Building a DIY Music Server

[Nenon]

1 minute ago, Dev said:

yes but boot can consume over 200watts.

But you only had 140W (28V / 5A).

[Dev]

1 minute ago, Nenon said:

But you only had 140W (28V / 5A).

 

yes, that is why a 10A choke might help in this situation. As I said, it was a boot issue and once it gets passed, the 5A choke might just work 😁

[Nenon]

Just now, Dev said:

yes, that is why a 10A choke might help in this situation.

And/or a higher voltage transformer. But yes, I agree your case is different. 

 

There is an easy test everyone can do. Here is a cheap $20 Wattmeter from Amazon - https://www.amazon.com/gp/product/B07VPTN8FZ/. Measure what your computer draws from the wall with an efficient PC PSU.  You will get an idea how much power it needs during boot, when playing music, and when idle. 

In my case, I had close to 200W peak at boot, 95W when playing music, and 65W when idle. 

If I was close to 150W while playing music for example, I would not consider having enough headroom with 200W. But at 95W, I have no concerns. This is with two 80W THP Xeon CPUs. I would expect that most CPUs up to 95W TDP would be okay with 200W PSU. 

 

But 120W+ CPUs with active cooling and heavy upsampling will obviously require more current. 

 

[Dev]

8 minutes ago, Nenon said:

And/or a higher voltage transformer. But yes, I agree your case is different. 

 

I think one also need to strike a balance between heat dissipation and voltage. A higher voltage means higher heat because none of the components work at that high voltage. If you have a stable current source, the preference is to use a lower voltage which will run cooler. Most designs don't and hence they are compensated with higher voltage and larger heat sink. Again, I am not talking about Taiko DC-ATX but in general.

 

 

12 minutes ago, Nenon said:

There is an easy test everyone can do. Here is a cheap $20 Wattmeter from Amazon - https://www.amazon.com/gp/product/B07VPTN8FZ/.

 

I have been using P3 Kill-A-Watt for years now.

[Dev]

43 minutes ago, Nenon said:

I would expect that most CPUs up to 95W TDP would be okay with 200W PSU. 

 

The i9-10900k when un-throtlled to its highest performance and can be kept to run cooler at the same time, sounds amazing. I have had other i7/i9 variants and also a 8-core/16 threads Xeon E and none come close to the performance of i9-10900k. The hdplex struggled to keep it cool enough when un-throttled and now I have 4 external fans blowing over it (thanks to @ASRMichael). I am also planning to use the left side-panels and will have something in coming weeks. All these with a cheap trans from Antek. So with a proper designed unregulated lps with high quality parts, I am expecting some more to be had from the SQ. Time will tell 🙂

 

 

[Exocer]

Today I measured power consumption of my server during boot. I witnessed a very quick peak around 135Watts but it consistently hovered around 53watts during boot, dipping into the high 40s at times, eventually settling down to 37W.

 

I didnt catch the peak with my camera but perhaps I'll record a video and snap a picture from there.

Edit: here is peek power consumption during boot:

 

During music playback AND idle, this is what I hover around, 37.5W:

 

I9-9900k OCed to 4.6ghz. 

Edit: I do not upsample with HQPlayer.

[Mags]

13 minutes ago, Exocer said:

Today I measured power consumption of my server during boot. I witnessed a very quick peak around 135Watts but it consistently hovered around 53watts during boot, dipping into the high 40s at times, eventually settling down to 37W.

 

I didnt catch the peak with my camera but perhaps I'll record a video and snap a picture from there.

Edit: here is peek power consumption during boot:

 

During music playback AND idle, this is what I hover around, 37.5W:

 

I9-9900k OCed to 4.6ghz.

With an i9 9900K overclocked to 4.6GHz, do you still run into issues with some HQP filters and modulators? 

[Exocer]

Just now, Mags said:

With an i9 9900K overclocked to 4.6GHz, do you still run into issues with some HQP filters and modulators? 

Good question.

 

I should have mentioned I do not upsample. I mostly use Euphony Stylus to stream at default sample rates or for PGGB file playback. I occasionally use HQPlayer but mostly to check whether it sounds better than Euphony stylus (at default sample rates).

[Mags]

7 minutes ago, Exocer said:

Good question.

 

I should have mentioned I do not upsample. I mostly use Euphony Stylus to stream at default sample rates or for PGGB file playback. I occasionally use HQPlayer but mostly to check whether it sounds better than Euphony stylus (at default sample rates).

OK, thanks.

[AngeloVRA]

This is a follow-up to my initial post on my server build as I continue to work on it (seems never ending!) : https://audiophilestyle.com/forums/topic/58164-building-a-diy-music-server/?do=findComment&comment=1147932

 

The Taiko ATX has a 5VSB / 5A rail that has much lower ripple specs (0.01mVpp) than its other rails. How does that compare to an LPS when powering my JCAT USB XE card?

I made a molex cable to power the JCAT USB XE card from the Taiko ATX 5vSB rail using Mundorf SG115 wires, same material as the DC cable from the SR7T 5v rail. 

Result: I was surprised that it was a big SQ difference in all areas and particularly in terms of dynamics, detail and resolution in favor of the external LPS. The Taiko ATX 5V SB rail was disappointing in this particular use case.

 Hence, the JCAT NET XE and USB XE cards will remain powered by two 5v rails from the SR7T

The JCAT USB XE card has a power consumption of about 700 mA @ 5V. It does seem that the SR7T LPS gives better SQ than the Taiko ATX at low currents. 

Next, I wanted to compare powering the CPU from the LPS vs Taiko ATX. Since the Taiko ATX was consuming about 2.0 Amps @ 19v, I was estimating that the CPU alone would be  consuming about 1-1.5 Amps @ 12V, making it a very viable candidate to be directly powered by an LPS. 

First step would be to fabricate an 8P+4P cable using the same Mundorf SG wires as I used for the EPS cable from Taiko ATX.

I decided to use 1 15.5AWG + 5 18AWG wires per side. Crimping all the wires in a WBT copper cable sleeve and then soldering, with the bigger 15.5AWG wire extending past the sleeve and then soldered to the chassis connector.

For safety and “best practice”, I decided to change the female chassis connectors to male. This is so that the connector at the destination side of the DC cable would be female sockets. Male pin plugs are prone to short circuits when the LPS is powered on and the destination connector wasn’t plugged in.

 

Be mindful that the relative position of pin 2, which in my case is used as the positive terminal, is reversed with pin 1 when changing from a male to female chassis connector. You will have to swap the positions of your + and – wire.

 

And above is the new EPS cable as installed.

 

Some current measurements:

-       185 mA with server off but LPS on

-       A max of 2.168 Amps during boot. It is possible that the actual max is much higher than this but my meter wasn’t fast enough to capture it.

-       Settles down to around 1.12 Amps when playing music

This is for a Xeon W-1290 CPU running at its base frequency of 3.2 Ghz, 10 cores, HT off, Turbo off, C-states off.

The SQ with the SR7T supplying 12V to the CPU was hands down much better than the 12V from the Taiko ATX. Both used the same wires and both have been burned in for at least 2-3 months.

 

Some more current measurements, this time on the JCAT NET XE card with a) Euphony’s normal playback and b) Euphony’s “playback with network disconnected”. Current consumption goes down by about 265 mA as Euphony disables some functionalities of the card but doesn’t totally shut it off.

 

 

And my last experiment was to test whether there were audible differences in the length / type of PCIE riser cables used to connect the JCAT NET XE and USB XE to the motherboard in my build.

Above is HDPlex’s 20 cm x16-wide riser cable. The cable is very wide and is probably hampering good ventilation to the cards. The x16 connector is also too long for the JCAT (x1) cards and are sort of unstable.

Below pic shows shorter 5cm (to USB card) and 15cm (to NET XE card), narrower x4 pcie riser cables from ADT Link. Motherboard connector is x16 so it gives a stable connection. X4 connectors to the USB/NET XE cards are also a better fit and more secure. Ventilation around the cards should be much better.

As a bonus, the shorter, narrower (and different brand) of PCIE riser cables made for a positive and audible SQ difference. Relatively, the HDplex riser cable had more lower treble energy ans had less body.

When it finally becomes available, it might be a good step up for me to change to the HDPlex H5 case so I can plug in the cards directly to the motherboard.

Next on the horizon would probably be upgrading the standard footers.

 

Once again, I thank all AS guys who have been generously posting and sharing their experience and knowledge!

 

[Nenon]

@AngeloVRA Very nice report! Thank you for sharing. And definitely try to get rid of the riser cables as soon as you can and connect the PCIe cards directly. I went to the extent of making my own pure silver wire custom riser cable. That sounded much better than other alternatives. But no riser cable comes even close to direct PCIe in my experience. 

 

Your findings about powering the JCAT USB card with the 5V from the Taiko are inline with mine. I also prefered to power up mine with a LPS. And the Sean Jacobs ARC6 DC4 prototype I had really transformed the way that card sounded. 

 

As for the Taiko ATX vs. LPS on the CPU - there were some discussions about that earlier in this thread. First, you need a super low impedance energizer to make the Taiko ATX work better than a good LPS. Even if the SR7 is super low impedance, the cable and connectors I see on the pictures would kill that. Every foot of the 15.5 AWG MSG wire adds about 3 mOhms. You can hear differences in SQ even with changes that impact the impedance by less than 1 mOhm. Second, in my experience the Taiko ATX sounds much better with higher voltage. 19V and those cables and connectors are really crippling the capabilities of the Taiko ATX. 

 

@Exocer was probably the first person to do that comparison. And he prefered his CPU powered by his PH LPS as well. I made the same comments to him (in this thread) and he built the v2 of the unregulated LPS I posted about. I believe that changed his perspective and he is planning to revise that comparison again. It will be interesting to see what he thinks now that he has the DIY unregulated LPS I recommended running at a higher voltage powering the Taiko ATX. 

 

In any case, good work and very nice report.

[Exocer]

54 minutes ago, AngeloVRA said:

When it finally becomes available, it might be a good step up for me to change to the HDPlex H5 case so I can plug in the cards directly to the motherboard.

Next on the horizon would probably be upgrading the standard footers.

Totally agree with this idea. Getting rid of the riser cables will be an improvement.

Also, I haven't ventured too far down the footer journey but what I will say is:

 

1. Gaia footers for the PHSR7T was absolutely worth it.

2. The Herbies Tenderfeet I use for the server were also an improvement.

 

Sounds like you're headed in the right direction. Wonderful job with wiring BTW.

14 minutes ago, Nenon said:

@Exocer was probably the first person to do that comparison. And he prefered his CPU powered by his PH LPS as well. I made the same comments to him (in this thread) and he built the v2 of the unregulated LPS I posted about. I believe that changed his perspective and he is planning to revise that comparison again. It will be interesting to see what he thinks now that he has the DIY unregulated LPS I recommended running at a higher voltage powering the Taiko ATX. 

Correct - you just reminded me that I should test going back to the CPU being fed by the PHSR7T along with the V2/V3 comparisons. I did initially prefer the SR7T powering the CPU via the 12V rail. But even before I switched to the Duelund 20awg solid core copper EPS cable, it became apparent that there was a loss in separation/speed/transparency/detail retrieval going to the CPU directly from the SR7T. It did sound a bit weightier in the lower registers but somehow with enough break-in the Taiko which initially sounded a bit thin, filled out tremendously with break-in and eventually I had all the slam I needed + the positive traits of the SR7T.  In the end it still was about as weighty and more nibmle/transparent for me, which I why I never switched back to the CPU direct to SR7T. Plus, the transparency may have highlighted issues elsewhere in my system... (I had some really crappy sounding interconnects mucking up the sound more than I realized).

 

V2 from the moment I powered it on was even more transparent and effortless than the SR7T. @AngeloVRA, I would be very interested in your impressions and hope you do entertain building one at some point. You certainly have the technical abilities!

 

[Soul Analogue]

5 hours ago, AngeloVRA said:

 

This is a follow-up to my initial post on my server build as I continue to work on it (seems never ending!) : https://audiophilestyle.com/forums/topic/58164-building-a-diy-music-server/?do=findComment&comment=1147932

 

The Taiko ATX has a 5VSB / 5A rail that has much lower ripple specs (0.01mVpp) than its other rails. How does that compare to an LPS when powering my JCAT USB XE card?

 

I made a molex cable to power the JCAT USB XE card from the Taiko ATX 5vSB rail using Mundorf SG115 wires, same material as the DC cable from the SR7T 5v rail. 

Result: I was surprised that it was a big SQ difference in all areas and particularly in terms of dynamics, detail and resolution in favor of the external LPS. The Taiko ATX 5V SB rail was disappointing in this particular use case.

 Hence, the JCAT NET XE and USB XE cards will remain powered by two 5v rails from the SR7T

The JCAT USB XE card has a power consumption of about 700 mA @ 5V. It does seem that the SR7T LPS gives better SQ than the Taiko ATX at low currents. 

Next, I wanted to compare powering the CPU from the LPS vs Taiko ATX. Since the Taiko ATX was consuming about 2.0 Amps @ 19v, I was estimating that the CPU alone would be  consuming about 1-1.5 Amps @ 12V, making it a very viable candidate to be directly powered by an LPS. 

First step would be to fabricate an 8P+4P cable using the same Mundorf SG wires as I used for the EPS cable from Taiko ATX.

I decided to use 1 15.5AWG + 5 18AWG wires per side. Crimping all the wires in a WBT copper cable sleeve and then soldering, with the bigger 15.5AWG wire extending past the sleeve and then soldered to the chassis connector.

 

For safety and “best practice”, I decided to change the female chassis connectors to male. This is so that the connector at the destination side of the DC cable would be female sockets. Male pin plugs are prone to short circuits when the LPS is powered on and the destination connector wasn’t plugged in.

 

 

Be mindful that the relative position of pin 2, which in my case is used as the positive terminal, is reversed with pin 1 when changing from a male to female chassis connector. You will have to swap the positions of your + and – wire.

 

And above is the new EPS cable as installed.

 

 

Some current measurements:

-       185 mA with server off but LPS on

-       A max of 2.168 Amps during boot. It is possible that the actual max is much higher than this but my meter wasn’t fast enough to capture it.

-       Settles down to around 1.12 Amps when playing music

This is for a Xeon W-1290 CPU running at its base frequency of 3.2 Ghz, 10 cores, HT off, Turbo off, C-states off.

The SQ with the SR7T supplying 12V to the CPU was hands down much better than the 12V from the Taiko ATX. Both used the same wires and both have been burned in for at least 2-3 months.

 

Some more current measurements, this time on the JCAT NET XE card with a) Euphony’s normal playback and b) Euphony’s “playback with network disconnected”. Current consumption goes down by about 265 mA as Euphony disables some functionalities of the card but doesn’t totally shut it off.

 

 

Thank you so much for your finding and listening impressions of Taiko ATX and LPS

it concurs with my concept and rationale of using true LPS for all rails

the Taiko ATX is a very good and relatively economical solution to supply a high power music server, approaching the result of full LPS

But for me, I surely will opt for full LPS for ultimate sound quality and fidelity =)

I just managed to complete 3 L input chokes for 3.3v, 5v and 12v at 8A spec.....cant wait to have the whole project completed 😍

 

 

[Darryl R]

17 hours ago, AngeloVRA said:

When it finally becomes available, it might be a good step up for me to change to the HDPlex H5 case so I can plug in the cards directly to the motherboard.

 

 

Like others, thanks for the meticulous documentary of your build (I always admire anyone who can keep the thermal goo neat on the side of a Streacom ;).  Keep in mind Anwar's thread that Streacom is finally going to make an FC10 rear plate for the desired brackets by year's end.

[elan120]

For the people have done building Unregulated power supply for the Taiko DC-ATX, what transformer secondary voltage output you are using for your build?  I recall somewhere Nenon mentioned higher voltage will provide better sounding, but can't recall what the optimum output should be, and I am thinking to order a transformer that will provide close to 40V of output.

[ASRMichael]

1 minute ago, elan120 said:

For the people have done building Unregulated power supply for the Taiko DC-ATX, what transformer secondary voltage output you are using for your build?  I recall somewhere Nenon mentioned higher voltage will provide better sounding, but can't recall what the optimum output should be, and I am thinking to order a transformer that will provide close to 40V of output.

@Exocer? I believe it’s 36v is optimal but Rob can confirm. 

[elan120]

38 minutes ago, ASRMichael said:

@Exocer? I believe it’s 36v is optimal but Rob can confirm. 

Sounds good, I will wait for the confirmation from Rob.  I plan to use this power supply for my upcoming endpoint build experiment, hopefully it will provide the same magic in the server application.

[Exocer]

1 hour ago, elan120 said:

For the people have done building Unregulated power supply for the Taiko DC-ATX, what transformer secondary voltage output you are using for your build?  I recall somewhere Nenon mentioned higher voltage will provide better sounding, but can't recall what the optimum output should be, and I am thinking to order a transformer that will provide close to 40V of output.

 

1 hour ago, ASRMichael said:

@Exocer? I believe it’s 36v is optimal but Rob can confirm. 

Transformer secondary we used is 24VAC. After rectifier it gives around 36.5-37VDC

[Tatomek7]

Hi  @elan120,

In my case it’s 400VA with primary 230VAC, secondary 24VAC. 

[elan120]

@Exocer, @Tatomek7,

 

Thank you for the confirmation, I will get the parts ordered and see how it works in my endpoint build application.

[Exocer]

7 minutes ago, elan120 said:

@Exocer, @Tatomek7,

 

Thank you for the confirmation, I will get the parts ordered and see how it works in my endpoint build application.

Awesome.

Are you coming from a multi-rail regulated LPS?

[elan120]

30 minutes ago, Exocer said:

Awesome.

Are you coming from a multi-rail regulated LPS?

Are you talking about the power supplies used for the endpoint?  If yes, then no, I don't have anything at the moment, everything still in planning / research stage, but I do want to get the unregulated power supply project moving, and try to get some of the long lead time parts ordered.

[NanoSword]

Hi folks ,

 

I'm using the Power from @Soul Analogue with Taiko ATX.

The performance is very promising.

It's the first time I've turned it on before the transformer and cables are settled and everything is new .

The stage is wide the sound is more holographic with new layers and the background is black. The resolution took a big leap . The natural sound is very excellent vocals are beautiful too. I've never tried any server power before, but I can say Soul Analogue ULPS is pretty good, and the parts he using look amazingly large, From the first sight and from the weight I can tell that the parts are carefully selected.

 

The connector is really big and heavy military grade .

 

The chassis is very heavy, the quality is excellent. I like the black color, especially the signature It looks great and there is manual soft start design I need to push the button before I pull the toggle down I think it is more safety.

 

 

What I know about Soul Analogue ULPS is that all the pieces he uses are custom made according to his request and personal taste.

 

 

I'm using a Ghent cable to power my computer at the moment and hope one day I can try MSG. I trust Nenon a lot. He helped me a lot in making many choices and I am confident in his taste. 

 

As for the power I got from the Soul Analogue from round one, I didn't find the sound too thin or bulky, it's well balanced. My system is more transparent than before.The shape of the musical instruments also looks great more real. There is also an improvement in control, speed and dynamics.It's a huge improvement for me.

 

Thank you everyone, I have benefited a lot from the experiences in this topic.

 

My server has 3950x processor, Jcat XE and SR4T, Euphony + PGGB.

[Soul Analogue]

6 hours ago, ASRMichael said:

@Exocer? I believe it’s 36v is optimal but Rob can confirm. 

No!! for 36vac secondary from transformer… it will be around (36x1.4 - 2) =48.4vdc for cap input topology….very risky

24-28vac secondary for cap input psu

 

[Soul Analogue]

1 hour ago, NanoSword said:

hi Nanosword

thank you for your favourable impressions on my work

there will be further improvement of sound quality in the coming weeks when the parts are settled..

 

I would like to make clarifications on the components used:

 

Transformer - custom made from my local partner supplier up to my specifications and requirements

Choke - best materials available for this purpose in the market, hand assembled and adjusted with my hands

Capacitors - selected from my years of experience in my diy journey and favourable feedback from friends… they had gone through my proprietary treatments for better performance

Cable - wires are copper core and also selected from my experience… the wires are carefully braided for low inductance and noise rejection… i think MSG wires may perform better in some areas.. but not on naturalness and musicality…

better wire i can think of is duelund tin plated wires in cotton

 

Happy listening 😄