Jump to content

Building the ideal(ish) Music Server


Recommended Posts

1 hour ago, kukynas said:

..... to upgrade/update my 4 years old work PC anyway but I wanted to fulfill gab between now and Chanh's atx board and test it's capabilities, unfortunately with unexpected result :unsure:  

Daniel - I think we have discussed these via pm quiet extensively at length.... It appears lowest dc ripples might not always reflect better SQ, as dc-ripples are not the primary attribute in SQ contribution. I could be wrong here?! I think the ability/rates of current release/supplies are equally weight in the overall SQ differences and among other things...!? 

 

As for ATX board, I have shipped the spare boards and also the fully assembled one to someone in CA USA. Hopefully, we will get a third party feedbacks from him soon? 

 

IMG_5140.JPG

Edited by Chanh
  • Like 2
Link to comment
Share on other sites



fully agree, ripple isn't the only attribute for perfect SQ but what confuses me in terms of measurements and final performance of Corsair is that it should in every aspect perform better than those two other SMPS I tried and yet doesn't deliver expectations but again until I'm 100% sure it's only caused by SMPS I can't confirm anything

 

Link to comment
Share on other sites

so little conclusion, I exchanged streamers as it was easier than mess around with cables from DAC and can report that it's night and day, beebox just sound much better, pinpoint sound stage, relaxed and accurate presentation and clean background, these are most significant differences audible immediately, it's funny how immediate is the difference if you go back from bad to better, this also means it's not just smps as beebox sounded better with it's brick smps against i5 streamer, it seems to me that more processing power and board with its faster components doesn't guarantee better result, lesson learned...

I was waiting for months to get a chance to buy IMB-155 board but it seems it's still unavailable to general public even if asrock slowly working on their new boards with Apollo lake (N4200) chipset, fortunately (I wasn't aware of it) Gigabyte in the meantime released their industrial version of N3160TN chipset quietly for very good money so finally I have a chance for decent comparison, hopefully within next week ;)          

Link to comment
Share on other sites

Thanks for the findings - Daniel! 

It appears your findings coincided with ours and the argument we have constantly facing between atom mobo vs high power i7 or i5 CPU....., with regards to SQ's influential! 

 

My further findings have have also found, number of cores in a CPU did not contribute to SQ but rather compromised and worsen the SQ. However, I found fixed speed at higher frequency improves timing, resolution/focus, and imaging to say the least. 

 

Again - not everyone can appreciate these findings from their desktop analysis. 

Edited by Chanh
Link to comment
Share on other sites

Guest rmpfyf
40 minutes ago, Chanh said:

Thanks for the findings - Daniel! 

It appears your findings coincided with ours and the argument we have constantly facing between atom mobo vs high power i7 or i5 CPU....., with regards to SQ's influential! 

 

My further findings have have also found, number of cores in a CPU did not contribute to SQ but rather compromised and worsen the SQ. However, I found fixed speed at higher frequency improves timing, resolution/focus, and imaging to say the least. 

 

Again - not everyone can appreciate these findings from their desktop analysis. 

 

I think we conflate too many issues in Atom vs i5/i7/etc. It's less about the architecture and more about what's done with it.

 

High power and more schedulers (HT, multicore) are bad in general but that doesn't mean you have to run an i7 at high power, or that the pipeline differences to an Atom don't make a useful difference, or that having more cores and using them well doesn't increase SQ if used well.

 

Certainly power suppliers are more manageable at low power levels but it's not the only thing going.

 

I should get the current clamp out and see where I'm at with mine, BTW. Prolly too high :D 

Link to comment
Share on other sites



On 28/09/2016 at 2:13 PM, realysm42 said:

Guys I want to ID the clock I should be replacing on my new motherboard.

 

For the record it's a GIGABYTE GA-H110M-S2H - aka an LGA1151 board with an H110 chipset. According to Paul Pang, that means I should need a 24MHZ clock. It doesn't hurt to double check though - I've not got it yet and this is the highest resolution image (attached) I can find on the web (I've highlighted the three potential clocks in red).

 

Untitled.jpg

 

Questions:

  • Any ideas on which one I should be removing (if I were a betting man I'd say it's the one in the middle) and
  • what it should be replaced with (in terms of MHZ clock)

 

@JDWest Finally racked up the guts to buy a soldering iron and tackle this upgrade. Been sat in a box for (way too many) months...

 

My girlfriend actually did the soldering! And... my god it actually works!

 

Right now it actually sounds a little rough around the edges but that doesn't surprise or bother me; can't wait for it to burn in. Will report back.

 

Next for me will be proper sata cables and all silver wiring on the inside to replace the generic stuff. Then I'll stop.

 

Maybe.

Link to comment
Share on other sites

Guest rmpfyf
7 minutes ago, realysm42 said:

@rmpfyf thanks for the advice on which clock to replace mate; turns out you were right!

 

Glad the forum could help - I've learned heaps here, am still learning and am glad to contribute in some way :) 

Link to comment
Share on other sites

Update: I left the system on all night (to help the new clock burn in).

 

Yesterday I commented that the card sounded slightly rough around the edges. The track I briefly listened to-to make that initial assessment (Jason Isbell, California Pines (GF's choice)) still sounded like that today.

 

On reflection I think there's a number of factors at play:

 

  1. These cards take time to burn in and really open up
  2. That track isn't a great recording anyway and actually, the greater transparency allowed by the clock upgrade is widening the gulf between great / good / bad recordings.
     

To test point 2, I've listened to a couple of known fantastic recordings and I'm getting more of the SOtM goodness as what I can only assume is further reduced timing errors / jitter; less digital sound, more analogue. Although I'm no stranger to lowering the noise floor (by any means necessary!) and what it does for sound, it remains a great pleasure and surprise to discover more of the music; those little nuances and small touches. The steps that helps us shift the hi fi into real performances. I can't wait to get a couple of hundred hours on the clock!

 

One 'interesting' side effect is that some video games no longer work and cause the PC to hard lock (everything else works perfectly btw). This PC has fist and foremost always been an audio machine, (that happens to be able to play games (on low settings)) so I'm not heartbroken, but it is annoying nonetheless. I guess this is always a risk when modifying things with 3rd party components. Having said all this - I wouldn't take it back :)

 

That's three discrete SOtM cards under one bonnet, they've really played their part to unlock the music for me.

Edited by realysm42
Link to comment
Share on other sites

After more than 4 years of on going spending, tweaking, investing, and spent undefined funding/time, I came to realise, who I am kidding with Computer Audio!!! As much I wanted it to become close to TT or matched a decent CD-Transport, it did not happen. The truth is it is a hyperbola graph......, you can spend all the time, funding, and level of effort(s), you want, you can only get closer and is never apple to apple in term of SQ.

Yes - the excitement at time when a teak or a modification made the SQ sounds differently, not necessarily be better..., though we have the tendency of judging the SQ differences as better SQ....!  

I could be wrong here but the only thing I can really appreciate from this journey is CA is only for the conveniences....! :D

 

Having said that, I am now investing my resources into a decent Turntable Setup...! Any recommendation for the darkside would gratefully appreciated!

 

 

Edited by Chanh
  • Like 1
Link to comment
Share on other sites



After more than 4 years of on going spending, tweaking, investing, and spent undefined funding/time, I came to realise, who I am kidding with Computer Audio!!! As much I wanted it to become close to TT or matched a decent CD-Transport, it did not happen. The truth is it is a hyperbola graph......, you can spend all the time, funding, and level of effort(s), you want, you can only get closer and is never apple to apple in term of SQ.
Yes - the excitement at time when a teak or a modification made the SQ sounds differently, not necessarily be better..., though we have the tendency of judging the SQ differences as better SQ....!  
I could be wrong here but the only thing I can really appreciate from this journey is CA is only for the conveniences....! [emoji3]
 
Having said that, I am now investing my resources into a decent Turntable Setup...! Any recommendation for the darkside would gratefully appreciated!
 
 

EMT Chanh.
Link to comment
Share on other sites

Ouch, what you have been spending on amps and speakers, I am shuddering how much you will need to pay for a TT to make you happy :sorry:.

All jokes aside, I have Consonance amps and they are good value for money. 

Link to comment
Share on other sites



@Chanh whose that cheeky bastrd converting you to pure analog? :D 

 

I was thinking the other day about similar how to maximize my current gear to sound as analog/natural as possible without loosing too much of accuracy and details but not that drastic approach like going straight to TT :) , I found out after many attempts that the biggest drawback in my digital (PC) to analog conversion is USB > I2S transfer, when I bought gustard to try out different DAC chip and made comparison to my existing DAC the biggest difference was digital glare or as I call it cold and unnatural presentation, after further analysis and back and forth comparisons I found it's all caused by USB converter in this case Xmos chip, one might think there shouldn't be anything wrong with that chip if clocking and psu is made as perfect as possible, nope or at least nope in my experience, I got 3 xmos converters in different price bracket with different clocking and psu and none of them can cope with FPGA found in Amanero, I don't know why such difference and why is Amanero sounding so good (natural) but for sure it's not "only" caused by higher/lower jitter otherwise xmos found in gustard with it's flip-flop reclock and accusilicon xtals should sounds the best, I believe owners of Ian's FPGA reclock board came to the same conclusion...

 

I'm for sure done with my front end so no more boards, PCs, PSUs and I don't know what else, I might get slightly different result with this or that (as Chanh mentioned) but is it worth it? not really, too much time and effort spent for trying and failing and I'm still going back to the old solution :)     

Link to comment
Share on other sites

I can't believe this return to Vinyl given the limitations of the medium https://www.st-andrews.ac.uk/~www_pa/Scots_Guide/iandm/part12/page2.html \

The so called glare with the Gustard disappears over time - 200 plus hours and can be tempered if paired with a valve preamp. You will see numerous claims on the Gustard head-fi site re improvements over vinyl, especially when processing DSD upsampling, and the advantages of singxer Su-1 with XMOS XU208 output to i2s.

Link to comment
Share on other sites

1 minute ago, Triplefun said:

I can't believe this return to Vinyl given the limitations of the medium https://www.st-andrews.ac.uk/~www_pa/Scots_Guide/iandm/part12/page2.html \

The so called glare with the Gustard disappears over time - 200 plus hours and can be tempered if paired with a valve preamp. You will see numerous claims on the Gustard head-fi site re improvements over vinyl, especially when processing DSD upsampling, and the advantages of singxer Su-1 with XMOS XU208 output to i2s.

 

nope it won't, I have way more than 200 hours on mine but even if I don't I had the same trouble/experience with my current Sabre DAC when fed by xmos and btw. I'm talking about A20H not X20 ;)

why would be anyone buying gustard DAC with its pretty good THD+N and SNR figures if all would be consequently destroyed with valve preamp? to tamed it's flaws? I don't wanna judge here but that's not my approach ;) , if I wanna do something similar it would be easier to buy DSD capable NOS DAC and still getting better result

Link to comment
Share on other sites

2 hours ago, kukynas said:

@Chanh whose that cheeky bastrd converting you to pure analog? :D 

I'm for sure done with my front end so no more boards, PCs, PSUs and I don't know what else, I might get slightly different result with this or that (as Chanh mentioned) but is it worth it? not really, too much time and effort spent for trying and failing and I'm still going back to the old solution :)     

I am not entirely out of the wood just yet! Still can appreciate DSD into my Accuphase DC-950 pretty darn much. Most of all, still assisting fellows SNAer(s) with their server buildings. I have already built one for a Gent in NSW, another from NSW with remain parts on their way, one from Vic, and a local SNAer. :love  

 

I think, together We've managed to bring down a commercialised Server from our Kiwi-friend (Antipodes DX) based on our collaborations. I have also received feedback(s) from members via pm, on how they did a shootout vs Aurender W.20, and found it performed better against the $20k commercialised unit. Another findings that is more publicly posted can also be found here, http://jplay.eu/forum/index.php?/topic/1022-aurender-w20-versus-my-pcjplay/?hl=aurender

 

  • Like 2
Link to comment
Share on other sites



2 hours ago, rmpfyf said:

So Chanh... no power supply boards?

No @rmpfyf. I have been seeking advices from the appropriate departments/authority in W.A, and it was recommended that not to do PCB group buy. We are dealing with a variation of  225VAC to 260VA in Australia, and it can be lethal/deadly.

 

For those interested in DIY-Modular unit(s), I think @kukynas, is your man to talk to. He has recently discovered a shunt regulator that can handle up to 8A current, from a guy in Korea. May be more infor from him?  

Edited by Chanh
typo
Link to comment
Share on other sites

5 minutes ago, kukynas said:

@rmpfyf @Chanh are you referring to whole PSU unit or ATX board?

 

if anybody interested I can share the source for custom made isolated R-core trany and PSU boards... 

If I understood it correctly the whole psu. 

ATX board is still pushing ahead Guys!

Link to comment
Share on other sites

Still wondering if someone else is going to run a miniATX 1151 on pure DC ") from a solar maintained battery/capacitor bank

Anyone ?

 

IMG_4978.JPG

ignored again :) and dismissed :D and will stay my secret weapon. 

 

Edited by Guest
Link to comment
Share on other sites



  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...
To Top