Practical suggestions for improving Power supply/electrical issues affecting Sound Quality

[Audiophile Neuroscience]

Hi to everyone,

I am assembling a HTPC/media server and looking for practical suggestions for improving power supply/electrical issues affecting Sound Quality. I have placed the thread on the general board as I thought the question goes beyond system building per se.

 

I am particularly keen not to 'spark' (pun, ha ha) a "bits is bits" argument so respectfully request tolerance from the naysayers. I would also like to request keeping the language as technically simple as possible so I can hopefully understand.

 

So, I roughly understand that power supplies can introduce both physical and electrical "noise" and that there are other sources of electrical interference. I have heard of things like ground loops, Radiofrequency interference, Electromagnetic Interference, ripples (presumably in the voltage regulation) but not really sure how they affect SQ..apart from say an audible hum etc.

 

In the case of SPDIF-out from an on-board sound card what can be done to isolate from electrical noise?

 

In the case of external USB DACS I am aware there are products like iUSB iFi-audio iUSB that purport to clean up the power supply.

 

I am also aware there are "power conditioners" for the mains power.

 

My analogue components and main external DAC have their own dedicated "power supplies" that came with them, so presumably they are having the intended beneficial effect.

 

I know some favour linear power supplies but I am not aware if these are available to power a pc? Some say switch mode is okay too. Or is it really just about buying a good quality PSU with good specs ?

 

Like to hear how you have approached this issue and what effect it has had on SQ. Thanks for any thoughts or advice.

 

Cheers,

David

[sandyk]

David L.

Please check your PMs.

Kind Regards

Alex

[wakibaki]

It is entirely possible to run a computer from a linear PSU, but the high wattages involved and poor efficiency of linears make this problematic. It is more commonly done with low power 'fanless' types using processors such as the Atom. A fanless type is worth considering, not because of its greater suitability for operating from a linear supply, but because the majority of problems encountered by users of DAWs centre around the noise of the computer fan, rather than any deficiencies in recording or playback.

 

A computer is a difficult environment in which to operate sensitive analog equipment, because it is full of digital circuitry racketing away at frequencies in the microwave bands. The higher the frequency the smaller the antenna, and the greater the tendency for radiation to occur, and the greater the likelihood of reception taking place.

 

Many desktop computers are assembled without the benefit of EMC compatibility testing. This means that their radiated emissions can be high despite the metal case.

 

What does all this mean in practical terms?

 

It means that if you intend to use an external DAC you are less likely to encounter problems if you buy a laptop or a desktop assembled by a reputable manufacturer, as in both cases it should have been tested for radiated and conducted emissions.

 

It needs to be borne in mind, however, that the majority of these effects take place well above the frequencies regarded as ultrasonic, and that other effects have to come into play for these radiations to become audible. A mobile phone is purpose built to deliberately output a few watts of RF power, but you cannot hear any effect on the audio output other than the intended call.

 

The designers of both internal and external sound systems are not unaware of the potential problems of operating in close proximity to or connected to a computer.

 

Unless you become aware of specific problems with your off-the-shelf setup, you should not take any measures other than the normal ones of plugging the equipment in and operating it. Only if there are very gross effects present will you be able to reliably detect improvement or degradation of the sound, other than by instrumented test.

 

w

[sandyk]

David

There will soon be a range of Linear PSUs designed by well respected EE John Swenson made available for the Mac Mini. Paul Hynes also makes very high quality Linear PSUs to order, but there can be several months lead time.

Regards

Alex

[Audiophile Neuroscience]

Hi Wakibaki and thanks very much for your input

It is entirely possible to run a computer from a linear PSU, but the high wattages involved and poor efficiency of linears make this problematic.

yes, this is what I have heard. My problem is I want/need a more powerful computer to handle full madvr video rendering eg with at least a nvidia GTX660 gpu and for it to be a true media server, multiple Hard discs etc. I have resigned that a small fanless pc will likely not cut it. I am not concerned by a small amount of fan noise as the server can be placed behind a false wall if required.It would however mean more electrical noise introduced into the system. I have heard of sotm products to combat this.

 

 

 

Many desktop computers are assembled without the benefit of EMC compatibility testing. This means that their radiated emissions can be high despite the metal case.

 

my problem here is that I am assembling the parts but can only guess it wouldn't be any worse than a pre-assembled product. ??

 

 

 

It needs to be borne in mind, however, that the majority of these effects take place well above the frequencies regarded as ultrasonic, and that other effects have to come into play for these radiations to become audible. A mobile phone is purpose built to deliberately output a few watts of RF power, but you cannot hear any effect on the audio output other than the intended call.

 

The effects of ultrasonics on SQ has been debated elsewhere...as has all aspects of subjective vs objective which I do not want to evoke here. I respect opinions from both sides, but in truth believe both sides equally subjective for different reasons. Importantly it is the folks that have made simple real world mods to their 'regular' pc's that I want to hear from, whether qualitatively or quantitatively making a difference.

[Audiophile Neuroscience]

Hey Alex

David

There will soon be a range of Linear PSUs designed by well respected EE John Swenson made available for the Mac Mini. Paul Hynes also makes very high quality Linear PSUs to order, but there can be several months lead time.

Regards

Alex

 

what watts will they weather (sorry, carried away with alliteration :-)

[sandyk]

Hi David

It will be mentioned in one of the threads with very long names started by Alex C. (Superdad)

If you PM Alex C. he will be able to give you a lot more in depth information, or point you to the relevant page of one of those threads.

 

Regards

Alex

 

P.S. See from post 4

 

http://www.computeraudiophile.com/f10-music-servers/why-linear-power-supply-18929/

[wakibaki]

my problem here is that I am assembling the parts but can only guess it wouldn't be any worse than a pre-assembled product. ??

 

No. This is precisely what I mean by 'assembled without the benefit of EMC compatibility testing'. You cannot, unfortunately, be assured that it will be no worse than a pre-assembled product. As simple a feature as the way that internal cables are dressed can affect whether an item passes or fails. That is why they have to be tested.

 

This, of course, renders moot the question of installing a linear PSU. What point is there in installing a quiet PSU when:

 

1. Radiated emissions may already be drowning out conducted.

 

2 Conducted emissions are probably not markedly reduced by using a linear PSU.

 

Any device rejects some artifacts onto the mains. Similarly, any device in the computer rejects some artifacts onto it's power supply, so the power may be corrupted despite the fact that it comes from a quiet source.

 

Power supplies (in the routing sense) should be arranged so that noisy components should not share tracking with quiet ones, but in digital equipment the noise margin is much greater, so issues such as these may well not be as well attended to in a computer as might be thought desirable from an analog designer's POV. Who knows exactly what route the power takes on it's way to the soundcard or USB port, or whether it is different on a different motherboard?

 

Do not despair, however. I run several PCs around the house, all self-assembled, they are all used for music from time to time and I notice no problems. They all use the regular PSUs that were built into the cases, everything connects to or is in close proximity to a 2 or 3 GHz motherboard.

 

Audio device manufacturers will anyway attempt to control the noise on the power supply by e.g. decoupling capacitors, lossy inductors (ferrites) or local regulation. Similarly, radiated emissions may be shielded from some areas of a PCI soundcard by Faraday cages, or traces in 4-layer or greater boards may be buried and run in between solid copper planes. If, for example, a soundcard manufacturer claims a SNR of XYZ dB, then he must accept that independent testing may attempt to verify this in non-ideal (noisy) conditions, and he better build a device capable of aquitting itself in the kind of environment which it will typically be employed, else his reputation will inevitably suffer.

 

Build the system as standard, make a neat job of it by shortening (folding and cable tying or clipping any wires or ribbon cables) run it for a couple of weeks. If you, your family and friends do not hear any problems, distortion, noise (hiss), hash or changes in timbre with onscreen activity, forget it. Life's too short.

 

One of my systems makes use of a soundcard and small (4-channel) mixer. The connections from the sound card to the mixer are coaxial. This means that signals for record or playback are not exposed to digital noise until they have been processed and amplified (to line level). I thoroughly recommend such a system, which is perfectly adequate in a domestic situation, and affords the opportunity for high-quality recording with the addition only of some microphones, the mixer providing phantom power. There is an integral headphone amplifier.

 

w

 

I think you may be mistaken about the fanless PC, see here for example:- Mini-ITX 100% fanless HTPC: MPC-HC+LAV+MadVR 1080p OK ==> 13W power under WIn7

 

At these power levels a linear PSU becomes much more of a why not, even though the benefits might be inaudible. Giving way to the desire for a certain amount of overkill is understandable where the cost impact is small. Of 2 PSUs of equal cost and capability we will always choose the quieter.

 

There are some measures which are more likely than others to be cost effective if it emerges that EMI is a problem, but I will not go into that here.

[sandyk]

Any device rejects some artifacts onto the mains. Similarly, any device in the computer rejects some artifacts onto it's power supply, so the power may be corrupted despite the fact that it comes from a quiet source.

 

As E.E. John Swenson has pointed out even Linear PSUs are also capable of injecting some artifacts into the mains.

John typically uses a series 22nF and 330 ohms across the outers of the secondary windings of the transformer for this purpose.

Fred

Have you read some of the reports from John Swenson in this forum about his new Linear PSUs for the Mac Mini, including a surprise result when using a Kelvin Sensor with one version ?

Incidentally, I was serious about hearing about your headphone amplifier designs, as this area is presently exploding due to those stylish but non natural sounding Beats headphones favoured by many of the younger generation.

A high quality headphone amplifier coupled with a much technically higher quality pair of headphones can be a revelation to many, as well as a fine complement to a decent home stereo system.

 

Regards

Alex

[mayhem13]

Also, there's quite a few third party sites testing ATx PSU's frequently for ripple and other parameters. On a site recommendation I purchased my 600w PSU and scoped it.....ripple is lower than most other comparable PSUs available to date.

 

Are there other factors involved with a switching supply that might effect SQ as Alex eludes to?........can't say as I've never experienced anything to support it either way, but that's only one man's experience. Scientifically, the case supporting a linear supply holds water for a reduced noise floor.......but from what I've measured and experimented with, said reductions have always been 20-30db below audibility and way below the average rooms ambient noise floor.

 

600w PSU. ATX intel board i5 integrated graphics, 5 mixed hard drives acting as a server with networked AV and direct connection for high end desktop listening station.

[Bunny]

Hi to everyone,

I am assembling a HTPC/media server and looking for practical suggestions for improving power supply/electrical issues affecting Sound Quality. I have placed the thread on the general board as I thought the question goes beyond system building per se.

 

I am particularly keen not to 'spark' (pun, ha ha) a "bits is bits" argument so respectfully request tolerance from the naysayers. I would also like to request keeping the language as technically simple as possible so I can hopefully understand.

 

So, I roughly understand that power supplies can introduce both physical and electrical "noise" and that there are other sources of electrical interference. I have heard of things like ground loops, Radiofrequency interference, Electromagnetic Interference, ripples (presumably in the voltage regulation) but not really sure how they affect SQ..apart from say an audible hum etc.

 

In the case of SPDIF-out from an on-board sound card what can be done to isolate from electrical noise?

 

In the case of external USB DACS I am aware there are products like iUSB iFi-audio iUSB that purport to clean up the power supply.

 

I am also aware there are "power conditioners" for the mains power.

 

My analogue components and main external DAC have their own dedicated "power supplies" that came with them, so presumably they are having the intended beneficial effect.

 

I know some favour linear power supplies but I am not aware if these are available to power a pc? Some say switch mode is okay too. Or is it really just about buying a good quality PSU with good specs ?

 

Like to hear how you have approached this issue and what effect it has had on SQ. Thanks for any thoughts or advice.

 

Cheers,

David

Yes EMI/RFI all affect sound quality including hiss and hum in your speakers.

 

I have the best solution for you.

 

Underclock your computer. And undervolt it, both CPU and motherboard. The lower the wattage the better the sound quality. If you have a large monitor screen or TV, run the long power cord and plug on the outlet far, far away from your audio system. This reduced noise infestation. You can thank me later.

 

 

If you have a large dedicated video card in the system, remove it and use a little dinky one. Minimal number of peripherals and low power consumption will help minimize noise in your computer system. Even if you use an external USB DAC you still need to do this. They are all interlinked.

 

 

And finally, go get da power cables. Upgrade to high end. This is the most effective upgrade in this universe. Even the aliens will tell you to upgrade power cables. Depending on what model you choose and what brand you get it could cost anywhere from $500~ $2 million. You just gotta look around and research and research and research.

 

 

No don't use power conditioners. They hamper sound quality. Connect the high performance power cables directly to the wall outlet if you want better sound.

Use powerstrip only if you run out of outlets.

 

 

bunny

[coxhaus]

Yes EMI/RFI all affect sound quality including hiss and hum in your speakers.

 

I have the best solution for you.

 

Underclock your computer. And undervolt it, both CPU and motherboard. The lower the wattage the better the sound quality. If you have a large monitor screen or TV, run the long power cord and plug on the outlet far, far away from your audio system. This reduced noise infestation. You can thank me later.

 

 

If you have a large dedicated video card in the system, remove it and use a little dinky one. Minimal number of peripherals and low power consumption will help minimize noise in your computer system. Even if you use an external USB DAC you still need to do this. They are all interlinked.

 

 

And finally, go get da power cables. Upgrade to high end. This is the most effective upgrade in this universe. Even the aliens will tell you to upgrade power cables. Depending on what model you choose and what brand you get it could cost anywhere from $500~ $2 million. You just gotta look around and research and research and research.

 

 

No don't use power conditioners. They hamper sound quality. Connect the high performance power cables directly to the wall outlet if you want better sound.

Use powerstrip only if you run out of outlets.

 

 

bunny

 

I basically agree with you on your recommendations for using a good power cord. I have added a Bybee power plug with a high end Purist power cord. It makes a considerable different in sound quality. I have never heard a linear power supply on a PC so I don’t know if it would make a difference but if the power supply like most linear power supplies are small I would be leery as to whether they can deliver punch for a big system.

 

As far as power conditioning I have tried a Tripp lite 1200 watt power conditioner not the battery kind but the transformer ones and it was thin and lacked bass. I use a 3 large power conditioner for my home stereo and they work great. They weigh over 120 pounds each and run off 220 volts and live under the house as they hum. I have my PC plugged into one circuit which has 2500 watts available which helps the sound a lot.

 

My opinion is do not use any cheap power strips as they affect the sound in a negative way.

[Paul R]

Hi David - Buy a Mac Mini and wait for Alex C. and John S.'s new power supply.

 

-Paul

[astrotoy]

I was using a Mac Mini from late 2009 with 8GB of RAM to run A+ and JRMC19. I had some major issues with running multichannel files (ripped from my SACD's). The files would pause for about 1 second or less and then continue, every few minutes. I am using an Exasound E28 to decode the DSD mch files. I emailed George of Exasound explaning the situation and my equipment. He emailed back that I needed a more powerful Mac Mini - an i7. I bought one with 8GB of RAM and everything is working fine.

 

PS. I didn't have any problem with running stereo files, only multichannel.

 

Larry

[Audiophile Neuroscience]

Hi and thanks to all

As simple a feature as the way that internal cables are dressed can affect whether an item passes or fails. That is why they have to be tested.This, of course, renders moot the question of installing a linear PSU. What point is there in installing a quiet PSU when:1. Radiated emissions may already be drowning out conducted.2 Conducted emissions are probably not markedly reduced by using a linear PSU. Any device rejects [injects?] some artifacts onto the mains. Similarly, any device in the computer rejects [injects?] some artifacts onto it's power supply, so the power may be corrupted despite the fact that it comes from a quiet source.

 

Yeah , good points. I have always taken care physically separating interconnects from power cables/sources with my analogue components and using well shielded cables. This reasoning wasn't lost on me for assembling a pc, just that it wasn’t at the forefront of my mind. I will endevour to source some shorter internal sata III cables screened from outside and IIRC internal cable screening.

 

Who knows exactly what route the power takes on it's way to the soundcard or USB port, or whether it is different on a different motherboard?

 

I am trying to find ways of isolating/cleaning up critical audio pathways eg involving usb or spdif .To this end maybe something like JLH on a pcb power supplies that Alex talks about or the iFi products.Maybe some sound card with intrinsic power filtering for spdif or other spdif converter that also cleans power.

 

Do not despair, however. I run several PCs around the house, all self-assembled, they are all used for music from time to time and I notice no problems. They all use the regular PSUs that were built into the cases, everything connects to or is in close proximity to a 2 or 3 GHz motherboard.

That was what I was intending to do and not worry about it but I got a bit of audiophile nervosa….you know,jeez,I better look into fixing power issues before I build, not sure if it will make a difference but do it anyway…..

 

Audio device manufacturers will anyway attempt to control the noise on the power supply by e.g. decoupling capacitors, lossy inductors (ferrites) or local regulation.

I hope Alex doesn’t mind me mentioning that he uses a C-L-C filter (Capacitance, inductance,capacitance) for hos C drive and JLH ina spare drive bay supplying his optical drive (for ripping and burning). This is exactly the type of measures I want to explore.

 

Similarly, radiated emissions may be shielded from some areas of a PCI soundcard by Faraday cages

I know you're not suggesting it but probably beyond my capabilities. I won't even go into does it make a difference and how do you know. If it's reasonably straight forward to implement and not too pricey I have nothing against trying it. As you also say life is too short to obsess about it, but then again, it can be all part of the hobby and fun

 

I think you may be mistaken about the fanless PC, see here for example:- Mini-ITX 100% fanless HTPC: MPC-HC+LAV+MadVR 1080p OK ==> 13W power under WIn7

 

Have you read some of the reports from John Swenson in this forum about his new Linear PSUs for the Mac Mini

 

Hi David - Buy a Mac Mini and wait for Alex C. and John S.'s new power supply.

600w PSU. ATX intel board i5 integrated graphics, 5 mixed hard drives acting as a server with networked AV and direct connection for high end desktop listening station.

 

Yes EMI/RFI all affect sound quality including hiss and hum in your speakers.

 

I have the best solution for you.Underclock your computer. <snip> If you have a large dedicated video card in the system, remove it and use a little dinky one. Minimal number of peripherals and low power consumption will help minimize noise in your computer system. Even if you use an external USB DAC you still need to do this. They are all interlinked.

 

I am led to believe from the Jriver forum that a minimum of a nVidia GTX660 is required for full madvr rendering.It may be even more than this with recent additions to Madvr. I fancy the new Hasswells for their low power consumption but their onboard graphics don't come with Iris pro (5200, GT3e).Even the GT3e may struggle with Madvr full capacity but would be at least a compromise. I came across the GIGABYTE BRIX Pro: A First Look at the Intel i7-4770R with Iris Pro HD 5200 but figured Id rather do my own thing and include a video card.

 

I am always going to be at odds with building a dual high end video server competing with the lesser requirements of an audio design like caps or other only requiring small fanless clean power supllies, no cooling fans,spinning discs, noisy graphics cards and the like. Again it gets back to my goal or question, can the beast be tamed to isolate off/filter/protect the sensitive audio bits.

 

 

And finally, go get da power cables. Upgrade to high end. This is the most effective upgrade in this universe. Even the aliens will tell you to upgrade power cables. Depending on what model you choose and what brand you get it could cost anywhere from $500~ $2 million. You just gotta look around and research and research and research. No don't use power conditioners. They hamper sound quality. Connect the high performance power cables directly to the wall outlet if you want better sound.

bunny

 

I basically agree with you on your recommendations for using a good power cord.<snip>As far as power conditioning I have tried a Tripp lite 1200 watt power conditioner not the battery kind but the transformer ones and it was thin and lacked bass. I use a 3 large power conditioner for my home stereo and they work great. They weigh over 120 pounds each and run off 220 volts and live under the house as they hum. I have my PC plugged into one circuit which has 2500 watts available which helps the sound a lot.

.

 

 

 

 

Yep, I was mad enough in the nineties to buy some very expensive power cables, run 20 amp lines to gryphon monoblocks etc. I also got some custom made power leads with cheap but industrial grade shielding. I believe I heard distinct differences in power cables (I confess, I confess). These days semi retired and slightly less mad...but still have the power cables hanging around so why not use them, eh. Also have a giant power mains conditioner from the nineties (transformer type I think)...the one that improves sound not destroys it coz it was expensive (HA HA)...I can't say I can hear a difference. Some say that the power treatment in things like gryphon monoblocks with huge toroidal? something or others should not be plugged into power conditioners…...they are plugged into dedicated 20 amp lines instead. As you can see my audiophile nervosa had me in its grip in the nineties, lol.

[sandyk]

Yep, I was mad enough in the nineties to buy some very expensive power cables, run 20 amp lines to gryphon monoblocks etc. I also got some custom made power leads with cheap but industrial grade shielding. I believe I heard distinct differences in power cables (I confess, I confess). These days semi retired and slightly less mad...but still have the power cables hanging around so why not use them, eh. Also have a giant power mains conditioner from the nineties (transformer type I think)...the one that improves sound not destroys it coz it was expensive (HA HA)...I can't say I can hear a difference. Some say that the power treatment in things like gryphon monoblocks with huge toroidal? something or others should not be plugged into power conditioners…...they are plugged into dedicated 20 amp lines instead. As you can see my audiophile nervosa had me in its grip in the nineties, lol.

 

Hi David

I suspect that those measures may be far more worthwhile for our American cousins due to their lower mains voltage, and the resulting much higher current draw.

 

Regards

Alex