Jump to content
IGNORED

A novel way to massively improve the SQ of computer audio streaming


Message added by The Computer Audiophile

Important and useful information about this thread

Posting guidelines

History and index of useful posts

Most important: please realize this thread is about bleeding edge experimentation and discovery. No one has The Answer™. If you are not into tweaking, just know that you can have a musically satisfying system without doing any of the nutty things we do here.

Recommended Posts

6 hours ago, jean-michel6 said:

I am extremely interested to know if you did compare powering your server with the 4pin atx only and 4 pin +24 pin atx. 

 As per my previous reply to you I could not use the 12V - 4 pin alone due to the low Ampage but what I could do is test the 24 pin connection alone without the 4 pin, therefore omitting the 12V 'CPU rail'. I could hear a reduction in quality, less seperation and a slightly more grainier sound, not huge but noticeable.

Link to comment
2 hours ago, Johnseye said:

 

It was immediately obvious, especially in the low end.  I'll let @austinpop, @lmitche or @Forehaven share their opinion.  This was powering the sMS-200ultra.  I am hoping Rajiv and Eric's tests do not confirm what we heard, but if they do I won't be buying a multi-rail PSU again.  I will also be able to duplicate the test with the SR7 and sPS-500, but those supplies are probably too different to draw a true conclusion whether the difference is attributed to the shared transformer.

 

Rajiv is right in that is was fast, but for me, it was immediately noticeable.  Better bass and sl. more resolving...quiet?  Either way, I too would only stick with mono rails.  The last few mos. have shown to me how important and quierky clean power can be no matter where it's applied.

Ryzen 7 2700 PC Server, NUC7CJYH w. 4G Apacer RAM as Renderer/LPS 1.2 - IsoRegen/LPS-1/.2 - Singxer SU-1/LPS1.2 - Holo Spring Level 3 DAC - LTA MicroZOTL MZ2 - Modwright KWA 150 Signature Amp - Tidal Audio Piano's.  

.

Link to comment
5 hours ago, LTG2010 said:

 As per my previous reply to you I could not use the 12V - 4 pin alone due to the low Ampage but what I could do is test the 24 pin connection alone without the 4 pin, therefore omitting the 12V 'CPU rail'. I could hear a reduction in quality, less seperation and a slightly more grainier sound, not huge but noticeable.

Thank’s for your detailed answer and take the time to test only the 24pin ATX ps.

On a previous and different mb set up I was powering separately the cpu and the 24pin atx connector (pico psu) and it did provide a good sound quality improvement over only using the 24atx power connector.

On my set up the mother board is not powering any accessory except the mother board itself (as your set up I think).

Before using the Sean Jacobs psu I was using the ifi 12v 1.8A smps with no problem. 

 

I agree with you that to have a dedicated psu for each part of our set up is always a move for better sound. 

For the sotm trio I have 

one sps 500 for the pcie USB ex

one sps 500 for both the sCLKex and that txUSBin

I have had also very good results with uptone lps 1.1 which gave better results on sotm trio than the sps 500 .

I plan to try the new lps 1.2 which seems to be recognized as clearly superior to the lps1.1. 

And as you suggest I will try with 3 separate psu for each piece of the set up. 

 

For the mother board I am sure it is also a good move to use the 24pin atx as you did but it is another layer of complexity on a system already quite evolved. I may consider this for future trial. 

 

By by the way going to a dual pc set up ( roon core , roon bridge  Fedora ) was a very big gap in music improvement. 

I have tried in my system sms200ultra with one pc as server and dual pc was clearly superior. 

Those Supermicro board are truly excellent for audio. 

 

  

PCserver Supermicro X11SAA under Daphile  ,Jcat pcie net card ,Etherregen,e-red dock endpoint,powered by LPS 1.2 , SPS 500 , Sean Jacobs level 3 psu,  DAC Audiomat Maestro 3, Nagra Classic Amp , Hattor passive preamplifier , Martin Logan montis

Link to comment

First of all, it's great to have Nuno from Innous joining us on this thread and thank you very much for your participation.

 

We should be able to understand if you weren't too happy about revealing "too many" details regarding that custom motherboard inside Statement. However, would you please kindly tell us just a little bit about bypassing the regulators of all motherboards in general?

 

For instance, we could feed 1.3V directly to RAM slots with the best linear regulators (i.e. 6 pieces of LT3045 @ 3A) with modifications while that seemed to result in a significant improvement

 

http://www.diyaudio.com/forums/vendor-s-bazaar/310402-fs-ultra-low-noise-power-supply-lt3045-based-pcb-6.html#post5353416

 

Similarly, feeding 5V directly to this PCI-Express card with LT3045 also yielded great results without having to replace the original system clocks with OCXO etc.

 

https://www.startech.com/Cards-Adapters/USB-3.0/Cards/4-port-usb-3-1-card~PEXUSB314A2V

Now it's only "natural" to ask the question, how far do we wanna go in terms of bypassing the regulators? While LT3045 should be fairly affordable to begin with, we still have to think about what do we put in front of LT3045.

Besides, is it actually feasible to feed the right voltage directly to something like Celeron J1900 inside X10SBA-L since the voltage is VERY low (only 1.0V or so) while the amperage should be fairly high.

Last but not least, is it OK to discuss more about measuring the noise pattern of so many motherboards with a detailed oscilloscope? Was X10SBA-L really THAT outstanding or did you find anything that's reasonably close? I'm just wondering if its big brother X11SBA-LN4F were also a decent choice or otherwise

https://www.supermicro.com/products/motherboard/X11/X11SBA-LN4F.cfm

Thanks again for chiming in and looking forward to your answers.

Link to comment
4 hours ago, nvitorino said:

The interesting part is that it still uses the 12V rail on both the ATX and the 12V input separately if the 12V input is available.

If both the 4-pin and 24-pin are supplying 12V, which components are being powered by each power supply? 

Pareto Audio AMD 7700 Server --> Berkeley Alpha USB --> Jeff Rowland Aeris --> Jeff Rowland 625 S2 --> Focal Utopia 3 Diablos with 2 x Focal Electra SW 1000 BE subs

 

i7-6700K/Windows 10  --> EVGA Nu Audio Card --> Focal CMS50's 

Link to comment
4 hours ago, nvitorino said:

 

Hi everyone,

 

I can provide a bit of feedback on this based on our experience with the Supermicro X10SBA-L. On this board there is a significant difference in sound quality when powering both the ATX and the 12V Input compared to 12V alone. If you connect the ATX supply, the board will not use it's own switching regulators to convert to 5V and 3.3V from 12V. The interesting part is that it still uses the 12V rail on both the ATX and the 12V input separately if the 12V input is available. We've tested 3 configurations:

 

1. ATX only (using a Pico-PSU with a clean 12V input)

2. 12V only

3. ATX + 12V

 

1. and 2. sound similar in our tests. 3 sounds significantly better. At first, we were expecting 2. to be the best, as it removed the Pico-PSU from the equation. But seems the board itself will do the same job as the Pico and doesn't seem to do it any better than that.

We've fthen further tested with separate rails using the ATX only and compared to using separate rails for ATX + 12V input. The result is the same, adding both ATX and 12V results is clearly superior sound with this board. If you don't want / can't get separate rails for ATX and 12V, getting at least another 12V supply and connecting a Pico-PSU will still yield benefits compared to feeding the 12V alone. If you are using a separate power supply for the SSD, 1.5A on ATX and 1.5A on 12V is enough for this board.

 

Nuno
 

Thank you for this very I teresting comment . Me too in my dual pc set up  I have used only the 12v 4pin connector as I felt it will provide the best sound quality since it removed the noisy pico psu from the system.  

 

Apparently i was wrong I will definitely try the dual 12vas I have spare picopsu available. 

 

Does going to a separate 3 linear psu to power the 24 pin atx provide a sizable progress over using the pico psu ?

PCserver Supermicro X11SAA under Daphile  ,Jcat pcie net card ,Etherregen,e-red dock endpoint,powered by LPS 1.2 , SPS 500 , Sean Jacobs level 3 psu,  DAC Audiomat Maestro 3, Nagra Classic Amp , Hattor passive preamplifier , Martin Logan montis

Link to comment
4 hours ago, nvitorino said:

The result is the same, adding both ATX and 12V results is clearly superior sound with this board. If you don't want / can't get separate rails for ATX and 12V, getting at least another 12V supply and connecting a Pico-PSU will still yield benefits compared to feeding the 12V alone.

Can you provide any photos, i dont understand, you connected some ATX PSU to 24pin and at the same time some another PSU to 12V mb plug, like this:

5ae1e1e4c2bc3_ScreenShot2018-04-26at16_27_36.jpg.f1975699651b296acafe1992e5ec0d35.jpg

 

dCS Network Bridge | Audio Note DAC2 Signature | Audio Note M5 Preamp | Audio Note Empress Silver Monoblocks | Audio Note AN-E/Spe HE Speakers

Link to comment
16 hours ago, austinpop said:

Once Eric's SR-7 is here, we'll compare.

 

In light of @nvitorino's post, - thanks for your input, Nuno! - and the fact that @paulhynes is even willing to build multi-rail SR-7 supplies, I feel this topic (of cross-rail interference, if we can call it that) needs to be examined more carefully and deeply.

 

I believe @limniscate's SR-7 is (will be) a dual-rail supply, with each rail rated at 6A (called HD in Paul's spec sheet). I assume Paul sizes the transformer appropriately to accommodate both rails, and makes engineering assumptions of capacity, i.e. simultaneous load on both rails. In addition, I have an SR-4.

 

To replicate the experiment we did at @Johnseye's, we would baseline his system, powering 2 components (tX-USBultra and dX-USB HD Ultra):

  1. first, with 2 rails of the SR-7
  2. next, one powered by a single SR-7 rail, and the other with the SR-4.

Both components' current demands will be under 1A, just like in the experiment in Chicago.

 

Now we just need to hurry up and wait for Paul to get to #10 on his build list. :)

Link to comment

Would it be advantageous to remove the graphics card in my wkst that is headless and running ROON server/HQP (which is located in laundry room)?  Wouldn't appear to provide any real benefit as well as being power hungry (Quadro 4000). 

 

Not sure if the OS (WIn10 Pro) would require it to boot and or if it is even possible to disable it in the BIOS.  I'm pretty sure there is no onboard graphics either so if I had boot issues, I would have to reinstall "a" graphics card.  I may have a cheapo graphics board lying around I could use but just thought I would ask if anyone else does this.

My rig

 

Link to comment
2 hours ago, paulhynes said:

I have to admit I am impressed by the tenacity of purpose displayed by CA members in searching out the highest level of performance, and I also have to say that I think your heroic efforts are having a profound effect encouraging manufacturers to raise their game. Once a topic is raised there is a concerted effort to understand the topic more fully and to put the resulting conclusions into practice.

 

For those currently investigating power supply issues here is some information you may find useful.

 

The articles I wrote for Common Ground magazine during the mid 1990s explain why separate power supplies are required for each individual stage of signal processing for the finest performance. Yes I did say each individual stage of signal processing and not each individual item of equipment. Breaking this down to its logical conclusion does require a large number of power supplies in the average system. If you have an item of equipment with say 6 signal processing stages, the requirement is 6 separate power supplies to remove any power supply cross modulation between the processing stages. Scared yet? Think how many signal processing stages there are in your system. You might also consider that separate mains power feeds back to your consumer unit would also be a must to reduce mains interaction. For most enthusiasts this is about as far back on the power system as you could go without incurring a redesign of your local power grid.

 

I originally suggested in the articles that separate transformer windings were required for each power supply to provide galvanic isolation, and this is certainly an improvement over one common master power supply for all the stages, but further work on this proved that separate mains transformers were essential to avoid core interaction artifacts.

 

The Common Ground articles were written it three parts which Chris C kindly agreed to re-publish in the CA forum a while back. For those that have not read the articles I have added the link to the first part. The three articles are reproduced in three consecutive posts so scroll down to read them all.

 

https://www.computeraudiophile.com/forums/topic/28538-the-importance-of-power-supplies-generally/?page=2&tab=comments#comment-560512

 

To answer a few questions that have popped up since I last checked this thread.

 

The SR7MR Multirail power supplies were designed as a result of customer requests for high quality multiple power rails in a single chassis rather than having separate power supplies. This also reduced the overall cost as only one chassis was used and one mains transformer core. Of course there is a compromise running up to four rails on separate windings, on one mains transformer, in one chassis, but to follow the purist path gets expensive. And, just like the quality of design and implementation of any piece of equipment is important, the quality of the power supply is also important.

 

As an example, I designed the power supplies for a system commissioned by a client of Russ Andrews in the early 1990s. Each accessible signal processing stage in the system had its own power supply back to the separate windings on mains transformers to reduce power supply inter-modulation to as low a level as we could at that time. There were no regulator chips good enough back then for use in this system so all the regulator modules were my discrete component design that was the forerunner of the SR4 and SR7 regulator circuitry. All the resistors in the system were Vishay Bulk Foil and the printed circuit boards were manufactured from Teflon board. BHC slit foils were used throughout for energy storage. This system was tri-amped and the power amplifier output stages used six1000VA mains transformers. The power amplifier voltage gain stages each had their own 500VA mains transformer, as did all the stages in the preamplifiers, vinyl record deck and the analogue stages in the DAC on the laser vision player we used as the digital source. The mains transformers were custom manufactured on oversized cores of the highest quality.

 

This system was the centre of attention at one of the HiFi News audio shows at the Penta Hotel, London, and was regarded as best sound at the show by many. HiFi Reviewer Paul Messenger brought his own personal test records for audition and described the system as being all things to all men as it provided exemplary performance with all types of music. The system filled half of the demonstration room and cost around £80,000 back in the early 1990s. I guess you could quadruple that figure to reproduce the system nowadays. Whilst there are HiFi enthusiasts able to cope with such costs they are few and far between. So we manufacturers have to work in the real world for most people and the design brief for products is to provide the best performance we can for a given level of cost.

 

The regulator circuitry and component quality used in the SR4 is the same as used in the SR7. The SR4 also uses a balanced mains isolation transformer (smaller of course due to the finite chassis size) and the same amount of energy storage as used in the HD and EHD rails. This is why the performance of the SR4 and SR7 are close. The single rail SR7 will out perform an SR7MR Multirail but the cost escalates when you start separating the supplies into single rail power supplies.

 

I am currently finishing off orders 8 and 9 ready for shipment, Order 6 was delayed due to a specification error with one of the mains transformers. I have tried to contact you via e-mail, number 6, but my e-mail keeps being rejected by the mail system. Perhaps you have an alternative e-mail we can use. I am expecting a replacement mains transformer next week and will progress this as soon as the transformer arrives.

 

I will also be starting Limniscate’s order number 10 next week.

 

Regards to all,

Paul

 

Thank you for the information Paul, and confirming what we heard from your supplies.  This validation, combined with actually hearing the difference, demonstrates the impact of a supply beyond what I heard when I first introduced the SR7 into my system.

 

I think it's very important to note that you call out a benefit can be had by isolating the supply by stage, not just device.  This is something that a few people here in the forums have been doing with their PCs, resulting in noticeable improvements.  Everyone will have their cost/space thresholds, but at least for me, now a bar has been set.

 

 

Link to comment
2 hours ago, paulhynes said:

I have to admit I am impressed by the tenacity of purpose displayed by CA members in searching out the highest level of performance, and I also have to say that I think your heroic efforts are having a profound effect encouraging manufacturers to raise their game. Once a topic is raised there is a concerted effort to understand the topic more fully and to put the resulting conclusions into practice. 

 

Hi Paul,

 

Thanks for your gracious comments. Speaking for myself, that is indeed my primary objective - to get manufacturers to take note of, and analyze, what we here are reporting from our scientific, but admittedly empirical efforts. I greatly admire the DIYers who fearlessly mod their systems in their quest for quality, but I do feel the biggest impact comes from having manufacturers analyze these findings with the resources only they have at their disposal.

 

2 hours ago, paulhynes said:

 

For those currently investigating power supply issues here is some information you may find useful. 

 

 

Thanks for reminding us of these epic posts. Highly recommended reading!

Link to comment
On 4/24/2018 at 9:41 AM, Johnseye said:

 

Let me know if the sPS works.  I'm tempted to buy this Synology just because it can be powered externally.  My current NAS is an 8 bay with internal supply.

John, Qnap supports NIC bridging out of the box which I hear is a :

 

A novel way to massively improve the SQ of computer audio streaming.

Pareto Audio aka nuckleheadaudio

Link to comment
23 minutes ago, lmitche said:

John, Qnap supports NIC bridging out of the box which I hear is a :

 

A novel way to massively improve the SQ of computer audio streaming.

 

LOL, thanks Larry.

 

Since you've opened Pandora's box, I currently have an 8 bay Synology with internal SMPS.  This can be improved, which is why I was asking about the 12v.  I had no idea you could power a NAS with 12v, but it's a 2 bay array.  I'd probably only use it to store music for listening while I keep a backup copy on the Synology.

 

That said, is a 2 bay NAS feeding music over ethernet to the server, going to be better than a USB connected drive?  I have 8TB of music which will only grow, so that limits my options.

Link to comment
1 hour ago, Johnseye said:

 

LOL, thanks Larry.

 

Since you've opened Pandora's box, I currently have an 8 bay Synology with internal SMPS.  This can be improved, which is why I was asking about the 12v.  I had no idea you could power a NAS with 12v, but it's a 2 bay array.  I'd probably only use it to store music for listening while I keep a backup copy on the Synology.

 

That said, is a 2 bay NAS feeding music over ethernet to the server, going to be better than a USB connected drive?  I have 8TB of music which will only grow, so that limits my options.

John, having tried the NAS, SATA and USB connected drive options, I find the USB solution to sound best. With a USB hard disk enclosure one can control the 12 volt motor and 5 volt vbus to sata adapter power quality. I'd recommend this as the best approach with backups to the NAS. I have lt3045 modules on both power inputs to the enclosure, one 3 amp and the other 1 amp from USB card power injection.

Pareto Audio aka nuckleheadaudio

Link to comment
11 hours ago, paulhynes said:
Quote

I am currently finishing off orders 8 and 9 ready for shipment, Order 6 was delayed due to a specification error with one of the mains transformers. I have tried to contact you via e-mail, number 6, but my e-mail keeps being rejected by the mail system. Perhaps you have an alternative e-mail we can use. I am expecting a replacement mains transformer next week and will progress this as soon as the transformer arrives.

 

I will also be starting Limniscate’s order number 10 next week.

 

Regards to all,

Paul

 

Thanks Paul,

 

I am relieved to hear this as I saw #7 was delivered and I wondered where my #6 was!!

 

I"ll send you another email with an alternate address which might work better.

 

Cheers,

 

Geoff

 

PH SR7 > MacMini+Uptone MMK Mod > Audirvana 3.2 > re-clocked D-LInk switch/LPS1.1 > sMS-200Ultra/LPS1.2 > tX-USBUltra/PH SR7 > Chord BluDave > Focal Utopia(Norne Silver) or Voxativ 9.87/ Stereo REL G1 Mk II

Link to comment
10 hours ago, Johnseye said:

 

LOL, thanks Larry.

 

Since you've opened Pandora's box, I currently have an 8 bay Synology with internal SMPS.  This can be improved, which is why I was asking about the 12v.  I had no idea you could power a NAS with 12v, but it's a 2 bay array.  I'd probably only use it to store music for listening while I keep a backup copy on the Synology.

 

That said, is a 2 bay NAS feeding music over ethernet to the server, going to be better than a USB connected drive?  I have 8TB of music which will only grow, so that limits my options.

Probably. To my surprise the sound I get from Tidal through my sonic transporter i5 with Roon and an ultrarendu is far better than the one that comes from my HD connected through USB to the STi5...

Jensen VRD-iFF>Router>Rj45>opticalModule>
SFP>Buffalo2016>SFP>opticalModule >Rj45>

IZen Mk3>Rj45> Delock62619>Rj45>
etherRegen (Master Clock+ Mini-Circuits BLP)>SFP>opticalRendu>USB>IsoRegen>

USB>Phoenix>USB>OPPO 205 (Modded)>HMS “the Perfect Match”>Proac Tablette Reference 8 Signature.
 

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...