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

[flkin]

15 hours ago, Nenon said:

Let's talk about the SR7.

Yes, SR7 is unobtainable. I have never had a chance to try one. And I probably would never do. Sometimes I wonder if the SR7 is really that good or it's the fact that people can't get it that makes it so desirable to the point that it became legendary. There is a very small number of those power supplies ever made, and yet so many people talk about them. I bet most of those people, just like me, have never heard one in their system. 

 

So let's not split this into a thread of people with SR7's and everyone else (aka mortals). And let's not say that the only alternative for the mortals is a NUC solution. 

 

There are other power supply alternatives. Sean Jacobs at Custom HiFi Cables is a good example. He has learned a lot from his work with Innuos. In my opinion, the way the power supply on the Innuos Statement was designed might be even a better approach than using a SR7 with long umbilical cords. Here is a good article explaining the distributed design on the Statement:

http://www.the-ear.net/how-to/power-supply-design-innuos-statement

 

I am OCD when it comes to that stuff. I have a Sean Jacobs power supply. It sounds a lot better than the sBooster, LPS1.2, etc. in my system no matter what I power with it. A huge excellent quality oversized (400VA) toroidal transformer is a key part of it. I have also gone the Statement distributed power supply approach with the transformer and the recap modules in one chassis and the regulators in another chassis with the motherboard. I have 4 separate rails going to the motherboard (3 for the ATX connector and one for the CPU). My cables from the regulators to the motherboard are between 1'' and 3''. They are 15.5AWG Mundorf Silver/Gold - https://www.partsconnexion.com/MUNDORF-72180.html. Expensive but really good. Every wire has the JSSG360 shielding, even those that are 2'' long. Yes that requires soldering, but attention to every detail pays out. I have heard the difference between having a long wire from the regulators to the motherboard and a short wire. I can also hear a difference between the different types of wire I use there, no matter how short they are. Amazing stuff, and to be honest if someone told me those little things make a difference I would not believe that. 

 

So, the question is if someone can give me a SR7 with two rails, would I replace what I have now? I would have to install a Pico DC to DC ATX converter  and use one rail for the ATX. And another rail for the EPS connector. And use 1 meter long cables for that. I don't know. Actually, I highly doubt it. There are too many compromises in such a solution - the long umbilical cords, the shared DC rails, less DC isolation, all that cross-contamination, the DC to DC converter, etc. I would certainly try it if I had that chance. But my point is that the SR7 is not a perfect solution either. The transformer in the SR7 is half the size of the transformer I use. And no, I did not just get a big transformer from eBay, but an expensive custom one made to order with every trick that can help reduce noise and lower the impedance. The Mundorf capacitors Sean uses look a lot bigger and better quality than those on the SR7. Yes, the SR7 uses a magical custom regulator, that is supposedly the best in the world, but can that make up for all the other compromises taken here? Maybe it can, maybe it can't; I don't know. 

 

Sean Jacobs is just one solution out there. I am sure there are more. It just happened that's the one I am familiar with. I suggest instead of splitting this thread between mortals and immortals we go out there and explore the current solutions on the market, try, test, listen, and report back our findings. 

 

 

Paul Hynes and Sean Jacobs uses the same transformer supplier - Terry Monaghan from UK

[guiltyboxswapper]

These posts genuinely beg for a "reference LPSU that can be delivered within a reasonable timeframe" thread.  Lets face it, most if not all gains come simply from high quality power.

 

I have an Farad Super3 on the way to me, so can report on that though haven't a foggiest how it compares to the "chosen ones" SR4 / SR7s.

[Confused]

18 minutes ago, guiltyboxswapper said:

These posts genuinely beg for a "reference LPSU that can be delivered within a reasonable timeframe" thread.  Lets face it, most if not all gains come simply from high quality power.

 

I have an Farad Super3 on the way to me, so can report on that though haven't a foggiest how it compares to the "chosen ones" SR4 / SR7s.

We did have this one a while ago.  It never really got anywhere, but maybe it would be good resurrect it in hope of a Sean Jacobs versus SR-7 shoot out or something.

 

 

[LTG2010]

15 hours ago, sandyk said:

Are you able to post an internal photo of Sean's PSU with the 400VA toroidal transformer ?

My Supply has 2 smaller transformers and the regulators are all onboard in one box, with a single umbilical. @Nenon's supply is based on the Innuos Statement, which has outbox DC regulation,  situated next to the motherboard and dual umbilical for shorter signal paths, one large transformer, I guess the 400VA version has an even larger transformer to take care of the 5A 12V rail required by the i9 9900K, (the statement only needs 1.5A) you can see the 'Statement' supply below to get an idea of quality:

http://www.the-ear.net/how-to/power-supply-design-innuos-statement

 

[Superdad]

3 hours ago, guiltyboxswapper said:

These posts genuinely beg for a "reference LPSU that can be delivered within a reasonable timeframe" thread.

 

Well we build and deliver about 25 JS-2s each month.  Choke-filtered, dual-rail, user-adjustable output voltages, true 7.4 amp continuous (@12V). Even ranked by @austinpop as performing above SR4 on a actual audio component:

https://audiophilestyle.com/ca/contributors/my-quest-for-a-new-dac-part-1-the-mytek-brooklyn-dac-with-uptone-js-2-power-supply-r716/

“What impact did an external PSU make on the sound quality of the Brooklyn+?

 

I had on hand 3 external PSUs that could power the Brooklyn+ adequately at 12V DC. After trying them all, along with the built in SMPS (switching mode power supply), here is my ranking, in terms of decreasing sound quality:

1. Uptone Audio JS-2
2. Paul Hynes SR-4
3. SOtM sPS-500
4. built-in SMPS

I had done a similar comparison of PSUs powering the SOtM tX-USBultra, and found the SR-4 delivered the best sound quality. While the SR-4 had really shone on the tX-USBultra, on the Brooklyn DAC+, it was the JS-2 that pulled ahead, and by quite a margin. Compared to the JS-2, the SR-4 (and the sPS-500) seemed to lack dynamism, and sounded polite and unexciting.”

 

And I never request payment earlier than one week from a guaranteed shipment date.  

 

By the way, I’d put the nice (and completely silent) R-core trans we use up against any toroidal any day. The difference is obvious right away.  

 

But sorry, despite many requests, we have no plans to produce a big multi-rail ATX supply or to offer custom build services.  

[Chopin75]

11 minutes ago, Superdad said:

 

Well we build and deliver about 25 JS-2s each month.  Choke-filtered, dual-rail, user-adjustable output voltages, true 7.4 amp continuous (@12V). Even ranked by @austinpop as performing above SR4 on a actual audio component:

https://audiophilestyle.com/ca/contributors/my-quest-for-a-new-dac-part-1-the-mytek-brooklyn-dac-with-uptone-js-2-power-supply-r716/

“What impact did an external PSU make on the sound quality of the Brooklyn+?

 

I had on hand 3 external PSUs that could power the Brooklyn+ adequately at 12V DC. After trying them all, along with the built in SMPS (switching mode power supply), here is my ranking, in terms of decreasing sound quality:

1. Uptone Audio JS-2
2. Paul Hynes SR-4
3. SOtM sPS-500
4. built-in SMPS

I had done a similar comparison of PSUs powering the SOtM tX-USBultra, and found the SR-4 delivered the best sound quality. While the SR-4 had really shone on the tX-USBultra, on the Brooklyn DAC+, it was the JS-2 that pulled ahead, and by quite a margin. Compared to the JS-2, the SR-4 (and the sPS-500) seemed to lack dynamism, and sounded polite and unexciting.”

 

And I never request payment earlier than one week from a guaranteed shipment date.  

 

By the way, I’d put the nice (and completely silent) R-core trans we use up against any toroidal any day. The difference is obvious right away.  

 

But sorry, despite many requests, we have no plans to produce a big multi-rail ATX supply or to offer custom build services.  

Looks like i may be able to use the JS-2. Now how does it compare with HDplex 200W if only single rail is used ?

[rickca]

I'm considering a JS-2 to power the EPS12V 8-pin on a motherboard.  I'd be surprised if it didn't outperform a 12V rail from an HDPLEX 400W ATX LPS.  I have no idea how it would compare to a Sean Jacobs DC-3 supply with Mundorf caps and the 5A high current boost module which prices out around 1000 euros.  This would be a very useful comparison.

 

An interesting issue.  Many of the latest gaming motherboards which people are considering due to good VRM etc. now have two 8-pin connectors because they're designed for overclocking.  How do you handle such a thing?  Do you use two 12V rails with half the amps for each connector or what?  I'm not sure whether such motherboards are stable if you only use one of the two 8-pin connectors.

 

What is more sensitive to the quality of LPS ... the 24-pin on a motherboard or the EPS12V?  I'm thinking of using a separate LPS for each of these.

[bobfa]

17 hours ago, austinpop said:

Now, let's get back on-topic to listening experiences again. 

PLEASE and Thank You.

 

I have been comparing Roon +StylusSQ and Stylus against each other on my Xeon.  I have to start some processor speed limit testing next.  I still feel that Stylus sounds better.  There is more dynamics and there is more detail.  I am going to get some other folks over to listen!

 

My son likes Stylus more for the simplicity of using their web interface and the speed of the UI.  I still like Roon for their look and feel.  

 

I will also bring the NUC back up and listen to it I have to validate the differences again.

 

NOTE: using the HDPLEX 400 and an FMC to the endpoint.  I am not trying to change network speed due to FMC.  The Xeon now has an Element-X  USB card.

[guiltyboxswapper]

1 hour ago, bobfa said:

PLEASE and Thank You.

 

I have been comparing Roon +StylusSQ and Stylus against each other on my Xeon.  I have to start some processor speed limit testing next.  I still feel that Stylus sounds better.  There is more dynamics and there is more detail.  I am going to get some other folks over to listen!

 

My son likes Stylus more for the simplicity of using their web interface and the speed of the UI.  I still like Roon for their look and feel.  

 

I will also bring the NUC back up and listen to it I have to validate the differences again.

 

NOTE: using the HDPLEX 400 and an FMC to the endpoint.  I am not trying to change network speed due to FMC.  The Xeon now has an Element-X  USB card.

 

Do you mean Element M card?  I got one of these to try in my Intel NUC, but it requires 12v and 3.3v from the PCIE bus.  The adapter I tried provided 12v.  Will try the JCAT in time (no power from PCIE bus required).  

[rickca]

2 minutes ago, guiltyboxswapper said:

Do you mean Element M card?

Matrix Element M and Element X are streamer/DACs.  I think @bobfa means the Element H, which is the latest Matrix USB card.

[guiltyboxswapper]

2 minutes ago, rickca said:

Matrix Element M and Element X are streamer/DACs.  I think @bobfa means the Element H, which is the latest Matrix USB card.

 

Apologies I have the Element H myself (ready to go back to the seller) due to it requiring PCIE bus voltage (as noted above).  In bobfa's case may not be a big deal to be honest given system is already powered by LPSUs, though I'd prefer to isolate power from system with the JCAT card as mentioned.

 

I confirmed this by setting external power and connecting a 12v DC feed, but disconnecting the PCIE 12v feed on the adapter for the NUC; would not power up at all.  Shame.

[bobfa]

1 hour ago, bobfa said:

NOTE: using the HDPLEX 400 and an FMC to the endpoint.  I am not trying to change network speed due to FMC.  The Xeon now has an Element-X  USB card.

I cannot type : Element-H USB card.  As a side note I got an email blast that Element stuff is available at Moon-Audio in the USA.  I will report on any notes over the USB card on the thread about USB cards.

[rickca]

1 minute ago, guiltyboxswapper said:

I have the Element H myself

So how did it sound, and what did you compare it to?  It's around $250, so about half the price of the JCAT USB Femto card.

[bobfa]

2 minutes ago, rickca said:

So how did it sound, and what did you compare it to?  It's around $250, so about half the price of the JCAT USB Femto card.

There are a few posts about trying to get a comparison done over here:

 

 

 

 

[guiltyboxswapper]

8 minutes ago, rickca said:

So how did it sound, and what did you compare it to?  It's around $250, so about half the price of the JCAT USB Femto card.

 

TBH I'm not going to even try.  PC "server" is not fully optimised, will require moving into a different room (and dirtying up the rest of the power feed on Hifi).  From past experience of temporarily powering up a JCAT on ATX internal power, it was ... ugly.  External was night and day.

 

If anyone can confirm the Element H can be powered externally only i will happily try it... but I don't think you can.

[Nenon]

2 hours ago, LTG2010 said:

@Nenon's supply is based on the Innuos Statement, which has outbox DC regulation,  situated next to the motherboard and dual umbilical for shorter signal paths, one large transformer,

 

Let me clarify something. Sean did not do that. It was a DIY project I did. I used his regulators and recap modules and placed them the way I described (similarly to the way it's done on the Statement). I ordered a large transformer by his specs for this project. I wanted it to be massively oversized and the best build quality. 

As far as I am aware Sean does not offer what I have done. It's a tedious job to mount and wire everything. It's not that complex but takes a lot of time to do.

 

1 hour ago, rickca said:

Many of the latest gaming motherboards which people are considering due to good VRM etc. now have two 8-pin connectors because they're designed for overclocking.  How do you handle such a thing?  Do you use two 12V rails with half the amps for each connector or what?  I'm not sure whether such motherboards are stable if you only use one of the two 8-pin connectors.

 

I use one 12V rail. Although you have an 8-pin connector, four of the pins are the DC+ and the other four are the DC-. I mean they are paralleled on the motherboard and are not separate inputs. At least that's how they are on my Asrock gaming motherboard. I think the reason they use 8-pins instead of 4 is to end up with a higher gauge wire when you combine four parallel wires from the power supply to the motherboard.

 

2 hours ago, Superdad said:

By the way, I’d put the nice (and completely silent) R-core trans we use up against any toroidal any day. The difference is obvious right away.

 

I respect you and John a lot, just from the posts I have read. And I would never even try to go against anything John says about that stuff. Not to argue (don't really want to do that), but to provide another perspective. I had an argument about a R-core transformer (that looks very similar to yours on the outside) with a DAC manufacturer. And I tried his transformer. To make a long story short, his R-core transformer worked better on his DAC, but not on the computer (seperated in a different chassis like I have described earlier). One of the differences was that while his transformer was oversized, it was still much smaller than the toroidal I used. And the toroidal I used was much more expensive, had electrostatic shield, GOSS band, and every trick in the book to reduce noise (some of which might have been done on the R-core too). This does not tell ANYTHING about YOUR "R-core" transformer, but it convinced me as an end user (not a manufacturer) not to look back. Also, given that the power supply I use took about 3-4 months to completely break-in, one probably needs about 6 months of parallel running to determine the true performance of each transformer. I did not spend 6 months on this test, so take my word with a grain of salt. But for me, I am sticking to the oversized toroidal. 

 

[rickca]

19 minutes ago, Nenon said:

I use one 12V rail. Although you have an 8-pin connector, four of the pins are the DC+ and the other four are the DC-. I mean they are paralleled on the motherboard and are not separate inputs. At least that's how they are on my Asrock gaming motherboard. I think the reason they use 8-pins instead of 4 is to end up with a higher gauge wire when you combine four parallel wires from the power supply to the motherboard.

You misunderstood me.  The motherboards I'm talking about have dual 8-pin EPS12V connectors or one 8-pin plus one 4-pin, not just one 8-pin connector with 4x12V pins and 4xground.  For example, ASRock Phantom Gaming 9 Z390 has both an 8-pin and a 4-pin.

 

Some of the latest PC SMPS have two cables each with a 4+4 connector for this situation.  They may have a single 12V rail design.  For example, Seasonic Prime Ultra Titanium 1000W.

[Superdad]

25 minutes ago, Nenon said:

One of the differences was that while his transformer was oversized, it was still much smaller than the toroidal I used. 

 

Of course.  I was just referring to an apples-to-apples comparison of similar VA-rated toroid versus R-core—swapped into the same supply. I did just that during development of the JS-2 and it was not even close.

[LTG2010]

28 minutes ago, Nenon said:

 ordered a large transformer by his specs for this project. I wanted it to be massively oversized and the best build quality. 

You used your own transformer?

[dctom]

I have been following the thread for a while and  took the plunge a couple of months ago.

To start with, my original diy fanless Asus mini-ITX was  utilised as a server feeding a 7i7dnbe in Asaka plato case. Both machines running on Ubuntu, the ITX directly linked to the NUC.

 

This immediately produced an uplift in SQ. Replacing the USB  HDD externally connected to the Asus with a Synology NAS was another improvement. Next came the HDPLEX 200w. The Plex currently powers the NUC at 19v, the NAS 12v, DP link switch 9v, and ITX ssd 5v with CANARE 4S6s  dc cables. This has seen another increase in SQ.

 

As the psu  has proved so beneficial I have ordered a Sean Jacobs 5amp DC3  and Hynes SR 4 to, hopefully raise the bar further.

 

It has been something of a journey trying to improve on the linux ITX. I borrowed an Aurilac G1 and although it seemed slightly more transparent, initially  with my external USB HDD connected. However it stopped recognising the HDD and would only work with my ITX acting as a server connecting to the G1 via ethernet. In this mode it sounded worse than the just the ITX  + exHD connected to my DAC.

 

The current set up has left the G1 in the dust suffice to say.

 

With the power supply seemingly ever more critical, I have replaced the RCBOs in the dedicated consumer unit, which supplies the HiFi system, with better quality ones. This has brought a subtle but noticeable improvement in the already excellent clarity, weight and bass.  Obviously  all the components in the system have benefited not just the server set up.

 

[RickyV]

52 minutes ago, Superdad said:

 

Of course.  I was just referring to an apples-to-apples comparison of similar VA-rated toroid versus R-core—swapped into the same supply. I did just that during development of the JS-2 and it was not even close.

Hi @Superdad

Can you tell a little bit of those differences, was that electricly or in sound quality?

i am asking because I have now cheap china r-core and I am waiting on a Lundahl 4x20V 2A, C-core transformer so I curious about the differences. 

[seeteeyou]

And then there's also massive overkill stuff like ROG Dominus Extreme or C621 AORUS XTREME with dual 24-pin ATX and quad 8-pin EPS, we could throw in Xeon Silver 4214 (85W TDP) since that's officially supported by the latest BIOS version

 

https://www.asus.com/us/Motherboards/ROG-Dominus-Extreme/HelpDesk_CPU/

https://www.gigabyte.com/Motherboard/C621-AORUS-XTREME-rev-10#support-cpu

https://www.provantage.com/intel-bx806954214~7ITEP6FK.htm

https://www.provantage.com/intel-cd8069504212601~7ITEP6HQ.htm

https://www.wiredzone.com/intel-components-processors-server-cd8069504212601-10030032

 

At some point they might even add Xeon Silver 4215 to the supported list and maybe Optane DC Persistent Memory Modules could get a chance?

 

Fanless chassis for E-ATX motherboards here

 

http://www.turemetal.com/product_en.html

https://item.taobao.com/item.htm?id=590791895684

http://www.fanlesstech.com/2019/05/upcoming-turemetal-cases.html

https://www.anandtech.com/show/14562/turemetal-fanless-chassis-for-diy-highend-pcs

https://www.todotech20.com/un-chasis-super-sin-ventilador-de-turemetal-para-bricolaje-y-0-estaciones-de-trabajo-db/

http://www.comptoir-hardware.com/actus/boitiers-barebones/39256-turemetal-le-reve-dun-boitier-diy-100-fanless-bientot-a-portee-de-main-.html

 

In that case even 85W Xeon Silver could get away with 4 × LT3045 × 10

 

https://www.mpaudio.net/power-supplies

[Nenon]

7 hours ago, rickca said:

I'd be surprised if it didn't outperform a 12V rail from an HDPLEX 400W ATX LPS.

me too

 

4 hours ago, rickca said:

You misunderstood me.  The motherboards I'm talking about have dual 8-pin EPS12V connectors or one 8-pin plus one 4-pin, not just one 8-pin connector with 4x12V pins and 4xground.  For example, ASRock Phantom Gaming 9 Z390 has both an 8-pin and a 4-pin.

Yes, sorry, I misread your message. Does it say anything in the manual what each is used for? I haven't used a motherboard like that yet. Have you checked with a multimeter if the DC+ on one of the EPS connector is connected to the DC+ on the other? It seems like you may need (benefit from) 5 rails for a motherboard like that.

 

4 hours ago, LTG2010 said:

You used your own transformer?

Yes. 

 

4 hours ago, Superdad said:

Of course.  I was just referring to an apples-to-apples comparison of similar VA-rated toroid versus R-core—swapped into the same supply. I did just that during development of the JS-2 and it was not even close.

Good to know. For my next project I may try multiple R-core transformers (one dedicated for each rail). Sounds like that would be a killer LPS. We, DIY-ers, can afford one-off exotics like that from time to time. It would be more difficult in a commercial solution where so many things need considering. 

 

 

[austinpop]

20 hours ago, Nenon said:

 

Yes, you can buy a custom LPS from Sean Jacobs. That's what he does. 

 

If someone's looking for a no-compromises ATX linear PSU, this could be one of the best options. Sean had priced one out for me a few months ago, and I think it would have been a bit under £3000, which is a very fair price, given you're getting a 4-rail (ATX - 3.3/5/12V & EPS - 12V) high-current linear PSU.

[austinpop]

On 6/22/2019 at 2:01 PM, limniscate said:

Rest assured that the Farad will be tested again against my SR7 soon.

 

Great teaser post from Eric!

 

I have acquired a Farad Super3 12V PSU from Mattijs, and am burning it in. Given how long it took for my SR-4 to fully burn in, and the level of interest I'm sure there is for this comparison, please be patient.  

 

Once I feel the Farad is fully burned in, Eric and I will conduct a PSU shootout between the Farad Super3, the LPS-1.2, the SR-4, and the mighty SR-7, all set at 12V.

 

It's an exciting time to be an, um, LPSU fan.