Linear v Switched mode PSUs - which is 'best' for audio?

All general audio posts go here.
User avatar
CN211276
Posts: 6522
Joined: Tue Aug 09, 2016 9:29 am
Location: Cardiff
Has thanked: 1401 times
Been thanked: 976 times
EUROPEAN_UNION

Re: Linear v Switched mode PSUs - which is 'best' for audio?

Unread post by CN211276 »

I think it is fair to say that the effect good power supplies have on digital source components is dependent on a number of variables and nothing is set in stone. The quality of the PS of the source component and the transparancy of the rest of the system being major factors. From personal experience my Topping E30 benefited most, but the PS cost twice as much. Next comes my Sonore streamers which are supplied without PSs. I believe that the literature states that they benefit considerably with high quality PSs and I found the Sbooster was a major step up from the MCRU. Although Chord state that their components do not benefit from LPSs, this is not true. The performance of the Qutest and Mscaler have been lifted, but not to the same extent as the streamers.
These users thanked the author CN211276 for the post (total 2):
Lindsayt (Sun Dec 04, 2022 12:21 pm) • Berty bass (Mon Dec 05, 2022 12:44 pm)
Main System
NVA BMU, P90SA/A80s (latest spec), Cube 1s, TIS, TISC(LS7)
Sonore OpticalRendu, Chord Mscaler & Qutest, Sbooster PSs
Network Acoustics Eno, ifi iPurifier3, AQ JB FMJ, Cisco 2940 & 2960
DH Labs ethernet, BNC & USB cables, Farnells cat 8 ethernet cable

Second System
NVA P20/ A20, Cubettes, LS3, SSP, SC
Sonore MicroRendu, Chord Mojo 2 MCRU PSs, AQ Carbon USB cable & JB FMJ

Headphones
Grado SR325e/Chord Mojo, Beyerdynamic Avetho/AQ DF Colbat

RIP Doc

User avatar
Fretless
Posts: 9290
Joined: Wed Nov 12, 2014 12:15 pm
Location: Somewhere in Holland
Has thanked: 1487 times
Been thanked: 2280 times
Netherlands

Re: Linear v Switched mode PSUs - which is 'best' for audio?

Unread post by Fretless »

Pi-based streamers are very sensitive to power supplies with noticeable differences between switched, battery and linear.
These users thanked the author Fretless for the post (total 2):
Lindsayt (Sun Dec 04, 2022 12:22 pm) • CN211276 (Sun Dec 04, 2022 12:26 pm)

Upstairs:
Vinyl
Pro-Ject 1.2 + Grado Sig Jr + Cambridge Alva Duo
DigiVolumio PC + Kiss DP-500 + Sabaj A20d
NVA: P50sa - Cube2 - SSP - LS6+ Sabaj A10a (x2)
Downstairs:
Vinyl
Logic DM101 + Syrinx LE1 + Grado Sig MCX
DigiDenafrips Ares II + Volumio PC + Cambridge CXC
NVA: P50 - BMU+ Aiyima A07 MAX (x2) + Arcam One
HP: HifiBerry Digi+ PRO + Sabaj A10d
Office:
Allo DigiOne SIG + SMSL M300se + Douk G4 (x2)
Mission 760 + Monolith THX AAA 887
Headphones: German Maestro, Monolith & AQ

User avatar
karatestu
Posts: 5945
Joined: Sun Jan 08, 2017 4:40 pm
Location: North Yorkshire
Has thanked: 1865 times
Been thanked: 1378 times
Great Britain

Re: Linear v Switched mode PSUs - which is 'best' for audio?

Unread post by karatestu »

I have done a lot of tinkering with CD players and found that power supplies are critical especially for the actual dac chip and following anologue stage. The purely digital parts before the dac are not as fussy in my experience.

My Naim CD3.5 was born with a single toroid based power supply. After thesmoothing capacitors the power rail branched off and went to 16 separate regulators. Nothing fancy, just plain old LM317. Of course you could add an external psu to power the anologue stage leaving the internal psu to power everything else (including the dac chip). I added a hicap and it did improve things until I bypassed the output stage altogether rendering it redundant.

The player went through many changes in the proceeding years. I tried fancy super regulators on the anologue stage opamps - that had a large effect. Biggest improvement came from fitting a pfm flea regulator with a tent lab's oscillator (clock) and giving it its own dedicated psu .

Then i isolated the dac chip and gave it its own psu. Well, three psu's for the dac actually - each power supply pin got it's in psu. That was an eye opener. I put several more fancy regulators in key positions. One of these went on the digital supply rail and consists of a FET based capacitance multiplier. It eats noise for breakfast. The few remaining LM317 regs were spruced up in various ways including replacing the Rset resistor with two green LED's in series. Gives a 5V output and is way quieter. It looks like Blackpool illuminations in my cd player.

I think what I am trying to say is you can't go far enough with power supplies. My CD3.5 has five of the buggers and that's without the ones i would require for the anologue output stage which I bypassed. They all use EI frame transformers which I prefer to toroid - probably due to them letting less noise through between windings.

Alot of the regulators respond well to improvements made upstream, even the fancy super regs. Certain things need low impedance power while others are not bothered. The clock I made has a fancy regulator with a capacitance multiplier on the front which has a rather high output impedance. It actually takes quite a while to get up to full voltage but it eats more noise as the resistor gets bigger. The FET based capacitance multipliers on the dac chip however resulted in a slow, sat on type of sound. Very detailed and clear but had no get up and go. They were replaced rather quickly.

It was a lot of fun and resulted in a unique player which I must say sounds damn fine.
These users thanked the author karatestu for the post (total 4):
CN211276 (Sun Dec 04, 2022 2:56 pm) • Lindsayt (Sun Dec 04, 2022 3:36 pm) • Fretless (Sun Dec 04, 2022 3:44 pm) • Berty bass (Mon Dec 05, 2022 12:44 pm)
DIY FREE ZONE

Post Reply