Building a PC to improve SMS-200 Ultra

[ismewor]

4 hours ago, Johnseye said:

  I know of a way to do it with either a PICO or HDPlex adapter

Thanks, Yes, i aware of this items couple years ago. I tried them both plus some. but the sound quality is not i'm looking for.

May be is my bad. I set the bar too high which it compare to my vinyl rig and extreme digital rig.

[LTG2010]

5 minutes ago, ismewor said:

The idea is True Linear ATX Power Supply with 9 rails output. And he replied with he can only done with 5 Rails and the cost is sky high 7k if i remember right. and this never happen.

Pink Faun's Streamers use linear ATX power supplies dont know how many rails but the cost is 7K for the top of the line Streamer and doubles for the 2 box solution.

[ismewor]

4 minutes ago, LTG2010 said:

Pink Faun's Streamers use linear ATX power supplies dont know how many rails but the cost is 7K for the top of the line Streamer and doubles for the 2 box solution.

Yes, He said that, and i also follow the SGM2015 server that Jord help to build. but he won't say detail about specification for his linear ATX design. And he only sell it with his server.

[GUTB]

Wtf are you guys talking about? HDPlex sells one with native ATX outputs for $800 and there's a TeraDak for $530 that uses an ATX adapter.

[ismewor]

2 minutes ago, GUTB said:

Wtf are you guys talking about? HDPlex sells one with native ATX outputs for $800 and there's a TeraDak for $530 that uses an ATX adapter.

Thanks, Yes, been there and done that. 

[LTG2010]

Just now, GUTB said:

Wtf are you guys talking about? HDPlex sells one with native ATX outputs for $800 and there's a TeraDak for $530 that uses an ATX adapter.

True but they are not in the class of Paul Hynes linear power supplies etc as recomended here by others who own them, unless they will build a 'specialist' one to a higher spec. Hence the recomendation to use a small factor motherboard with 12V dc power input.

[Johnseye]

13 minutes ago, GUTB said:

Wtf are you guys talking about? HDPlex sells one with native ATX outputs for $800 and there's a TeraDak for $530 that uses an ATX adapter.

 

Yes, I think the unspoken point was a very good LPSU.

[ElviaCaprice]

4 hours ago, LTG2010 said:

 

The SMS200/Ultra does close to what you describe when you connect a usb drive directly to it. It plays bit perfect files from the USB storage and 'bypasses' the ethernet connection. Software such as jriver are used simply for library management. I know roy described the bridge mode as superior for the ethernet connection but not sure if he tried the 'direct route'

The OP might want to try this connection before embarking on any new pc build to get the flavour of whats on offer.

There are obviously limitations as to software choice and use of dsp / convolution processing and no video, but it would give an idea of server to dac playback.

 

This has been tested by others and found not to equal the SQ results of streaming via Ethernet to the renderer.  Why?  Could be many reasons, software design and it's non flexibility with custom linux, inability to load to memory after decompressing the file or in case of WAV, ability to even load to memory or lack of memory? 

Point is, this headless server design you aiming at has it's deficiencies.  For a little bigger mobo but yet low power, you can defeat those deficiencies in a sCLK-EX server.

[bibo01]

8 hours ago, ismewor said:

Thanks, Yes, been there and done that. 

http://www.teradak.com/products/105.html

Is this what you used? Or a DC-DC adapter?

[TopQuark]

11 hours ago, LTG2010 said:

Hi Elvia and everybody, thank you all very much for your informative posts which helps us very much in making the right buying decisions.

The SMS200/Ultra does close to what you describe when you connect a usb drive directly to it. It plays bit perfect files from the USB storage and 'bypasses' the ethernet connection. Software such as jriver are used simply for library management. I know roy described the bridge mode as superior for the ethernet connection but not sure if he tried the 'direct route'

The OP might want to try this connection before embarking on any new pc build to get the flavour of whats on offer.

There are obviously limitations as to software choice and use of dsp / convolution processing and no video, but it would give an idea of server to dac playback.

 

 

Exactly.  Good point @LTG2010.  The OP already said he is keeping sMS-200ultra.

 

I don't recall Roy trying the subnetted mode also to totally isolate the server.

[TopQuark]

7 hours ago, ElviaCaprice said:

This has been tested by others and found not to equal the SQ results of streaming via Ethernet to the renderer.  Why?  Could be many reasons, software design and it's non flexibility with custom linux, inability to load to memory after decompressing the file or in case of WAV, ability to even load to memory or lack of memory? 

Point is, this headless server design you aiming at has it's deficiencies.  For a little bigger mobo but yet low power, you can defeat those deficiencies in a sCLK-EX server.

 

If one want to use Roon with convolution, this set-up is not going to work.  This was discussed already a million times in this thread.

 

Roy is searching for the ultimate of the ultimate.  When you say "not equal SQ", you have to listen it yourself or quantify it and not post it because Roy said this and because Roy said that.  Roy is in a different league in all of this.  His set-up is NOT the same as yours.  You have to listen to it and not just spread the gospel of the greatness of single server if you haven't listened to trifecta which you yourself admitted you have not heard!

 

Dang internet! What have thou done!

 

[Johnseye]

7 hours ago, TopQuark said:

 

If one want to use Roon with convolution, this set-up is not going to work.  This was discussed already a million times in this thread.

 

Roy is searching for the ultimate of the ultimate.  When you say "not equal SQ", you have to listen it yourself or quantify it and not post it because Roy said this and because Roy said that.  Roy is in a different league in all of this.  His set-up is NOT the same as yours.  You have to listen to it and not just spread the gospel of the greatness of single server if you haven't listened to trifecta which you yourself admitted you have not heard!

 

Dang internet! What have thou done!

 

 

LOL.  Wow, *facepalm*

[ismewor]

10 hours ago, bibo01 said:

http://www.teradak.com/products/105.html

Is this what you used? Or a DC-DC adapter?

Yes use this 400W 3yrs ago. Heat is so bad and died in a 1.5yr.

may be bad luck, bought an 600w http://www.teradak.com/products/106.html same result SQ change when it get so hot so easy so fast. 

[Lebouwsky]

On 19-10-2017 at 9:53 PM, Advieira said:

Chassis: Streacom F1C Evo Mini ITX

Motherboard: Jetway NF591-3160

Memory: Crucial 16gb 1333

Boot storage: Corsair Force MP500 120gb M.2

Cache storage: SSD Crucial 240gb Sata 2.5

 

Audiophile Upgrade
Jcat Femto Net Card
Sotm Sata filter
Roon OS (ROCK)
 

 

 

Hello @Advieira

 

Would like to respond to your question based on purely my  own experiences in a server/sms200 combo

 

* Windows 10 gives great flexibility and sounds good by itself in a server purpose

* audiophile optimizer and Fidelizer improve the sound of w10 even more

* hqplayer is a great piece of software, filters and upsampling in pcm to 24/192 sound way better then native mpd/dlna running for example minimserver

* hqplayer upsampling pcm runs on a lightweight pc like your celeron 3150 (you can use an even more lightweight cpu like my celeron j1900, 20% load max, less (distortion) is more in this case)

* 4gb is more then enough, again less is more, 50% load max

* power supply of 30 watt is more then enough, so a Sotm sps500 (50watt) will do

* cheap small embedded mb with just 1 lan port can be easely expanded with a usb to lan adapter

* sdd cards spread high frequency sound. I have not solved this, but Sotm filter or running w10 from an sd card might solve this

* my case is of similar size like yours. I can recommand a bigger one for more upgrade room. 

* the passive cooling of the embedded heatsink block of my 4 watt j1900 cooles more then enough.

* pioneering in the audio front, searching for the best of the best comes at a high price. If budget is limited like most of us, put your money on second best but proved and very very good sounding technology.

 

Looking forward to your findings and enjoy  your path to (affordable) audio heaven.

 

[phusis]

15 hours ago, TopQuark said:

 

If one want to use Roon with convolution, this set-up is not going to work.  This was discussed already a million times in this thread.

 

Roy is searching for the ultimate of the ultimate.  When you say "not equal SQ", you have to listen it yourself or quantify it and not post it because Roy said this and because Roy said that.  Roy is in a different league in all of this.  His set-up is NOT the same as yours.  You have to listen to it and not just spread the gospel of the greatness of single server if you haven't listened to trifecta which you yourself admitted you have not heard!

 

Dang internet! What have thou done!

 

 

Well put, TopQuark. 

 

I’ve pointed to the problem of this before - i.e.: copy-pasting the findings of poster @romaz in particular as universal gospel - and it’s a shame really if this should in any way reflect negatively on Roy’s contributions, which have been a help and an inspiration to many. To reiterate: help and inspiration (as in what would allow people to come to better sonic results also via their own interaction and experience), not merely blind following. I’m sure great, or even sublime results can be had with the single server approach (not least perhaps when the monetary ceiling is, well.. lofty?), obviously, but when considering the context of different aspects, including price, one option over the other (vs. Ethernet-based playback) may not see easy preference going by sound quality. Indeed, battling over and being a strong proponent of which route to take bears the scent of brand- or approach-territoriality, which is not that far off from worshipping individual efforts.. In any case, isn’t it wonderful there are different paths to come by in search of great SQ? 

[Johnseye]

 

42 minutes ago, phusis said:

 

Well put, TopQuark. 

 

I’ve pointed to the problem of this before - i.e.: copy-pasting the findings of poster @romaz in particular as universal gospel - and it’s a shame really if this should in any way reflect negatively on Roy’s contributions, which have been a help and an inspiration to many. To reiterate: help and inspiration (as in what would allow people to come to better sonic results also via their own interaction and experience), not merely blind following. I’m sure great, or even sublime results can be had with the single server approach (not least perhaps when the monetary ceiling is, well.. lofty?), obviously, but when considering the context of different aspects, including price, one option over the other (vs. Ethernet-based playback) may not see easy preference going by sound quality. Indeed, battling over and being a strong proponent of which route to take bears the scent of brand- or approach-territoriality, which is not that far off from worshipping individual efforts.. In any case, isn’t it wonderful there are different paths to come by in search of great SQ? 

 

While I think a lot of people respected and anticipated the results of Roy's tests, to the point where many copied what he was doing, I think worship is a bit extreme.  Actually it's weird to even insinuate it.

 

Many people have tried a large variety of options to improve sound quality.  What I've found is that most people here are just looking for the next best thing to try.  As soon as someone says they found some hardware, software or configuration that makes a positive impact then they want to try it.  I would place a large bet that much of the equipment you use was purchased based on what you read in a review by someone here, another forum, a local store or in a magazine review.  To blindly purchase something in this hobby is foolish.

 

When someone suggests not to take a leap forward, but to stick with a previously tried and known model as what I've been hearing in this thread from TopQuark, then I ask why.  Why would you limit yourself and not try different solutions.  The single server approach as @ElviaCaprice and I have been discussing isn't expensive at all.  You don't need a Chord DAVE for this approach.  I personally don't understand any resistance to trying a server with sCLK-EX modded clocks.  Adhering to an endpoint approach without knowing anything about a direct approach is folly.  In the end, both approaches could result in the same sound quality because they are both using the sCLK-EX.  If that's the case why the resistance?

[ismewor]

1 hour ago, Johnseye said:

  I personally don't understand any resistance to trying a server with sCLK-EX modded clocks.  Adhering to an endpoint approach without knowing anything about a direct approach is folly.  In the end, both approaches could result in the same sound quality because they are both using the sCLK-EX.  If that's the case why the resistance?

Well, I been there and done that as well. I would say using the end-point is much easier for many of us. It is simple plug and play. And you can use any computer to feed in. 

The Modded clock solution require to pick out a mobo, send it in to sotm or their distribution, than mess with finding a all other components, installing window, AO, Ssetup, Testing and who know where it went wrong in the process.

So the timeline, skill and money involve is alot higher to enjoy in music. and endpoint sounding is acceptable too.

This is my founding. Single Server done right sound better than just any machine to an endpoint.

But Double the Good server will sound better than a single server.

[ElviaCaprice]

28 minutes ago, ismewor said:

Well, I been there and done that as well. I would say using the end-point is much easier for many of us. It is simple plug and play. And you can use any computer to feed in. 

The Modded clock solution require to pick out a mobo, send it in to sotm or their distribution, than mess with finding a all other components, installing window, AO, Ssetup, Testing and who know where it went wrong in the process.

So the timeline alot longer to enjoy in music. 

This is my founding. Single Server done right sound better than just any machine to an endpoint.

But Double the Good server will sound better than a single server.

You've been there and done that, but really you haven't, because you failed to properly send out a clean primary stream from a server.  If you would have succeeded then all your additional fixer suggestions would have ceased. 

 

If you can never repair the damage from poor clocking and power in the primary server regardless of the multitudes of fixers within the streaming path, let alone the software limitations and format capabilities one has to endure, why wouldn't you want to solve the damage from the primary server to begin with?

 

If one fixes the primary server stream from being damaged, why would you need a second server via renderer or NAA to possible damage again the stream that is perfect initially to start out with from the primary server?  Answer, you don't. KISS

 

The server, when done properly is the ultimate endpoint.

 

I set out on that path long before Roy's findings or better comparisons with the trifecta.  In fact, I published my findings before Roy did.   But unable to make any comparison to the trifecta, I was not able to conclude the superiority of the sCLK-EX server, until Roy did so.  I could only conjecture at that point from what I was hearing.

 

SOtM has taken the guess work out of the mobo clock replacement and put the responsibility on themselves.  Yes, you need to send them the motherboard for doing the work, but it is worth the expense (which basically is cheaper than adding fixers) at the end of the day in comparison to the multitudes of fixers one needs to even get close to the capabilities of the sCLK-EX server.  Also to get close to the sCLK-EX server your going to have to use the "trifecta" of sCLK-EX fixers and that too may require sending in the components to SOtM for modification.

 

If one finds building a PC as difficult, installing software and or adding components.  Then yeah, I would suggest overpaying for fixers and/or a boutique server to forgo the process and remain inflexible due to one's lack of basic computer skills.  If your in to computer audio, you should be able to perform some basic skills, it's not that hard.

[motberg]

11 hours ago, ismewor said:

Yes use this 400W 3yrs ago. Heat is so bad and died in a 1.5yr.

may be bad luck, bought an 600w http://www.teradak.com/products/106.html same result SQ change when it get so hot so easy so fast. 

I have been using the 210W ATX Teradak unit nearly 24/7 for about 3 years with no problems, all SSD's and the PPA card are on separate LPS's and the server has no MB fans and no video card... I do have a small LPS driven computer cooling fan in place near the LPS fins for summer days when ambient is over 30C or so.. I am also using Teradak's to power 2 Uptone LPS-1's, a bunch of cooling fans, and with some tube preamp/buffers in an environment often well over 30C ambient without a single failure... I also found that a better Teradak (The DC30 range and up) when followed inline with an iFi DC iPurifier has a sound quality generally similar to the Uptone LPS-1.  

[ismewor]

5 minutes ago, motberg said:

I have been using the 210W ATX Teradak unit nearly 24/7 for about 3 years with no problems, all SSD's and the PPA card are on separate LPS's and the server has no MB fans and no video card... I do have a small LPS driven computer cooling fan in place near the LPS fins for summer days when ambient is over 30C or so.. I am also using Teradak's to power 2 Uptone LPS-1's, a bunch of cooling fans, and with some tube preamp/buffers in an environment often well over 30C ambient without a single failure... I also found that a better Teradak (The DC30 range and up) when followed inline with an iFi DC iPurifier has a sound quality generally similar to the Uptone LPS-1.  

The low voltage might work better, but I’m not aware they make a 210w atx LPS. I been working on 400w and up. Might have to try this one. But not sure same result will come. http://www.teradak.com/products/55.html

[tedwoods]

15 hours ago, Lebouwsky said:

 

Hello @Advieira

 

Would like to respond to your question based on purely my  own experiences in a server/sms200 combo

 

* Windows 10 gives great flexibility and sounds good by itself in a server purpose

* audiophile optimizer and Fidelizer improve the sound of w10 even more

* hqplayer is a great piece of software, filters and upsampling in pcm to 24/192 sound way better then native mpd/dlna running for example minimserver

* hqplayer upsampling pcm runs on a lightweight pc like your celeron 3150 (you can use an even more lightweight cpu like my celeron j1900, 20% load max, less (distortion) is more in this case)

* 4gb is more then enough, again less is more, 50% load max

* power supply of 30 watt is more then enough, so a Sotm sps500 (50watt) will do

* cheap small embedded mb with just 1 lan port can be easely expanded with a usb to lan adapter

* sdd cards spread high frequency sound. I have not solved this, but Sotm filter or running w10 from an sd card might solve this

* my case is of similar size like yours. I can recommand a bigger one for more upgrade room. 

* the passive cooling of the embedded heatsink block of my 4 watt j1900 cooles more then enough.

* pioneering in the audio front, searching for the best of the best comes at a high price. If budget is limited like most of us, put your money on second best but proved and very very good sounding technology.

 

Looking forward to your findings and enjoy  your path to (affordable) audio heaven.

You can also try "process lasso", if you haven't already, and set priorities for hq player to real time. It worked for me.

 

[rocl444]

hi all - i've had my ultra for a couple of months. got it with the ethernet mod. had it connected up to the home network and getting files from my windows 10 pc. quite happy with the sound - previously i had a home built atom based, sotm usb card, singled sided ram etc. the ultra was a real step up.

 

i tried a direct usb connection to an ssd in a powered dock as it seemed it would be easier but sq dropped. went to ali-express and got a little fanless celeron win 10 on msata mini comp for $125. loaded the ssd  and pressed go.

 

straight up a good step up again. lps's powering everything. spent the other evening doing comparisons with my $10000+ vinyl setup and was very impressed. different but equally enjoyable sound wise. an absence of clicks, pops etc puts it ahead atm. very happy indeed.

 

got another msata on the way and will give roon rock a go and see how that sounds. maybe a linux server distro. no urgency as i'm very content.

 

 

[Lebouwsky]

2 hours ago, rocl444 said:

I tried a direct usb connection to an ssd in a powered dock as it seemed it would be easier but sq dropped. went to ali-express and got a little fanless celeron win 10 on msata mini comp for $125. loaded the ssd  and pressed go.

Not sure if I understude you correctly. You wanted to put you music library in the server by connecting a ssd card in powered dock, connected by usb and your sound quality dropped. Compared to what, to the music files stored on an internal drive in the server? Why would you want it outside the server?

 

And you bought a 3rd computer (server =1, ultra =2, cheap celeron =3) to store your music on? To me that sounds like that cheap pc acting like a NAS for music library, feeding the server, feeding the ultra.  Is that correct?

 

My personal setup actually is quite the same. My w10 server feeds the sms200, but library is on a nas but can also be on ssd of the server. I compared library storage (nas vs ssd server), but could not hear a differance. I’ll compare again this weekend.

 

None the less, the ssd in my server is noisy and want to replace it by w10 running on sd card. Not sure where to put my library yet. Actually want to get rid of both the nas and the noisy ssd in server. A large 512 gb sd card might solve that.

[TopQuark]

On 10/25/2017 at 3:18 PM, Johnseye said:

When someone suggests not to take a leap forward, but to stick with a previously tried and known model as what I've been hearing in this thread from TopQuark, then I ask why.  Why would you limit yourself and not try different solutions.  The single server approach as @ElviaCaprice and I have been discussing isn't expensive at all.  You don't need a Chord DAVE for this approach.  I personally don't understand any resistance to trying a server with sCLK-EX modded clocks.  Adhering to an endpoint approach without knowing anything about a direct approach is folly.  In the end, both approaches could result in the same sound quality because they are both using the sCLK-EX.  If that's the case why the resistance?

 

Who said I stick with tried and known model?  I am not advocating a particular set-up. I am just answering the OP who already has sMS-200ultra.  Your suggestion to replace his sMS-200ultra with a single box server IS more expensive because he has to sell his sMS-200ultra and his mac mini at a loss for a cheap generic Jetway mobo with Realtek ethernet that you are advocating.

 

On 10/25/2017 at 3:18 PM, Johnseye said:

I personally don't understand any resistance to trying a server with sCLK-EX modded clocks.

 

I gave you my answer. He has it already in sMS-200ultra. Don't play with your "endpoint" semantics. The sMS-200ultra contains an audiophile mobo + cpu + ram + usb + sCLK-EX.  This is better than your generic Jetway mobo + cpu + ram + usb + sCLK-EX because the sMS-200ultra was designed for audiophile audio from the ground up with better filters and capacitors and not the desktop Jetway mobo you are advocating that uses generic components. The difference is really only the software.

 

Again, don't take this out of context on what the OP raised from the original thread.  The title says "Building a PC to improve sMS-200 Ultra". He has sMS-200ultra and wants Roon.  It's bad advise to let him sell what he already has - mac mini and sMS-200ultra - and propose a solution that was tried by you and Caprice because it is the best in what you have tried.  Worst is, as far as I can tell, none of you have listened to the trifecta set-up yet.

 

[rocl444]

20 hours ago, Lebouwsky said:

Not sure if I understude you correctly. You wanted to put you music library in the server by connecting a ssd card in powered dock, connected by usb and your sound quality dropped. Compared to what, to the music files stored on an internal drive in the server? Why would you want it outside the server?

the ssd dock was plugged straight into the sms200ultra. the sound was as good as when the ssd was in my pc connected via ethernet. but then i do have the ethernet mod.

 

only 2 comps. the pc is no longer in the loop. the files are on the ssd inside the chinese celeron running windows, and that feeds the ultra via ethernet. great sound.

 

will try a linux distro on the celeron comp instead of windows. i am running the free Fideliser on the windows