A novel way to massively improve the SQ of computer audio streaming

[RickyV]

No by neutron star 2 clock, done by me. 

The only thing I modded on the ISO regen. 

[RickyV]

https://www.ebay.com/itm/Hi-end-3-way-S11-Ultra-low-noise-linear-Power-supply-DC7-5V-DC12V-DC19V/123063908030?hash=item1ca72e36be:g:u6AAAOSwIwJZf2au

[RickyV]

2 hours ago, BigGuy said:

I subscribe to this threadand yet have not gotten any notifications since I last posted on 04/28 tho' the posters here continue to be quite prolific!

Anyone else with this issue?

 

Hi BigGuy,

no i do not have that problem.

Sometimes you get a reaction sometimes you do not. 

[RickyV]

At least you know now the notification machine works 

[RickyV]

Or power cables.

 

Or supra cat8 cable.

I did it with copper foil tape and braided tinned copper sleeve. 

With the copper foil is not the best idea, the cable gets stiffer this way

[RickyV]

2 hours ago, rafa said:

Hi guys, my friend replaced crystal oscillator motherboard with NewClassD NeutronStar. Is extremally happy with sound effect. Especially the soundstage is bigger, more clear and precision. The system is stable.

 

But integrated graphics card (Intel CPU) seems to don't like new clock. Time to time screen get black or flashing on the bottom (it is not a monitor problem). When screen is off music still plays as normal, so that's why it looks like system is still stable.

Interesting that simple switching lights on in the room may destabilize graphics card.

The NeutronStar is powered by NewClassD PSU supplied with a clock. Clock voltage is setup to 3.3V

 

Any ideas what can make those problems?

 

Any pics ?

[RickyV]

1 hour ago, Johnseye said:

Does anyone know where to find very small diameter sheaths that could be placed around an RF cable like this:

https://www.digikey.com/product-detail/en/te-connectivity-amp-connectors/2015357-3/A36228-ND/1249192

 

Also, what kind of shielding could be used?

 

I'm guessing this is too wide.

https://www.amazon.com/gp/product/B01BIBQE1S/ref=ox_sc_saved_title_2?smid=A38Z015AJ7I755&th=1

 

I'm looking to shield the cables used to connect to the sCLK.

 

http://www.techflex.nl/metaalafscherming-c-146.html?language=EN

 

 

[RickyV]

23 hours ago, rafa said:

Hi guys, my friend replaced crystal oscillator motherboard with NewClassD NeutronStar. Is extremally happy with sound effect. Especially the soundstage is bigger, more clear and precision. The system is stable.

 

But integrated graphics card (Intel CPU) seems to don't like new clock. Time to time screen get black or flashing on the bottom (it is not a monitor problem). When screen is off music still plays as normal, so that's why it looks like system is still stable.

Interesting that simple switching lights on in the room may destabilize graphics card.

The NeutronStar is powered by NewClassD PSU supplied with a clock. Clock voltage is setup to 3.3V

 

Any ideas what can make those problems?

 

The only thing I can think of is, did you use a good ground for the clock signal cable. So a neer by groundplane that measures 0 ohm with mobo mounting screw. 

 

"Interesting that simple switching lights on in the room may destabilize graphics card." If the room light switch has influence on the graphics card I think you may have an ground problem with computer, audio and maybe tv (antenna).

[RickyV]

It also maybe smart to put some shielding on the clocks pcb's and power supply. Computers are noisy emi.

 

What frequentie is the neutron star?

[RickyV]

7 hours ago, lmitche said:

Hi Bricki,

 

For some cables I have used clear shrink wrap for insulation, on others shrink wrap + a layer of techflex for more distance between layers.  The Lush cable is a stiff cable to start, and adding the three to four more layers of stuff makes it much worse. The techflex makes things a bit looser in a good way.

 

At the ends I have just let run the insulation short by about 3/8 of an inch so the two layers of tinned copper naturally join under the pressure of one last short, 3 inch,  layer of shrinkwrap on top.

 

Sorry, but pictures will have to wait as I am not at home.

 

Larry

 

Hi Larry,

The reason why i do not use shrinkwrap is that it shrinks around the cable but only applies a little pressure on the cable, so bad contact. For spacing is a good idea

That is why i use the linen tape used in the automotive industry for wiring harnesses.

Maybe a combination of the two, first pressure then spacing.

[RickyV]

3 hours ago, BigGuy said:

<lmitche> has commented  on his discover of the Gotham Ultra Pro 4/1 and its use in fabbing a JSSG360 but pondered why it is demonstrably better than JSSG with wire.

 

Found this article which shows measured differences between various shield configurations...

<http://www.isgroup-international.com/pdfs/Wire_and_Cable_PDFs/wire_and_cable_electrical.pdf>

 

Very little has been said about muMetal shielding except, IIRC, someone using muMetal foil. 

Magnetic Shield Corp owns the copyright for both muMetal and Co-NETIC tradenames and sells braided shield...

<http://custommagneticshielding.magneticshield.com/viewitems/magnetic-shielding-for-wiring-applications/co-netic-aacableshield>

Prices/foot are $11.95, $12.95, $16.85, and $20.45 for 1/4, 3/8, 1/2 and 3/4 diameters respectively. 

They do have a $100 minimum but that should not be difficult.  ?

 

I could definitely see a JSSG360-MM being the next leap.

 

 

 

WoW good find 

[RickyV]

12 hours ago, Cornan said:

 

Be careful what JS says. It is just his own experiences, meaurements vs common knowledge. The variables are just to great to collect into a single know-how compartment. I am quite sure that there is more than RFI/EMI or inductance at play here! ?

 

Yep, magic fairy dust

[RickyV]

2 hours ago, elan120 said:

Network switch clock mod - part-2 (Big thanks to Alex C.)

 

With input and help from @Superdad, I decided taking some time this morning to give Crystek CCHD-575 oscillator a test drive, replacing the Conner Winfield OH4 I retrofitted a week ago.  Again, after having the oscillator mounted on the carrier board, its measurement look good, and the rest was quite straightforward.  After the new oscillator was retrofitted, initial listening impression is very positive, but I will wait for some additional warm-up and burn-in time to see if there are going to be changes...hopefully even more improvements.

 

Below are some pictures if anyone is interested in finer details...

 

Crystek CCHD-575 initial power up measurement:

 

Completed carrier board:

 

Completed new oscillator retrofit:

 

Carrier board close up:

 

Good job Elan120, very intresting.

[RickyV]

6 hours ago, elan120 said:

Network switch mod - part-3

 

This is my ZyXEL network switch mod part-3, and likely be the final phase other than perhaps little tweak here and there in the future. 

 

The main focus for this mod is in the input power supply section.  Stock switch MB has two switch mode regulators (AP1513), it has fairly high ripple noise at around 22~24mV level, obviously they both are replaced with LT3045-A (1A regulator board), I also removed the stock input bypass cap since it will no longer be used, removed stock output feedback resistor dividers to help lower output current consumption, replaced both output bypass electrolytic caps with low ESR aluminum organic polymer caps, added thermal pads on the bottom of both LT3045-A's to better conduct heat to the switch enclosure, and finally, added EMI/RFI absorber sheet on inside top cover.

 

How do all these work in the end?  An obvious step up in background darkness level, both clarity and detail retrial improved, gained both depth and width in sound stage, makes the music playback more engaging and relaxing.  I am quite pleased with the end result, which proves once again how damaging those low cost switch mode regulators can negatively impact overall SQ.

 

As before, in case anyone is interested in the details, here are some pictures:

 

Stock step down power supply section:

 

 

Input bypass cap, output bypass caps, and inductors removed:

 

 

Both stock switch mode regulators removed, and new input bypass caps installed:

 

 

EMI/RFI absorbor sheet installed on inside of top cover:

 

 

Current carrier board with newly added input aluminum organic polymer cap:

 

 

 

 

 

 

 

 

Nice mod elan.

Do you have a modem/router upstream from this switch? Try giving that one a good psu. I put a ifi ipower on it and liked sq jump. So if that works better regulator and clock would work too, wouldn't it?

[RickyV]

I am thinking if i have the guts to remove the outer sleeve of lush cable and see whats what, and then do a 360 on it.?

[RickyV]

Does anybody know if it is possible to passievely split the pci bus of the supermicro X10Sba, support bifurcation? 

I may what buy one but it needs to support my jcat usb card and i would like to use a optane ssd for the os. Is that possible? 

[RickyV]

Thanks @LTG2010, pitty. The next best thing would be to replace the clock on a normal ssd. But i can’t do that myself.

How is your powersupply comming allong? Do you have some pics?

Would you buy the x10sba again or would buy something different today.

[RickyV]

3 hours ago, hurka said:

iso regen feeding bluwave think not so good.better you use independent power both(I tested with singxer).  Sigma 11 arrive soon,i hope  better than lt1084 board. If Sparky usb and auralic mini similar sq(don't know),then x10sba mobo big jump in SQ.

ucaps powering lps1.2 insane category.....

 

Hi hurka 

you can power the 12v and the 5v from two s11's. The s11 can't do 3,3v so use //lt3045 from 5v to 3,3v. 

I am planning to do the same but with 3045's for the 12v and 5v too.

[RickyV]

Or you can do this with neotech occ wire.

[RickyV]

4 hours ago, lmitche said:

Got it! Thanks John.

Yes it is copper foil tape wraped around NEOTECH STDCP-24 red and NEOTECH STDCP-22 black. I thought it was a good idea to use different gauges in star quad config with two layers of Braided shielding all connected at both ends.

I used it to make an active cable, with a 1A LT 3045 in line, after a lps1 powering my iso regen.

[RickyV]

I was more thinking of

 

But I am cat person ??

[RickyV]

Here is another interesting idea for a diy naa endpoint but it only converts to i2s or spdif. 

https://www.engineered.ch/products/nmr/

 

Thay allso have upsample modules to 384kHz or DSD at 2.8224MHz or 5.6448MHz.

[RickyV]

41 minutes ago, lmitche said:

Well for now. I am staying on the single machine hqplayer dsd512 upsampling bandwagon. Maybe when we have the new Uptone switch and learn how best to deploy it I'll change my mind. I am open minded but skeptical that the impact of upstream processing can be eliminated. John, you have the tls switch, will you be experimenting?

 

A year ago I built a low power, j1900 based single machine Linux/roon rig for a friend. Given Roy's experience, the mini-itx motherboard and case will be replaced with a nuc using the motherboard and emmc memory Roy's describes above. It will be an interesting experiment and good preparation for later possibilities if warranted.

 

Gotta run, the mumetal shielding just arrived!

 

Larry

Hi @lmitche,

Where did you buy the mu metal braided shield from and do you have the specs?

I found this one https://hollandshielding.com/Flexible-EMI-shielding-tube.

It shows also the shielding performance of E-field and H-field.

Also found this one but is not as effectieve in low fields, https://cdn.techflex.com/assets/pdfs/specification-sheets/fsn.pdf.

[RickyV]

10 hours ago, jean-michel6 said:

Ś

 

Mumetal is a blend of nickel and iron  and I really do not understand what is mucopper described by the holland shielding company. 

 

I think the copper shield advertised are effective for RFI interférences for EMI that is another story as you have to use a mumetal foil which is usually quite expensive. 

Yes I found that one too but I see no specs so how effective is it in comparison with the mu copper.

[RickyV]

I had a thought yesterday with analog signals we are trying to avoid ferro/ magnetic metals in signal connectors isn't it. Also near analog cables?

In mu metal alloy  17% or something  is ferro metal. What does it mean for digital signal?

Just a thought.

Somebody else any thoughts on this.