Jump to content
IGNORED

Building a DIY Music Server


Recommended Posts

On 3/30/2022 at 8:54 AM, sakso136 said:

I started with jcat usb/ net ,then added the taiko dc-atx ,then the ulps to taiko.

Then usb xe .

The Xe is now powered by keces p8 thru dxp 3A.

Biggest impact in my system was the xe then taiko then ulps.

I tried ferrum hypsos lps into the xe,i was floored by results...

Yes, The JCAT USB XE powered by an MPAUDIO ALS-HPULN was in an entirely different league from the generic Startech USB card I had been using, to my ears. It also bested another really solid USB card with the same LPS. Switching to a DC4 for the JCAT USB XE card was yet another sizable gain in performance. 
 

The Taiko DC-ATX vs the HDPlex 800w was a shocking and unexpected improvement as well.

Link to comment
On 4/1/2022 at 3:58 AM, sakso136 said:

Thks for the update.

The ferrum i tried on usb xe is quite expensive and i didnt pull the trigger yet.

While the dc 4 im sure is the best of the best,would you be more precise describing the gains versus the mpaudio.

I know the best ulps should go to usb xe,but i need at least 3 rails ,and maybe the mpaudio can fill the blanks . 

Thks

Hi Sakso,

it has been quite a while since then. I do remember feeling that there was an across the board improvement. Image size was more focused, but larger. Instruments sounded more real and presented themselves in a more tangible way. Dynamics improved in a big way.

 

That being said, ALS-HULPN is no slouch and I use it, happily, for the Femto NVME. It does a swell job in this use. I don’t think you’ll be disappointed unless you heard them back to back.

Link to comment
25 minutes ago, sakso136 said:

one day i will upgrade the bg7 to it or maybe the ref120.

You and I both. Very happy with the Bg7TBL powered by LPS 1.2. I’ve tried removing them for a sanity check… they both end up right back in the chain. There is some value add there for me.

 

Ive had my eye on the Afterdark Emperor Double crown as a stopgap for the Mutec. Soon, I will have an SR7T rail available to power the BG7TBL so will observe any improvements going that route. I wonder how close the AfterDark + SR7T will get to the Mutec (subjectively)… keep everyone posted.

Link to comment
16 minutes ago, Tatomek7 said:

 

I had a chance to test in my system the LPSs from Jcat: Optimo Duo with 2 x 5V and Optimo S ATX.

 

First comparison was between Optimo S ATX and ULPS + Taiko ATX to power my server based on Asus Sage with two Xeons.

Jcat had no problem to power it and temperature of the LPS was quite moderate. I powered only ATX and CPUs. At this stage of the test LAN and USB card was powered by SJ ARC6 DC4 & DC3. In my case it’s Jcat Lan Femto and Jcat USB XE.

Every disconnection of power cord from S ATX requires 2 -3 days for stabilisation of the sound.

My ULPS was built initially based on Nenon’s V2 and after V3 scheme, but at the final stage I went to slightly different option and I have one 10A custom made choke and 4 x 22 000uF and 2 x 47 000uF Mundorfs HC caps. Bypass capacitors are also different, only the last one in the chain is VCap Odam, but I will order more of them in the future.

Coming back to S ATX and how it sounds. It’s really great LPS, what is really outstanding is 3D soundstage and top end registers. Bass is also very good. If anybody is looking for a LPS to power his server and consider to build ULPS to power Taiko ATX, Optimo S ATX could be nice option particularly, you are not forced to tinker yourself the ATX wires.

Differences between ULPS + Taiko ATX and S ATX? Of course there are, but thanks to both of them, you will be able to achieve high quality of the sound.

Further improvement was achieved, when S ATX was moved on the platform, what improved every aspect of the sound. I will have chance to say more, when I will disconnect it and come back to my ULPS + Taiko ATX.

 

Second set of comparisons I did; Optimo Duo vs SJ ARC6 DC4 & DC3.

 

In the past I compared Keces P8 with Jcat Optimo Duo and latter was much better. I was really delighted by the increase it offered over Keces. Now I had a chance to compare it with SJ ARC6 DC4 & DC3. Comparison was done on the Jcat USB XE card ( DC4 ) and Jcat LAN Femto  ( DC3).

 

I compared DC4 and Optimo connected with the same DC cable, which was JSSG 360 Mundorf SGW solid wire terminated on one side with Neutric SpeakOn and Oyaide DC2.1 on the other. DC4 is better than Optimo Duo. ARC6 DC4 internal wiring is also based on Mundorf SGW,  what makes initial expensive project even more expensive, but it returns with better sound.

I did also comparison of DC3 and Optimo Duo. The results was;  as DC4 is better, Optimo is better by significant margin over DC3.

 

So my ranking is Keces P8 > ARC6 DC3 >> Optimo Duo > ARC6 DC4

 

All tests were done with DC cable based on Mundorf SGW. DC cable really makes the difference and if you have possibility to make it, please do, it will pay you back with important uptick in the sound.

 

When I did the test of DC4 vs Optimo powering USB card, Lan card was powered by DC3. During the test of Femto card powered by DC3 vs Optimo, I discovered, DC3 was limiting the performance of Femto card, so I did one more comparison of USB card powered by DC4 or Optimo, when Lan card was powered by Optimo. The distance between was closer, but DC4 is still better, nevertheless I can understand some could prefer Optimo, particularly, if you have softly sounding amp. Optimo is just playing more contour sound, when DC4 adds some softness,. 

Very good post Tatomek! A lot of us were wondering how the Optimo will sound compared to the ULPS. In your case, powering the Dual Xeon CPU build. Hopefully I get to hear and compare the Optimo S ATX to the ULPS at some point as well for those of us running single CPU builds. Interesting feedback on the Optimo vs DC3 as well. I use a dual rail DC3 to power my pair of Buffalo switches but have not compared them to anything other than the stock SMPS. 

 

The DC4 ARC6 is also something I hope to hear some day :)

 

I now have an Optimo on my radar as I rotate out the Keces P3 (at some point).

 

Cheers,

-Rob

Link to comment
5 hours ago, TimF said:

 Building a dc cable next for the unregulated power supply to attach to the computer. Probably using  the Mundorf silver cable. Do you use one cable each for positive and negative or two? I.e. positive has two cables and negative has two as well. 


Thanks Tim

To fit the Taiko DC-ATX’s 6 pin input, I would suggest 3 for positive and 3 for negative.

Link to comment
1 hour ago, oneguy said:

Anyone here have a Z690 Aorus Master MB and is running a JCAT USB XE card? Marcin sent me a card with the firmware to run that card on a Z590/Z690 motherboard and I get the same result as with the other USB XE card I have….nothing. 
 

I have a Net XE that works and I swapped PCIE slots with the USB XE and the Net XE works in both slots so it’s not the slot. 
 

Same power cord used on both USB XE cards and Net XE so it’s not the power cord. 

What type of power are you feeding it? What’s the voltage of that power supply?

Link to comment
11 minutes ago, oneguy said:

I am feeding it the 5v output from the HDPLEX 300w LPS. It’s rated at 3a max current. I checked the voltage while troubleshooting about month ago and I believe it was outputting 5.15v. 

5.15v is good. I do not see why this wouldn’t work. Does your DAC require 5v input? If so, the jumper will need to be left on the card.

Link to comment
  • 4 weeks later...
6 hours ago, ASRMichael said:

Even Opto USB connected to mobo is very good, as long as Opto is power by good LPS

I was forced to go this route for a few weeks and I was surprised how good the Opto sounded, being fed data/power at the input via the motherboard , with a DC4 powering the last leg.

 

While it does sound great, the JCAT USB XE is still the winner on its own. I will mirror @ASRMichaeland say that the two sound incredible together. At that time, I had actually somewhat defeated the isolation by feeding both sides of the Opto power from a single DC4 rail. Probably not a good long term solution if you plan to keep the CX module enabled which has a 2a limit.

 

More experimentation on the way.

Link to comment
16 minutes ago, elan120 said:

When you running JCAT USB XE + Opto USB, what LPS you use to power the Opto USB receiver with?

I was using a single DC4 rail. To power everything.

 

21 minutes ago, elan120 said:

From the conversation I had with Dan, I plan to test Single mode fiber optic cable (up to 10M) first, where phase noise is better.

Interested in the results of this test! I plan to do the same.

Link to comment
10 minutes ago, ASRMichael said:

Thermal Grizzly - Kryonaut thermal paste - For cooling all processors, graphics cards and heat sinks in computers and consoles

This is what I use as well. It is much better than what came with the last version of the HDPlex H5. The improvement gradually improved as the paste settled.

 

It will stain the rough surfaces of your HDplex H5 though. With some elbow grease it’ll come off but yeah it’s very goopy and messy to work with. Works great though!

Link to comment
4 hours ago, Gavin1977 said:

Your logic is sound, but I don't recall it having much/any effect - the effect of isolating is more pronounced.  I will try again at some point and let you know.

I completely agree with this statement. The benefits of isolation are apparent and more pronounced than any clock degradation to my ears.
 

I’ve also validated the importance of keeping your best LPS closest to your dac. I tried powering the JCAT with DC4 and Opto receiver with ALS-HPULN, and this sounded really good. Swapping these LPS’s sounded substantially better.

 

 

Link to comment
14 minutes ago, Superdad said:

With a great PS it is not within a hair's breath of my Pareto NUC/Roon setup

Did you mean “now” instead of “not”?

 

Great post! Interesting experiment. I now have this on my to try list:  FIBBR Alpha Optical USB 

Link to comment
On 5/3/2022 at 4:40 PM, ASRMichael said:

@Exocer what was your findings/differences? 

I've found that keeping the best LPS at the last leg resulted the best sound. It was not as close as you’d think, all else equal. I do not have enough rails to disable power from the JCAT to the first leg of the Opto, to feed the Opto with another dedicated rail at the transmitter. Looking forward to your tests!

 

I suppose I could borrow a rail from the Revelation SSD temporarily to test whether there are additional benefits to be had by going this route. A part of me feels the test would really only be worth it if I had another rail at the same quality level as the DC4. Otherwise, i'd be raising the question of if sharing the DC4 with the JCAT/Opto transmitter is better than using dedicated lower quality rails for each. Not the best test condition :D.

 

I'd really like to try this "https://fibbrcable.com/products/alpha-usb-a-b-optical-fiber-digital-audio-cable?variant=41474949349576" but:

 

1. The micro USB is kind of annoying but I suppose I could just use this to make my current cable compatible :https://www.amazon.com/AAOTOKK-DC5-5x2-1mm-Connector-Cellphone-2Pack-Micro/dp/B07XZ8BD6J/ref=asc_df_B07XZ8BD6J/?tag=&linkCode=df0&hvadid=385179140364&hvpos=&hvnetw=g&hvrand=16749794830653768090&hvpone=&hvptwo=&hvqmt=&hvdev=c&hvdvcmdl=&hvlocint=&hvlocphy=9067609&hvtargid=pla-829034086986&ref=&adgrpid=82240853201&th=1 (ashamed to admit I use something similar to connect the DC4 to the Opto transmitter and it still sounds glorious.

 

2. My longer Mudorf SG DC lead, ~ 12", would never reach the Micro USB input. I would need to relocate my DAC in order for power to reach the FIBBR device.

 

Curious if others are willing to try it and report back!

 

Cheers all,

-Rob

Link to comment
  • 3 weeks later...

And in the spirit of continuing this thread: 

 

For those with the Noratel transformer, please try the ULPSv3 without the Softstart!

 

Credit to @Nenonfor the suggestion. 😎

 

Disclaimer: He did not suggest trying this with the Toroidy supreme V2. Haven’t asked why. Maybe he’ll answer here.

Link to comment
13 minutes ago, Nenon said:

If that's the typical behavior after shipping, then we can never hear the Extreme's full potential at a 3-day high-end show :(.  Unfortunately that's the case with a lot of other equipment. 

We have cable cookers, why don’t we have “server cookers”? 😂. Euphony has a handy stress test mode. Would such a mode work towards speeding up performance stabilization? Prime95 24/7 to speed up the Taiko burn-in process perhaps. (There are probably much newer solutions by now…).

 

I don’t want to turn my server off for a week to find out… it’s sounding too good these days.

Link to comment
2 hours ago, Dev said:

For cooking servers components, I use HQP upsampling with the most demanding filters playing 24x7 for couple of days. Admittedly its a misuse of HQP and Miska wouldn't like this but it works like a charm.

This is closer to a broil 😂. All jokes aside, you must have your cooling under control if you’re able to accomplish this with stability! Kudos

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...