Best linear power supply for Nucleus now?

You probably know this, but you don’t need multiple ethernet cables from your router. You can split the cable with a non-managed switch.

I have a ethernet gigabit switch in the mix as well. The Elemen X support DSD512 (for some weird reason only up to DSD256 is supported via Roon, but irrelevant since no content exist) and PCM up to 768kHz. MQA also supported (but prefer hi-res music without MQA).

I also have Synology NAS with local music connected via ethernet cable to my local network. Using Roon to combine local content with streaming from Tidal + Qobuz. Actually Matrix Audio confirmed they plan to support Spotify Connect as well by October 2020 (not that it will be used, but a nice possiiblity).

1 Like

I actually have multiple managed Cisco switches at my place. The router also got a gigabit switch, but I have too many devices in my system.


Is there any advantage in using a linear power supply with the Nucleus ?
If so, can I use a power supply from Pro-Ject Audio with a 20 V output (Power Box RS) and not 19 V as the standard Nucleus power supply ?

Hi @Armando_Neto,

I’ve moved your topic into here, there’s lot of information on LPS here, well worth a read.

I doubt using a 20 volt supply rather than 19 volt would create an issue, but the use of any 3rd party PSU is at the users risk.

No, I haven’t read the whole thread, but if nobody refuted this… I will. That is simply incorrect.

BTW… this whole thread is a complete waste of time. How many times will you all fall into the same trap. Somebody starts a troll by saying “X can’t possibly make any difference because of Y” and it begins all over again.

X can be power cords, linear supplies, interconnects, footers, ethernet switches, and so forth. Y is some “scientific” reason or measurement that “proves” their point.

since nobody can prove or disprove what somebody else hears the debate is endless and pointless.


What about there are only two types of switching supplies is incorrect? Is there DC coming into your home?

You should read some of the replies, and the statement that as a designer, measurements and models are all there are for designing things. Some anecdotal opinion doesn’t help me optimize a circuit or change operating points etc. When someone says, this cable sounds better, I can measure its electrical properties and that’s it. IF there is no difference in those properties between two cables, and someone likes one better, how do I change the loser? What parameters do I vary? We have measurement gear that is more sensitive and more reliable than our ears and have for a long time. As an engineer that is all I have. So a “this sounds better” and there is no evidence to back it up isn’t useful.


1 Like

it may not be useful to someone trying to design a circuit, but if it “sounds better” it is better. As an end user that’s all I care about. Should I not use something that sounds better just because I can’t explain why?

I don’t say this to be mean or condescending, but if you don’t know the difference between a switching power supply (SMPS) and a linear supply then a reply here is probably not going to help much however…

at its most basic a linear supply consists of 3 components;

  1. transformer to drop voltage down to the level we need
  2. rectifier to change AC into pulsating DC
  3. filter to smooth out the DC

so can be as simply as a transformer, diode, capacitor. It is a passive device. They are usually more complex in audio supplies to include better filtering and often regulation, but much different than an SMPS.

The SMPS has a high frequency oscillator (active circuit) . It is beyond the scope of this forum to teach you basic electronics so Google the difference to get more information.

regarding your position on measurements, it is easily refuted, but if you are in the camp that “if it can’t be measured it doesn’t matter” , or even worse “they measure differently but in my opinion you can’t hear the difference” then go over to Audio Science Review and join the Luddites over there.

  1. it assumes that the instruments we are using to take measurements are more sensitive than your hearing. This has never been proven as there is no way to definitively determine how sensitive our hearing is. For instance, there are frequencies we can feel (sense) with organs other than our ears. At least our brains don’t interpret it as sound as we usually think of it. I can feel very low frequencies even if my brain tells me I don’t hear them. I had an ultrasonic jewelry cleaner years ago that operated in the 30K + region. It made me nauseous when it was running even though I didn’t “hear” it. . That is not anecdotal, I threw up.

  2. it assumes we are measuring everything that matters. Since you can’t prove a negative, how do you know there isn’t a parameter that affects how we perceive sound that we are overlooking?

  3. it assumes we are measuring everything in a way that relates to how we hear. Since we don’t know 100% how the ear/brain hears, that is something we can’t possibly know


That’s still anecdotal; the last bit makes it even more so. It’s important to know what words mean, if you want to communicate rather than pontificate.

Points 2 and 3 raise in me the thought that you may not understand the connotations of “Luddite”, either. Just me, of course. YMMV.


You missed the point of my original comment. But in addition to that, you missed why the linear supply is called a linear supply in the first place. A linear supply operates with the components (the diodes and regulators) in their linear region, so they are dissipating significant power in the rectification, filtering, and regulation. Switching supplies have active elements are are either cut off or saturated, so they do not dissipate power and are generally much more efficient because of that. The front end of a switching supply is just a hot chassis “linear supply” with a rectifier and filter cap. But that does not negate that they are both taking AC signals and rectifying and filtering them. The linear one has all of its harmonic noise products in the band of interest, the switching one not so much. The original comment was made with respect to noise. Where a switching suppy’s harmonic noise components are well out of the audible band, but a linear supply is not.

Since you seem to be in a position of great knowledge, please let me know how to improve a simple circuit like say a line level preamp with moderate gain and very small power requirements with listening tests alone. Or should I just keep buying them until I find one that strikes my fancy?

That’s all I’m going to say on your screed.


how is it anecdotal?

  1. I turn the machine on
  2. I get nauseous
  3. I turn the machine off
  4. the nauseous feeling goes away

very repeatable i.e. based on the facts as they happened, not anecdotes

Luddite = unwilling to accept that the way they are currently thinking about may not be correct, that there may be a new or different way it can be explained that they are unwilling to accept because their current technology does not explain it, stuck in the past clinging to outdated notions.

if that is an incorrect use of the term Luddite then I stand corrected on that point. It does not take away from my conclusions which basically boil down to one simple point… . It is not scientific to state something is not happening because we can’t explain it with measurements. I apologize if I clouded my argument with too many examples. Assuming that we have reached the pinnacle of science and nothing is left to be discovered is a mistake that has been made often throughout history.

and Sheldon, no matter how you try to explain it or rationalize it, linear and SMPS are very different animals. it has nothing to do with harmonics. I did not miss your point, You very clearly said they are the same

A linear supply is in no way a switching supply. Go ask an electrical engineer. And that’s all I’m going to say about it since you insist on being wrong.

I never said that or even implied it. I never said measurements weren’t useful, I said they could not explain everything. There isn’t an audio circuit designer in the world who doesn’t fine tune their devices by listening to them and making changes based on those listening tests. If measurements explained everything they could just do it all on a computer, build it, and ship it… but they never do that… why? Simple… because measurements don’t explain everything.


This is one of the best LPS you can purchase:

Expect some delivery time. 30 day’s. Postage also take a bit time. It’s produce in a remote location :grinning:


Why use a linear power supply when there are only switched-mode power supplies for the components on the mainboard anyway? Makes no sense to me.


Perhaps less noise than from the external SMPS than might be present from a LMPS

1 Like

DC switch mode PS ? :joy:
You’re joking ?

What do you want to tell me?
If you know anything about electronics, then it is logical that there is no point in providing a super DC voltage if this is equalized by switching power supplies on the mainboard. The different voltages for the components installed on the mainboard are namely generated by small switching power supplies on the board.

Well, we’re obviously mixing SMPS used in AC to DC vs DC to DC regulators.

I wasn’t aware that small internal regulators in a motherboard was noisy.
I’m used to technology like LDO regulators (LM3045).

Among those building low noise PC, LPS of good quality is the normal way to go in order to achieve better SQ. Maybe even use several depending on voltage requirements. (Like you can power the SSD separately, or some PCI cards).

It’s certainly a widely touted audio shibboleth. I’m unaware of evidence, measurements, that demonstrates an improvement in SQ, though. Do you know of such?

1 Like

So are you saying that there isn’t a single thing inside the machine that runs on 19v? Say it isn’t so?!

Well, your ears is definitely the most important instrument :joy:

I once added Uptone Audio LPS1.2 to a Squeezebox Touch. The changes was crazy.

Power supplies can be measured. SQ can’t. Even your hearing can be measured, but the results has probably nothing to do with SQ.

It’s wildly reported over at that different power supplies matters.

Good designed DAC’s as an example that spends a lot in sever power supplies, and don’t mix powering digital and analogue circuits. They do for a reason.

You should google a bit in order to find the answer.

What equipment do you have that we can test a power supply on ?

Let me also add, I never add an audiophile PS to a PC, as I believe it’s an good idea to stay away from PC in audio. However I of cause have a Roon server, that may or may not benefit from a better PS.

But if so, I would need my SonicTransporter i5 to move closer to my audio gear.
I don’t think that PS matter at all when it’s remote placed. Hence why it’s remote.
(So is the Tidal source :grinning:)

And if I move it closer, I would probably separate the network by an opticalModule (I have one), which I would give a good PS. (Also have).

Since I already have the opticalRendu as well as the etherRegen with an external clock, the difference I would here in my case, may not exist at all.

So point here is things may be very system dependent. Even with good measurements…

What we’re doing isn’t an exact science. But there is some things that most people can agree upon. Like MP3 sounds different from hires. And where is that measurement ?

As it is accepted that low noise power supplies is best. And you can actually measure them.

Finally, it’s possible to build fantastic endpoints by using a PC with the correct parts. And that is definitely not standard SMPS.

You get what you pay for. Quite often.