All Linear Power -- Possible?

[GUTB]

Is building a PC or streamer working on all linear power, ie, no switching regulators, possible? Do you need to buy a streamer designed like that?

[seeteeyou]

Most likely this is the only way to go

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/153/?tab=comments#comment-724987

On 10/2/2017 at 8:09 PM, romaz said:

Finding a suitable motherboard was extremely difficult and it became s a matter of finding the motherboard with the least compromises.  DFI was willing to design a motherboard for me to my specifications, however, it would have cost $1,500 minimum to get the project off the ground.

 

https://audiophilestyle.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/page/241/?tab=comments#comment-773624

On 1/31/2018 at 6:50 PM, romaz said:

Having communicated with Mark Jenkins about the Gen3, he commissioned ASUS to build him a custom motherboard and apparently he did a lot of testing and listening before deciding on what he wanted on this motherboard.  Of interest, he found that in some locations on the board, switching regulators were better than linear regulators because they had faster response (I presume he's talking about the regulator for the CPU itself).

 

https://audiobacon.net/2018/10/10/rmaf-2018-innuos-zenith-statement-vs-zen-mk-3-music-servers/2/

Quote

Custom motherboard by Super Micro.

 

[marce]

Not really and why would you when SMPS's are the best option, heat, size, locality (critical), power density, speed of response etc. Getting an understanding of the requirements of building a power delivery system for a digital, digital/analogue or a digital/analogue/rf  design would help understand the complexities. 

[Ralf11]

marce, your use of expertise and electronics knowledge is anti-audiophool

[AnotherSpin]

Again, if some people want to upgrade from SMPS to linear PSU, what is the problem of those people who does not believe it will bring a difference?

 

[GUTB]

16 hours ago, marce said:

Not really and why would you when SMPS's are the best option, heat, size, locality (critical), power density, speed of response etc. Getting an understanding of the requirements of building a power delivery system for a digital, digital/analogue or a digital/analogue/rf  design would help understand the complexities. 

 

Switching power is the enemy of audio.

[Ralf11]

2 minutes ago, GUTB said:

 

Switching power is the enemy of audio.

 

 

yes, and that is why it is never used for sensitive laboratory research equipment, or for important medical equipment.

 

 

oh wait.... !!!

[GUTB]

Just now, Ralf11 said:

 

 

yes, and that is why it is never used for sensitive laboratory research equipment, or for important medical equipment.

 

 

oh wait.... !!!

 

Did you know that hearing is our fastest sensory input?

[Ralf11]

2 hours ago, GUTB said:

 

Did you know that hearing is our fastest sensory input?

 

No.  And I still don't

[fas42]

A better phrase could be that it's our most adaptable sensory input, capable of discerning variations in a massive dynamic range - with built-in, brilliantly engineered AGC, it can withstand staggeringly intense input for tiny periods of time, yet detect below our technical ability to create noise-free environments with ease. Which is part of the reason why it's hard to "trick it",

 

Most playback will fail at one or the other end, or both - not do quiet properly, not do loud properly. Part of the task is to get the setup to do the "extremes" right - this is very achievable; but won't, unless the efforts are made.

[marce]

9 hours ago, GUTB said:

 

Switching power is the enemy of audio.

No it isn't, with digital you need small footprint point of load supplies can supply the required power and more importantly not add excessive heat to the design. You have numerous power islands on a correctly design digital digital/analogue boards, some can have 20+, digital/analogue boards even more. A typical board will have a 3V3 SMPS for general supply and IO, from this separate low noise LDO supplies will provide the various voltages required by the main devices, 1 - 1.2V core voltage, 2.5V for other processing areas. You don't daisy chain multiple SMPS's, you will have one main SMPS feeding several LDO sub supplies and general digital, such as gates etc. The on board SMPS's are also run at different frequencies and more often theses days spread spectrum. On the analogue side on instrumentation/medical/space of design the main analogue supplies are from linear supplies, the current requirement is less  for analogue (usually 1-2A regs) with again any local supplies provided by low noise LDO's. Analogue supplies are usually +/- 15, 12, 5 or 6 for lower voltage stuff and sometimes quite high voltages depending on the output requirement, though high current output is preferable on a separate board as the design requirements are different...

So its not SMPS vs LPS is designing the power delivery system to fit the design, minimise interactions between the sections, separating sensitive areas of design from the digital core and also generic IO (Ethernet, USB).

😃

[marce]

5 hours ago, fas42 said:

A better phrase could be that it's our most adaptable sensory input, capable of discerning variations in a massive dynamic range - with built-in, brilliantly engineered AGC, it can withstand staggeringly intense input for tiny periods of time, yet detect below our technical ability to create noise-free environments with ease. Which is part of the reason why it's hard to "trick it",

 

https://www.frontiersin.org/articles/10.3389/fpsyg.2014.00725/full

[Ralf11]

15 hours ago, fas42 said:

A better phrase could be that it's our most adaptable sensory input, capable of discerning variations in a massive dynamic range - with built-in, brilliantly engineered AGC, it can withstand staggeringly intense input for tiny periods of time, yet detect below our technical ability to create noise-free environments with ease. Which is part of the reason why it's hard to "trick it",

 

Most playback will fail at one or the other end, or both - not do quiet properly, not do loud properly. Part of the task is to get the setup to do the "extremes" right - this is very achievable; but won't, unless the efforts are made.

 

are you talking about vision?  or smell?

[GUTB]

10 hours ago, marce said:

No it isn't, with digital you need small footprint point of load supplies can supply the required power and more importantly not add excessive heat to the design. You have numerous power islands on a correctly design digital digital/analogue boards, some can have 20+, digital/analogue boards even more. A typical board will have a 3V3 SMPS for general supply and IO, from this separate low noise LDO supplies will provide the various voltages required by the main devices, 1 - 1.2V core voltage, 2.5V for other processing areas. You don't daisy chain multiple SMPS's, you will have one main SMPS feeding several LDO sub supplies and general digital, such as gates etc. The on board SMPS's are also run at different frequencies and more often theses days spread spectrum. On the analogue side on instrumentation/medical/space of design the main analogue supplies are from linear supplies, the current requirement is less  for analogue (usually 1-2A regs) with again any local supplies provided by low noise LDO's. Analogue supplies are usually +/- 15, 12, 5 or 6 for lower voltage stuff and sometimes quite high voltages depending on the output requirement, though high current output is preferable on a separate board as the design requirements are different...

So its not SMPS vs LPS is designing the power delivery system to fit the design, minimise interactions between the sections, separating sensitive areas of design from the digital core and also generic IO (Ethernet, USB).

😃

 

Switching power is the enemy of audio because it gets high-frequency hash into the audio circuit and AC mains, has problems with ground loop formation, etc. 

[Ralf11]

why do you think it is used in critical medical equipment?

 

why do you think it is used in highly sensitive research equipment?

 

why are you confused about high grade SMPS and low buck crap?

 

 

[marce]

9 hours ago, GUTB said:

 

Switching power is the enemy of audio because it gets high-frequency hash into the audio circuit and AC mains, has problems with ground loop formation, etc. 

Jeez you are close minded. SMPS's are used on far more sensitive stuff than Audio as Ralf has mentioned. Without the small form factor SMPS's we wouldn't have all the technology we now have.

A lot of the reason for sights like this is to improve audio, looking at new (!) technology would seam to be one of the ways forward, not just repeating unsound mantras and rejecting a technology that powers most of today's digital... And digital control is everywhere.

What about the noise digital itself makes when switching, noise from the clocks etc. There is far more to this than just LPS good, SMPS bad.

[sandyk]

9 hours ago, Ralf11 said:

why are you confused about high grade SMPS and low buck crap?

 

 Perhaps that's because most affordable Consumer A and V equipment uses low buck generic Asian sourced SMPS crap similar to the attached schematic ,with a low value high voltage capacitor (or 2) between the Primary and Secondary sides ? 

 

 Professional grade SMPS are far better designed and can also be very low noise, with minimal leakage back to the A.C. Mains Sewer.

[marce]

You still need on board point of load SMPS's for optimum performance...

This is FPGA based but it is similar for processors and then you have to add in the power requirements for any memory, interfaces to the outside world etc.

https://www.embedded-computing.com/embedded-computing-design/an-introduction-to-fpga-power#

A lot of basic motherboards will cut down on the power delivery network, for cost reasons, and use pi filtering for power islands instead of LDO's and have more power sharing where voltages are the same.

Using small footprint supplies on board allows the power to be where its needed, next to a device, this minimises the coupling of the supply to other circuitry and other circuitry coupling to the supply. Also small high efficient SMPS's can provide the required currents (many 10's of amps) without the heat that LPS's would add in the same position, as well as an greatly increased footprint.

There is far more to powering a PC or streamer than just a front end LPS (or SMPS)...

[opus101]

12 hours ago, GUTB said:

 

Switching power is the enemy of audio because it gets high-frequency hash into the audio circuit and AC mains, has problems with ground loop formation, etc. 

 

Some clarification here wouldn't go amiss. Off-line switching power is the enemy of unbalanced audio interconnection for the reason of common-ground impedance coupling of ultrasonic hash into cable shields. Go balanced and there's no issue.

[AnotherSpin]

10 minutes ago changed original SMPS to custom made LPS to power Netgear switch employed between server and NUC endpoint. Started listening to the same album which listened before the change. Can not explain the difference in SQ. Only could tell it sounds like different recording, have difficult time believing it was exactly the same what I listened before.

[marce]

The question was is all linear power possible for a PC or streamer...

[sandyk]

3 minutes ago, marce said:

The question was is all linear power possible for a PC or streamer...

 

 Probably not unless you want a motherboard the size of a suitcase !

[PeterSt]

There is a difference between the main PSU (let's have at least two different ones for digital and analogue) and the regulators behind it.

 

I personally would never dive into changing the regulators - or at least not on motherboards. What Marce told, applies.

 

It is different for those main PSU's;

First off, they make too much of a difference, SQ wise, to not do anything about them. Btw, Marce made a differentiation on this too (main PSU LPS vs SMPS).

The main PSU can always be changed from SMPS into LPS. Space constraints are quite OK, BUT a lot of control as such will be required in order to maintain the PSU and everything related to it, sufficiently small.

 

IMO this is a skill in itself, and a foremost function is the speed. That too was mentioned, though not really in the audio realm, while for audio this is crucial (and hey, for a processor too of course). Don't believe that ? then try it out first. Anyway :

Where the SMPS easily wins, is on that speed. Remember that we are talking MoBo's and the like (say computers) and that the required sudden current draw can be "huge". Now, where we tend to counteract the more sluggish LPS with a crazy bank of capacitors so a suitcase sized box is needed after all (and it still will be slower), it is the big trick to avoid that while make the LPSU fast as h*ll anyway. This is not so easy (to design) but it can be done - and in a very small form factor (and I am not dreaming here).

 

Back to the regulators ...

I would avoid the switching ones at all times. But, this is in designs of my own, and probably it is only common "audio sense" to do it like that. What I am saying is that I will never interfere with how things have been set up on a MoBo because it is complicated, usually requires more space, and well, 100% of what Marce told.

When LDO's are used and used with common sense, hence appropriate matching so no 50W is dissipated where 1W is possible, I see no harm in the space/context of the PC (MoBo) "producing sound". IOW, a MoBo does not produce sound to begin with, BUT it is about coupling in noise at places we don't want it (think about what the sudden draw of a virtual infinite rise time of 1A++ implies). Normally this is the 240V/120V backdoor, but sadly also USB which is PE related. From there anything can happen.

And now it gets vague ...

Avoiding the switching main PSU is always a good thing *IF* it is sufficiently fast.

 

 

[Ralf11]

it might be frutifull to pull apart the very high-end DACs that have 3 different PSs in them - and see which PS is the highest quality one

 

I'd expect that the PS that serves very low level analog signals just after the D>A conversion would be most critical as the signal is smallest there

 

Likewise, I expect the PS in my Audio Research pre-amp is a pretty good one (it was $5k new, and they are known for hand testing and selection of caps, resistors, etc.)

[botrytis]

On 8/2/2019 at 6:58 PM, GUTB said:

 

Switching power is the enemy of audio.

 

All my instruments in lab, including an MS have switching PS. I guess you should tell the instrument manufacturers they are wrong.

 

Any measurements to back your claims up?