitepaper.pdf 363.97 kB · 7 downloads

Thanks for that, I'll give it a read and take it from there.  My experience is with power supplies for industrial equipment where there wasn't much concern about how it sounded.  :)

 

Link to comment

Well designed equipment should not be sensitive to external AC to DC supply's. But when we talk about audiophile equipment, it's a different ballgame.

But any differences will show up in the measured Signal-to-Noise ratio at the component's audio output.

This needs to be a wide band measurement, best done with a battery powered meter. Also an O-scope check for high frequency noise.

Link to comment
18 hours ago, fas42 said:

 

And so am I. If one uses simulation software, typically Spice, to model how real world PSes work, and how the circuits they connect to then behave, as I have done, then one discovers the far more complex interactions that occur. To give an example, a reasonable audio amplifier design generates a THD figure 100 times worse when working from real world supplies, as compared to ideal voltages being present. Most thinking in the audio world makes lots of assumptions that things work better than they actually do - it takes some effort to better analyse what's really going on - which is why it's usually not done; it happens to be an interest of mine to delve further.

 

 

That one can't just assume, "anything" - getting some brilliant numbers of output regulation for some PS design is highly likely to be only part of the story - in the real world, other factors may figure; pretending that they can't be relevant is not helpful for best understanding.

SPICE is a great tool, but my question is are we really overthinking this whole thing?  

 

A full wave bridge, basic filter network and you get a DC power supply with a bit of ripple.  No electronics, no chips, old school power.  It works, and it works well.  All of the comments to my question are with a solution that involves some type of custom designed circuit that in a complex way does the same thing.  

 

And, the question still stands.  That old school rectifier/filter with a bit of ripple, does it sound better or worse than a boutique solution?

 

Sounds like we need to have a showdown, pardner.  

media_22d_22d1936b-f0d7-4801-8efb-fd09f8aaad98_phpElbXtk.png

Link to comment
2 hours ago, SJK said:

SPICE is a great tool, but my question is are we really overthinking this whole thing?  

 

A full wave bridge, basic filter network and you get a DC power supply with a bit of ripple.  No electronics, no chips, old school power.  It works, and it works well.  All of the comments to my question are with a solution that involves some type of custom designed circuit that in a complex way does the same thing.  

 

And, the question still stands.  That old school rectifier/filter with a bit of ripple, does it sound better or worse than a boutique solution?

 

Sounds like we need to have a showdown, pardner.  

media_22d_22d1936b-f0d7-4801-8efb-fd09f8aaad98_phpElbXtk.png

 

 That old school power supply should only be the front end of a low noise PSU using very low noise voltage regulation.

 It doesn't even come remotely close to good enough by itself in today's  high performance DACs, Preamps etc..

 It has poor voltage regulation under load, accompanied by high levels of ripple at higher current draws .. It still isn't good enough these days in sensitive areas ,even when followed by the old workhorse LM317 voltage regulator

which can reduce lower frequency ripple by about  75dB at 120Hz. with a 10uF capacitor bypassing it's Adjust terminal but isn't very effective with even HF SMPS ripple..

 

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 13-11-2020

Link to comment
55 minutes ago, sandyk said:

 

 That old school power supply should only be the front end of a low noise PSU using very low noise voltage regulation.

 It doesn't even come remotely close to good enough by itself in today's  high performance DACs, Preamps etc..

 It has poor voltage regulation under load, accompanied by high levels of ripple at higher current draws .. It still isn't good enough these days in sensitive areas ,even when followed by the old workhorse LM317 voltage regulator

which can reduce lower frequency ripple by about  75dB at 120Hz. with a 10uF capacitor bypassing it's Adjust terminal but isn't very effective with even HF SMPS ripple..

On a scope, yes. Still waiting to hear how it makes my stereo sound better or worse. 
 

Remember, this is within the context of an external linear PSU, not something that is part of an integral design.
 

i apologize for not making that point earlier. .  

Link to comment
4 hours ago, SJK said:

SPICE is a great tool, but my question is are we really overthinking this whole thing?  

 

Depends in part on what it's feeding - if the the current draw is relatively steady, and only only low currents are required, then one can be less fussy. Once we get into the arena of power amps, then it's gets messy - old school power supplies are quite primitive in their ability to maintain the nominal DC voltage; it's quite easy to simulate almost any amount of voltage sag, noise you care to nominate as being possible, if real world circuits and parts are used.

 

 

Link to comment
42 minutes ago, SJK said:

On a scope, yes. Still waiting to hear how it makes my stereo sound better or worse. 
 

Remember, this is within the context of an external linear PSU, not something that is part of an integral design.
 

i apologize for not making that point earlier. .  

 

 It depends entirely on what the external Linear PSU is being used to power.

Even the type of filter capacitors used in this area can still affect the resulting sound AFTER going through a voltage regulator.

e.g. A low ESR Panasonic FC vs. an Elna for Audio of the same value and voltage ratings.

 Surprisingly, this even applies to items like a USB Regen which has it's own internal low noise voltage regulation ! 

 

 

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 13-11-2020

Link to comment
48 minutes ago, SJK said:

think you're talking through your hat. 

 Try telling that to Alex C , E.E. John Swenson and many others, especially the majority of the posters in this thread

 .https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/#comments

 I did exactly what you demanded recently for a well qualified participant in this thread. via a PM.

 In this case it was powering a USB memory stick instead of using the PC's crappy SMPS +5V USB power.

https://audiophilestyle.com/forums/topic/54909-usb-audio-transmission-isn’t-bit-true/page/18/#comments

 

I would offer you some examples, but given an attitude like yours I doubt that your equipment is likely to be good enough to either see or hear the differences anyway.

 

How a Digital Audio file sounds, or a Digital Video file looks, is governed to a large extent by the Power Supply area. All that Identical Checksums gives is the possibility of REGENERATING the file to close to that of the original file.

PROFILE UPDATED 13-11-2020

Link to comment
8 minutes ago, Superdad said:

 

Actually, very few large power amps have regulated linear power supplies.  The amplifier chassis would need significant additional heat sink area just for the the pass transistors required to go with the regulators.

That’s not to say that state of the art power amps all just have simplistic transformer>diode>capacitor supplies.  Some are quite sophisticated, and if you ask any serious designer they will tell you that the power supply is crucial to the performance of their amps.

 

Yes, I did an exercise some years ago, DIY chip amps, LM3886s; and was able to do regulated supplies for them - turned out there was a sweet spot for the regulator parts, and heatsinks, which allowed this to be done for a very reasonable cost. Was happy with how they performed, but if I wanted more power then it would have got quite involved.

 

A design that appeals to me is what the Magtech units, by Sanders, do with regulation - heard them perform very capably at the last audio show.

Link to comment
On 6/19/2020 at 4:07 AM, One and a half said:

An SMPS conducts noise 100kHz upwards, which capacitively couples to the power supply and the network within an amplifier or DAC, doesn't matter. Floating heatsinks 'conduct' noise to ground planes, depending on the atmosphere, can change and is very difficult to control. SMPS only advantage is its small physical size, for audio SMPS is a nuisance, don't use them.

 

Battery. Well, you can use them, their internal resistance changes, so their source impedance will also change, which may affect noise paths. Batteries also need to be regulated and charged, so I find 

 

So a linear power supply is quiet electrically, so why not use them. Rectifiers can cause noise, mitigated by Schottky diodes, or just plain caps across them. Regulation is nice but not essential to be within 0.0001%, typically +10-15% like most mains rated equipment. Unless probably in some DACs the clock circuit is critical, but that's a DAC designer's issue, not ours and cannot depend on a customer's supplied unit (see Lumin thread on 3rd party power supplies).

 

What's important for a linear supply is the recovery caused by rectifier pulses and sudden load changes, The voltage shouldn't sag and swell over milliseconds, this leads to instability and load malfunctions. Some linear power supplies, like the Uptone JS-2  and the Acopian have sensing terminals that monitor and adjust the output voltage due to long lead length, I use the Acopian Gold Box at 9V sensing cable to an ifi Micro USB3, works well over a 3m cable length.

 

I've included a paper from Acopian which discusses SMPS and Linear supplies as well as how noise is created not only from SMPS but circuit elements, transistors, CPU, MOSFETS, gates,  where the noise comes from in the first place. 

 

 

An Overview of Chip Level EMC Problems - IEEE Santa Clara Valley ...032010Radu.pdf 4.54 MB · 6 downloads Trouble_with_wallwarts.pdf 27.8 kB · 10 downloads Wurth Electronic Product_Training_CMC_100728.pdf 717.37 kB · 6 downloads Linear-Vs-Switching-Power-Supplies-Whitepaper.pdf 363.97 kB · 11 downloads

 

This is not correct IMO. SMPS's advantage are its small size, low weight, high efficiency, low heat radiation and most important low cost per watt.

 

It’s not that simple that a linear power supply is always quiet electrically. Many LPS are sensitive to DC offset and both hum and radiate a lot of EMI.

 

Am not saying that SMPS is better sonically, but a quality SMPS is normally better than a low quality LPS for the same watts and price.

 

To answer the OP question, can a linear supply actually make any difference? Yes it can, but that it is linear is no guaranty. If it was, no respectable hifi manufacturer would use SMPS.

 

Link to comment

Not to be pedantic but my original post and subsequent ones clearly states the focus was on external linear power supplies, not power supplies in general as relates to high end equipment.  

 

My point was, dare I bore you with repetition, that a standard old school full wave bridge rectifier with a basic filter network is likely just as capable as any boutique linear power supply available.  

 

What say you to that?  You can be abusive if you like, or you can stay on topic.  I'm fine with it either way.

Link to comment
6 hours ago, davide256 said:

I can't offer you explanations, only the observation that in my gear improving DC power supply has been crucial for improved resolution of lower db level signals

and elimination of USB digital irritants. And that the cost /benefit equation (~$500 per power upgraded component) has been far cheaper

than any next equipment upgrade

OK, now we're on the right track.  Someone who has made a change in an external power supply and believes that it has made a difference in the sound of the the system.  

 

I find it interesting on how people have a hard time staying on topic with this, and insist on wandering off mumbling about how this and that obviously has to make a difference.  

 

Davide, perhaps you could give us a bit more information on how the change in power supply made a difference to your system.  It doesn't have to be technical, we seem to have technical depth on the design side well covered.  

 

Remember, this is and Audiophile forum, not an Electrical Engineering forum.

Link to comment
19 hours ago, k27R said:

I think you’ have this in the wrong forum then.  Posting in the objective sub-forum will probably steer you closer to what you’re interested in.

Being interested in fact vs. fiction is something we should all be able to relate to.  

 

My challenge stands.  Use a standard full wave bridge like kids use to breadboard in school vs. that fancy linear power supply that cost hundreds of times more - and then explain in terms that all of us understand and insist should be there (if it's different, it's measurable!) exactly how one compares to the other.

 

And, I reiterate, my focus is on external swappable power supplies only, not on what might be found in high end equipment where the power supply is much more than a single regulated output.

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...