Building a DIY Music Server

[prinz]

Guys, sorry to jump in into your discussion.

quick question - is that Taiko ATX already available? If not, then when and where it can be purchased?

[prinz]

@Topk thanks for the info.

 

I just want to add that I begun my interest in PC audio 10y ago, playing that time with Audiophileo, Sonicweld Diverter, good USB cables etc. Finally, I was able to dig deep into music presentation and unleash all the details, but the sound was never "musical" or should I say - free of issues caused by poor clock/power/vibration performance.

User experience was even worse (one PC as source, second as server, playing music files using command line (WTF!?).

 

I moved to CD players and started to build a CD collection, I had flagship Ayon, Sony DAS-R1A with dedicated transport, CEC-TL2 and few others. Not SOTA, but I would say quite high end products.

 

Month ago a friend of mine visited me with his simple DIY server with Pink Faun USB OCXO card. Next day I sold my CD setup and week after I sold all CD collection.

While I was disappointed 10y ago with the performance of file music playback, now even TIDAL with Roon sounds better than the CD players I had contact with. Nice to be back and I am glad that this pc audio path developed in such a positive way.

 

All the best to you all.

[prinz]

41 minutes ago, Topk said:

Same experience here. There’s a constant battle for me against a cold and sterile sound. Any component that brings any of this, I do not include in my chain. And I didn’t want to say that publicly yet... but let’s say for now that I also use Pink Faun and I find it musical. But that’s only one piece of the puzzle. Again, personal taste, system synergy etc. I have Harbeth speakers so that says already quite a lot about my type of sound.

 

 

I also listen on Harbeth although mine are tuned to sound slightly as studio monitor

[prinz]

I am also burning in a simple PSU recently:

 

Toroidy 300VA transformer with 17VAC output --> shottky bridge --> 33.000uf Epcos Sikorel *B41550 --> Hammond 195G10 choke (5mH, 10A) --> 100.000uf Epcos Sikorel B41550

 

this is the unregulated part and then is a PCB board with 10A LT1083 voltage regulator board with additional 66.000uf Mlytic at the input,  pair of original LT1083 high current regulators followed by Elna Silmic 1000uf on the output.

 

Currently under burn in period.

It already makes amazing difference in low frequency response, bass control and sound authority. It's quite surprising to see and hear how a computer transport power supply creates an illusion, like I would replace power amplifier with a stronger one.

 

 

 

 

*(I advise you to read about these on the internet, mainly diyaudio, still 9pcs available on mouser: https://pl.mouser.com/ProductDetail/epcos-tdk/b41550e5339q000/?qs=yE89RSRv7gKTkC1o8%2bPmmQ%3D%3D&countrycode=PL&currencycode=PLN )

 

 

[prinz]

I made some comparisons between CLC and CLCLC filtering using PSU Designer and have to confirm that CLCLC provides much better voltage ripple smoothing and filtering.

Better in terms of numbers of course.

 

1,68mV   on the output of CLC filter and 70nV after CLCLC filter.  Not sure how it influence the sound, but math wise - CLCLC is 24x times less ripple. Hmmm....

[prinz]

component on the TAIKO, next to Mundorf Mlytic is not a choke. It is most probably a Mundorf EVO foil capacitor mounted into wooden holder, made from panzerholz.

 

Guys, do yourself a favor and learn how to use free tool called PSU Designer. It's easy, widely available and it will allow you to simulate the basic performance of your power supply. You can try different values, different schematics etc.

 

If someone wants to use LC type of filtering for whatever reason, the basic formula to calculate choke inductance is:

L = V (dc) / I (mA)

For example, if you have 30VDC after your bridge rectifier and your maximum load is 8A:

30V /8000mA = 3,75mH --> this is the minimum required choke value

The higher the load, the less inductance is needed.

 

 

 

 

 

[prinz]

This is my LPS prototype

for second choke I have used regular inductor for speaker crossover. I had 1.8mH on hand which fits and simulates very well.

 

[prinz]

Are there any users here that changed regular chipset oscillator / quartz on computer motherboard for Crystek or similar quality clock?

I changed LAN clock twice and worked like a charm, but I am unable to successfully replace clock on the Asus motherboard chipset. Both times I tried computer does not boot and nothing I tried (bios reset, different points of clock signal injection) helps.

 

what are other users experience?

 

[prinz]

I have quick question regarding apacer modules:

Is this one still generally recomended: https://allegro.pl/oferta/apacer-pamiec-ddr4-8gb-2666mhz-cl19-1-2v-12212625685 

I do have asus Z390M plus motherboard which works from what I remember only with nonECC ram.

[prinz]

Guys, have you tried going with only one CAP in PSU? Transformer+bridge+cap

I did and like the sonic benefit of not paralleling capacitors. 20-40kuf of capacitance is really enough even for very power hungry server.

[prinz]

4 hours ago, MarcelNL said:

are you sure? how do you define 'enough', with what goal in mind? 

"enough' means that from both engineering and empirical point of view adding more capacitance does not bring improvement.

 

First I played with PSU DESIGNER tool to simulate the beheviour of PSU and tried to design one with less ripple. Result was CLCLC with overall ~~200kuf and two chokes. Almost zero ripple, so technicaly perfect, but sound wise - too slow, low frequencies dominant, I also heard a bit of distortion due to effect of paralleling the capacitors.

 

Then I started to modify and listen. Several atteempts were made. Transformer and bridge, then the following:

 

1/        2x 33.000uf --> choke --> 100.000uf --> choke --> 66.000uf

2/        2x.33.000uf --> choke --> 100.000uf

3/        2x33.000uf --> choke --> 66.000uf (mlytic)

4/        1x33.000uf mlytic -->choke --> 1x 33.000 mlytic

5/        1x 33.000 uf mlytic

 

last one I found the best. Flat frequency response without any frequency dominant, faster transient response, dynamic, sharp presentation with very high resolution and much less distortion.

Well, to be honest this does not surprise me as in the best when I owned many SE tube amplifiers like audio note, kondo, verdier, sun audio , I modified, DIY, and designed my own amplifiers. Every time the result was the same - simple PSU - better sound, single cap (no paralleling)= better sound.

 

it worked for me, doesnt have to work for you. But I am pretty sure many of you here building your power supplies, just blindly follow posted instructions without doing your own tests, research and evaluation.

I believe its worth to check if "less is more" or if such complex PSU's which are posted here are REALLY needed for top level performance.

 

 

[prinz]

10 hours ago, Nenon said:

 

Yes, that's V1. If transparency is top priority, you can't beat that. A bit lean sounding for me and not as immersive as I would like it to be. Maybe good for a small nearfield bookshelf speakers system but not for a big full scale system in my experience. 

I am curious what amps/speakers you are using @prinz.

 

I use Harbeth 40.0 Mastering Monitor

Regarding amps, two months ago I sold Ypsilon Phaethon integrated and moved to Circle Labs P300 pre + M200 power amp (these are absolutely amazing for me as a former tube lover).

https://highfidelity.pl/@main-1140&lang=en

[prinz]

9 hours ago, Nenon said:

@prinz I just found my old post from over an year ago:

I think I agree with that description :). 

BTW, the reason you don't see a capacitor mentioned here is because there is one on the Taiko rectifier PCB board. 

 

OK. I understand.

In my build I skip the fuse and soft start, but it's true if you want to stay on the super safe side it's better to keep them.

 

I will never say that my point of view is the absolute best, I play with audio for years and faced hundreds of times that what works for one, doesnt have to work for others. It's normal :)

Luckliy for us all, we have a nice playground for tests. I want to share what worked for me.

 

My PSU does not sound lean, actually If it would - I would throw it away quickly :). I love full, massive, full bodied, colorfull sound.

I didnt used these mosfet bridge and have no idea how they work, but generally speaking I could face slower sound due to the fact that schottky bridge which I use is very smooth sounding, but it is not the fastest one, which means it could be difficult for the bridge to "charge" 200kuf of capacitance. Who knows.

on the other hand, what is slow for me can be lightning fast for others and opposite.

 

Last thing I want to check is if replacing Mlytic with Kemet ALS70 will bring the improvement for me. I am awaiting delivery.

[prinz]

11 hours ago, MarcelNL said:

I hear what you are saying, listening tests are important, and more so than what the simulation shows!

At some point I want to hear what a tube rectifier PSU does with the Taiko ATX...I only am beginning to play around with the PSU yet it clearly has quite an impact where it should not have on a CPU ;-)

 

 

tube rectifier PSU for high A current?

Can you please share more details about the design?

 

I ask as regular tubes in full wave rectifier mode can offer 300mA of constant current. If you use four tubes in greatz mode you can get around 1-1.5A max, but for PC audio use you probably need 5A minimum.

[prinz]

49 minutes ago, sakso136 said:

Do you mike the circle lab better than phaeton????

They are different in presentation. Phaethon is delicate, subtle, it is voiced to be very high resolution oriented, but lacks body, warmth and power. I still liked Phaethon more than shitty Air Tight ATM-211 tube monos I had before, but if we speak about Circle Labs - for me - they are simply amazing.

Speaker drive ability which I never experienced before, combined with warmth, body, constant foot tapping :), energy, high resolution.

All I can dream off.  Imagine yourself a really great sounding tube amp with almost unlimited power, zero distortion and without the need to replace tubes.

wholeheartedly recomended.

 

I dont want to make more offtopic, so If you have other questions please send me PM

[prinz]

someone who wants to go extreme, can check Mundorf zero-ohm inductors:

https://www.mundorf.com/audio/en/shop/coils/MCoil-VN390-Zero-Ohm-Coil/?shop_category=&card=2990&var=true

 

They are perfectly fine to be used in PSU, altough I tried them in speaker crossover years ago. BASS was out of this world :), but overall I highly prefer wax copper foil coils from Jantzen (Polink Poland).

[prinz]

11 hours ago, Gavin1977 said:

Tell us more - how do the two differ in subjective listening?

Its diffucult to compare as overall the tests were done 6y ago , in speaker crossover in woofer section.

To make long story short - low frequecies were strong, punchy, amazing dynamic, precise, with great control of the speaker. This is typical behaviour of high quality thick solid core wire (and such one is used in mundorf).

Unfortunately woofer was also influencing midgrade heavily and here we need the coil to offer colorful, rich, warm, high resololution sound.

At the end we used 8AWG Jantzen Wax speaker coil for woofer instead of Mundorf, as we prefered rich sound in the midrange and good (but not perfect) bass.

 

As said, these tests were done in speaker, not in audio PC PSU :), also, remember that for PSU you need very low DCR (low resistance). 0,3R is the maximum which I would use in PSU If someone would force me to use choke.