Linear Power Supplies

What's the deal with linear power supplies? It seems like many people believe that if you want to get the best performance out of your dac or amp, a LPS is pretty much mandatory. My common sense tells me that as long as the stock wall wart included with your equipment is working properly, there shouldn't be any detectable difference. What do you guys think?

 

You need to take a look at the conversation surrounding the Eddie Current Studio. Purrin says that there are differences between power transformers, and I don't remember who it was, but someone was using self-made power supplies for the Regen with positive results.

 

Probably this link here has something to do with your question:
http://superuser.com/questions/912679/when-do-i-need-a-pure-sine-wave-ups

 

Linear power supplies are better because they are inefficient and bulky. How dare you even consider using anything else for your Class A spaceheater of an amp.

 

Some devices are sensitive to electrical noise, which is normally orders of magnitude higher from a switch mode power supply.

 

It depends on what needs to be powered, but usually audio equipment always benefits from clean power.
Simple linear psu is not necessarily better than quality switcher.
Super-regulator or shunt has always made improvement in my experience.

 

Audiophilia Nervosa... is my theory.

But not always. It seems that a lot of laptop power supplies can be electrically noisy to an audible extent.

Is my brain playing tricks with me, or do I remember rightly that, twenty-plus years ago, PC power supplies had chunky transformers in them? (and they blew up if the 120/240 switch was wrongly set) So they would have been "linear," right? And some of them physically buzzed like crazy! If they make a physical noise like that, what is their electrical profile going to look like?

My feeling, as a non-electronics-educated end user is that I don't care whether they are linear or switching, but they should be quiet. First, physically, and second, electrically.

So, am I just doing an audiophillia bullshit thing here, as I do indeed, from time to time? No... because I believe that there are good and bad power supplies, but I am very hesitant to believe that the line is equivalent to linear/switching. Or the presence or absence (for PC equipment, etc) of an audiophile label.

By the way... I have an iFi device with their wall-wart which, according to them, is supposed to be so silent that we can hear angels' wings. This is not a high-price item. I'd like to think that their claims are true ...because I bought the thing!

Also... Is it true that, If a power supply is truly electrically noisy, it injects noise back into the mains affecting everything on that circuit? Heck! our fridges could be suffering!

 

Maybe and it depends...

In an perfect world I am sure that it makes a difference. The bigger question is does the cost of upgrading make sense. While using a LPS may maximize the performance of your amp or DAC, that doesn't mean you will outperform a higher end model without an LPS. It is all about maximizing the price/performanxe ratio and I am not convinced there are many scenarios where a LPS will do that.

I know this is a generalization, but unless you have reached a very high level system I think you would be better off spending the money elsewhere. Talking about an LPS for your OR5 in a $10000 system is much different than talking about an LPS for your $99 Schit Magni/Modi.

Is there anything specific you were referrencing?

 

The power supply is just as much a part of your "signal path" as the input stage, gain stage, output stage, etc. is. Although I suppose it all depends on the level of abstraction you want to use. I suppose it's easier for the consumer's mind to dismiss or accept certain variables based on what they see from a block diagram (how I connect things together) POV. If you can accept that those parts of the amplifier do make a difference it really isn't much of a leap to presume your given power supply can affect performance, without knowing the nitty gritty

That really is the short of it before you start diving into the technical discussion of HOW or the philosophical discussion of WHY. As to whether power supplies have their own "sound" or make "enough" of a difference can be argued ad infinitum, just as we already do with other components

 

Has someone asked you to believe this?

"Injecting noise back in to the mains" is not how I'd describe it. Electrical ripple, DC offset, rail stability etc... can affect the operation of devices in your signal path. To what extent and affect might be subjective or measurable.

 

My ears tell me whenever i replace usb power from a spdif converter or usb powered dac/amp there is a clear improvement. I'm using an older version of one of those Sboosters lps's. The sound gets more focused, less dirty and smoother.

 

More or less, yes... but here and now, no.

 

There is also more to the question than switching wall-wart vs. linear. As an example, David Berning uses switching power supplies of his own design for all of his amps, including his 75k/pair monoblocks. Clean, silent power is better than the other option, however it's achieved.

 

On switchers:

I keep hearing people say that a "good" and "quiet" switcher can be better than linear power...but I've yet to see any substance or explanation here. The virtues of linear power can be easily seen by any familiar with the actual differences between the topologies. Off the top of my head, I can only think of one particular scenario where the "unique" nature of a switching power supply could be...."nulled" for a given application I have in mind. But I really think the major reason for using a switcher in most applications is reducing cost, power consumption, heat, footprint, etc. And let's remember that all engineered products operate within a set of constraints and that NO DESIGN IS WITHOUT COMPROMISES. Even stuff that costs 75k.

Perhaps someone could give a cursory explanation or provide a link on how a switcher can offer performance advantages (measurable or otherwise) over linear power.

 

It depends on the circuit design. That's what PSRR is- power supply rejection ratio.

Of course, then you get to debate where the threshold of significance is.

 

only advantage i know of is better sustained efficiency and temperature control per watt compared to a linear. unless the switcher has a very well design power factor correction at the power supply then the switcher might suffer and drop in major output power capability under heavy sagging loads from you main electrical source/wall socket. that's why linears are usually preferred cause they can still optimally maintain good power under heavy sagging mains and will continue to output at least decent amount of current to maintain good performance...

another advantage is switchers can have quicker power bursts on demands cause they refill the caps 1000's of times per minute compared to slow refill of 80-100 or so with linear supplies providing energy for the capacitors. that's why they're usually recommended for sub duty where very high power bursts for kick drums or basslines is needed.

problem with switchers if not done right will provide more noise,produce more harmonics and will cause a major roll-off before or at 20khz if the filtering frequency is not high enough and can have some current limiting issues under heavy stress/sagging without proper power factor correction. that's why some compare lot of class d amps to tubes all the time if someone can't afford an actual tube amp or wants something similar to what they have due to the known inherent characteristics of some class d's/mos-fet designs and also lot complain that some class d's softens or darkens the top-end too much and so forth.

switchers can handle heavier reactive loads bit better too without strain or over heating if done correctly as well. that's why lot can handle down to 1ohms nominal no problem usually with not much heat issues since switchers disperse heat/energy at much faster rate. of course the power supply would still have be adequately robust designe and proper size to provide the amount of amps/current it is design for so, some smps can become pretty hefty size of it's own but, will still be far lighter and smaller to a comparable linear power supply.

lot of sworn audiophiles will continue to hate switchers and deem linear as best because linears can indeed filter main line noise bit better due to cancellation of proper frequency matching between power transformer of amp and power mains frequency but, lot of switchers can just be as quiet but, requires bit more complicating filter tactics to prevent main line noise issues and unwanted harmonic/radio noise(which in some cases can require bit extra negative feedback if in need, which is another reason why pure audiophiles stay clear of them). i think both can be very good if designed right. they each have an advantage and disadvantage to certain extent but, will perform the same if implementation is good.

 

PSRR only determines how well the amplifier can reject noise from the PS, rather than any inherent performance advantage of the switcher itself. Unless the argument is "a switcher is just as good as a linear if the PSRR is high enough". Which I might be able to buy, but it ultimately comes back to the issue of "good enough" while saving cost, power, heat, etc. Why not aim for high PSRR and the best power supply performance possible, eh? Oh yeah, gotta save on dat BOM

 

My 2 cents experience with PC PSU... Notebooks and desktops. I was skeptical at first, but a friend brought his 100w 19 volts at my house and cleared every doubt.
Everytime I have changed an atx PSU or a power brick with a LPS I have noticed clearer more defined sound with better detail retrievement.
So to me linear PSUs (even radio comunications or CNC oriented!) proved to be better than general switching ATX or powerbrick PSUs.

 

Source on any of this? Your statements in themselves don't make any sense. They only thing charging/discharging caps does is ruin their longetivity. How does that relate to handling reactive loads better or delivering power faster (i.e. power supply bandwidth)? You realize that linear regulation does not draw directly from the PS caps, but rather throws power away (via the pass device, whether it's series or shunt) so that it can maintain the voltage for any given current draw.