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

[seeteeyou]

http://thelinearsolution.com/ocxo_switch.html
http://thelinearsolution.com/ocxo_router.html
http://thelinearsolution.com/ocxo_lan.html
http://thelinearsolution.com/ocxo_usb3.html

Quote

- Military Spec OCXO
- 2oz Copper PCB +/- 10ppb
- Ultralow Jitter <1ps RMS

 

Same specs for all of the above, and then their USB 3 adapter was clearly showing that $55 Connor-Winfield OH4610LF-025.0M

 

http://www.conwin.com/datasheets/cx/cx259.pdf

http://www.conwin.com/product_locator.html?ModelFam=Oscillators&ProdType=OCXO&Recommended=Y&Pkg=14 Pin DIP
https://www.digikey.com/product-detail/en/connor-winfield/OH4610LF-025.0M/CW899-ND/5641634

 

The cat was already out of the bag so there were no beans to be spilled whatsoever.

[seeteeyou]

Reference clock:

 

10MHz

 

Word clock:

 

Multiples of 22.5792 MHz / 24.576 MHz

 

System clock:

 

Depending on what we're talking about (e.g. 25MHz for Ethernet, 12MHz / 24MHz for USB 2.0 etc.)

 

BTW, it's pretty much like dealing with Traveling Salesperson Problem

 

http://www.cs.princeton.edu/courses/archive/fall14/cos126/assignments/tsp.html

 

The longer route we take, the more degradation we'll have to face. Anyways, it's kinda like a "better than nothing" strategy that's less than ideal if you will. Paying a lot for an excellent 10MHz OCXO in the first place, power that with one of the best PSU (such as SR7) as well, of course throw in another expensive clock cable.

 

Degrade that all we want but that's still performing quite a bit better than the average system clocks, even if the same level of performance were achievable with an affordable (let's say $55) OCXO inside a switch.

[seeteeyou]

For some relatively "special" frequencies such as 19.2 MHz, we could still go for stuff like Neutron Star 2

 

http://www.newclassd.com/index.php?page=200

 

In some cases sCLK-EX could very well be the only choice whenever we actually need an "oddball" frequency. For instance, 14.318 MHz is required for motherboards but only cheapest of the cheap ones could be found on Mouser.

 

For some reasons I looked very hard for the clock with 14.318 MHz but this particular Asus motherboard might be an exception with "abnormally" low number of system clocks

 

https://www.asus.com/Motherboards/PRIME-B250-PLUS/

 

http://www.computeraudiophile.com/forums/topic/38199-sotm-sclk-ex/?tab=comments#comment-801085

On 3/30/2018 at 6:47 PM, Chrisc10115 said:

I can resolve the issue now. Confirmed by SOTM. 

 

For the maniboard ASUS Prime B 250 plus:

24 MHZ for the mainboard. The ehthernet comunication crystral  25 MHZ.

Voltage:  3,3 Volt.

 

More importantly, how do we actually determine what system clocks are present throughout an entire music server (and identify each frequency) without sending everything to SOtM? Even stuff like SSDs and optical drives etc. have their own system clocks when we crack them opened. 

[seeteeyou]

Granted we've got two different frequencies below but it's just a matter of putting prices into perspective

 

$55 for ±10ppb (25 MHz)

https://www.digikey.com/product-detail/en/connor-winfield/OH4610LF-025.0M/CW899-ND/5641634

 

$94 for ±5ppb (25 MHz)

https://www.digikey.com/product-detail/en/connor-winfield/OH300-50503CF-025.0M/CW943CT-ND/7691417

 

$760 for ±0.5ppb (10 MHz)

https://www.digikey.com/product-detail/en/abracon-llc/AOCJY6-10.000MHZ-5/535-11917-ND/3641393

 

$1,222 for ±0.3ppb (10 MHz)

https://www.digikey.com/product-detail/en/abracon-llc/AOCJY6-10.000MHZ-2/535-11918-ND/3641392

 

$1,800 for ±0.1ppb (10 MHz)

https://www.digikey.com/product-detail/en/abracon-llc/AOCJY6-10.000MHZ-1/535-11919-ND/3641391

 

If we're only looking for system clocks that are usable without "resorting" to the discipline of 10 MHz reference clock, the "only" one with ultra-high stability might be this one from Vectron that's good for anywhere between 4 MHz and 15 MHz

 

https://www.vectron.com/products/ocxo/dx-040.htm

 

Depending on the model that's ordered, they're offering either ±0.4ppb or ±0.2ppb but so far we couldn't figure out what the actual price really is without getting a quotation.

 

Other than that, the only useful frequencies from 4 MHz to 15 MHz should be 12 MHz for USB 2.0 and 14.318 MHz for motherboards respectively.

[seeteeyou]

5 hours ago, Johnseye said:

Is phase noise really the only thing we care about in a clock?  Nothing else?

 

We could also get an OCXO (27MHz) from http://www.oppomod.com and the phase noise turned out to be nothing to write home about

 

http://www.sbtron.co.kr/korean/product/pdfs/ocxo_sboc_25_sine.pdf

 

However, even its optical output would perform as good as the USB output from that ZENith SE + tX-USBultra + Mutec REF 10 combo when they're connected to either Chord DAVE or Hugo 2

 

https://www.head-fi.org/threads/review-comparison-of-5-high-end-digital-music-servers-aurender-n10-cad-cat-server-totaldac-d1-server-auralic-aries-audiophile-vortex-box.787020/page-78#post-14117821

 

Since DACs from Chord should be jitter agnostic, that might or might not be applicable for other brands. In addition, Toslink couldn't support anything above 192kHz while not all brands could accept DoP via Toslink.

 

---

 

Regarding Crystek, good luck with getting any worthy models with 25MHz unless we're buying in quantities

 

https://www.digikey.com/products/en?FV=ffec2025

https://www.digikey.com/products/en?FV=ffec741f

https://www.digikey.com/products/en?FV=ffec7420

 

https://www.computeraudiophile.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/?page=205&tab=comments#comment-751270

On 12/7/2017 at 6:02 AM, RickyV said:

It is indeed difficult to find/ buy one crystek 575-25  25 MHz (fs105 clock).

The one on eBay is 56 dollar! That is a bit expensive, i thought thay were 8 dollar or so.

You also need a good regulator module too. Darn it is getting expensive again.

 

---

 

I would love to get something from http://www.pulsarclock.com myself if I were looking for a word clock, its compatibility with DIL14 clock sockets should be fantastic for stuff like Vanity203 HD or Ian Jin's McFIFO + McDualXO etc. However, we've been talking about system clocks so far while they don't seem to offer any frequencies (at least the standard ones) that we actually need.

[seeteeyou]

46 minutes ago, rickca said:

Good specs mean good potential outcome but good design/implementation is critical to realize that potential.

 

IMHO even something as good as The Linear Solution OCXO switch could be further improved

 

http://thelinearsolution.com/ocxo_switch.html

 

They only mentioned "ultra stable regulator design" but most likely we could do better with LT3045.

 

Besides, they're including their own LPS with 9V output while the OCXO circuit would require 12V instead. In other words, it should be a idea to feed the OCXO part with LT3045 @ 12V separately. (i.e. avoiding the penalty of a single transformer with multiple rails while reaping the benefits of LT3045 such as ultra-low noise and ultra-high PSRR.)

[seeteeyou]

1 hour ago, Superdad said:

 

We have about 600 CCHD-575 @ 25MHz on the shelf.  How many do you want?  

 

Thank you very much, it's great to have another source for those of us who only need a clock or two.

 

OTOH, why don't we invest more (€ 135.2 shipped outside EU) to get Neutrino 2 since the 22.5792 MHz version (NOT 25 MHz) seemed to beat CCHD-575 handily

 

http://www.singxer.com/pd.jsp?id=19

https://www.computeraudiophile.com/forums/topic/29553-my-very-mini-review-of-the-singxer-su-1-ddc/?page=52&tab=comments#comment-788425

On 2/27/2018 at 6:33 AM, rafa said:

Before the changes hires, so 48/96/192 sounded better than 44kHz in many aspects. Mostly from sound stage perspective.

Currently with Neutrino 44 get incredible boosted. 3D dimensional. Vocals more natural. But amazing is, it sounds better than hires.

https://www.computeraudiophile.com/forums/topic/29553-my-very-mini-review-of-the-singxer-su-1-ddc/?page=52&tab=comments#comment-788745

On 2/28/2018 at 12:07 AM, RickyV said:

 

Nice job rafa,

i am amazed at the difference in SQ between the 575 and the neutrino. You sould do the 24 too.

Now the neutrino will be staying you can shorten that clock sig cable, cm not dm.. as short as possible. It directly related to SQ. On newclassd site it says 10 cm max

 

[seeteeyou]

Supermicro X11SBA-LN4F is the big brother of X10SBA-L with quad Ethernet ports, about 200 bucks a piece and it could be an interesting option for killing multiple birds with one stone

 

https://www.servethehome.com/supermicro-x11sba-ln4f-review/

https://www.newegg.com/Product/Product.aspx?Item=N82E16813182988

https://www.supermicro.com/products/motherboard/X11/X11SBA-LN4F.cfm

 

4 system clocks @ 25MHz for quad Ethernet ports while an OCXO only costs $55. We could use that as NAS / Roon Server / Router while using another X10SBA-L only as an endpoint. Obviously it's also fairly cheap to replace a system clock or two on the other side so we're going through superior 25MHz clocks twice in a row

 

https://www.digikey.com/product-detail/en/connor-winfield/OH4610LF-025.0M/CW899-ND/5641634

 

[seeteeyou]

FYI - we could get the bulk wires of Supra CAT 8 STP for 9 bucks per meter (with free shipping)

 

http://www.jenving.com/products/view/cat-8-stp-patch-frhf-blue-b150-1001800547

https://www.ebay.com/itm/Supra-CAT-8-STP-cable-bulk-for-DIY-Made-in-Sweden-BEST-FOR-NETWORK-AUDIO/152930653863

 

CAT 8 connectors

 

http://www.jenving.com/products/view/rj45-cat-8-male-ethernet-plug-bulk-1011001375

https://www.ebay.com/itm/Supra-RJ45-male-ethernet-plug-for-use-with-CAT-8-cables-STURDY-/401501783442

 

https://store.wireworldcable.com/collections/ethernet-cat8/products/rj45-plug-for-starlight-cat8-ethernet

https://www.amazon.com/WIREWORLD-Cat-Ethernet-Plug-RJ45/dp/B01M0V6CUE

 

http://www.cablematters.com/pc-938-62-6-pack-tool-free-cat-8-cat8-shielded-rj45-field-termination-plug.aspx

https://www.amazon.com/Cable-Matters-Tool-Free-Shielded-Termination/dp/B074HHWNVN

 

http://www.delock.com/produkte/G_86485/merkmale.html

https://www.amazon.de/DELOCK-Stecker-feldkonfektionierbar-Cat-8-Metall/dp/B073X53FGB

[seeteeyou]

First of all, it's great to have Nuno from Innous joining us on this thread and thank you very much for your participation.

 

We should be able to understand if you weren't too happy about revealing "too many" details regarding that custom motherboard inside Statement. However, would you please kindly tell us just a little bit about bypassing the regulators of all motherboards in general?

 

For instance, we could feed 1.3V directly to RAM slots with the best linear regulators (i.e. 6 pieces of LT3045 @ 3A) with modifications while that seemed to result in a significant improvement

 

http://www.diyaudio.com/forums/vendor-s-bazaar/310402-fs-ultra-low-noise-power-supply-lt3045-based-pcb-6.html#post5353416

 

Similarly, feeding 5V directly to this PCI-Express card with LT3045 also yielded great results without having to replace the original system clocks with OCXO etc.

 

https://www.startech.com/Cards-Adapters/USB-3.0/Cards/4-port-usb-3-1-card~PEXUSB314A2V

Now it's only "natural" to ask the question, how far do we wanna go in terms of bypassing the regulators? While LT3045 should be fairly affordable to begin with, we still have to think about what do we put in front of LT3045.

Besides, is it actually feasible to feed the right voltage directly to something like Celeron J1900 inside X10SBA-L since the voltage is VERY low (only 1.0V or so) while the amperage should be fairly high.

Last but not least, is it OK to discuss more about measuring the noise pattern of so many motherboards with a detailed oscilloscope? Was X10SBA-L really THAT outstanding or did you find anything that's reasonably close? I'm just wondering if its big brother X11SBA-LN4F were also a decent choice or otherwise

https://www.supermicro.com/products/motherboard/X11/X11SBA-LN4F.cfm

Thanks again for chiming in and looking forward to your answers.

[seeteeyou]

IMHO this is very so interesting for owners of DACs from Chord, provided that we're OK with the limitations of Toslink

 

https://www.head-fi.org/threads/chord-electronics-dave.766517/page-736#post-14206068

Quote

1) Mac mini running Roon Server > USB to Dave
2) Mac mini running Roon Server > AQ-SWITCH-SE > sMS-200ultra > tX-USBultra > USB to Dave
3) MacBook Pro (battery powered) running Audirvana Plus > optical to Dave

 

To cut the long story short these are my main findings:
1) Again the differences are not night and day but indeed solution 3 offered the most enjoyable, full and non-fatiguing sound. Unfortunately for several reasons this is not a sustainable solution (placement, storage capacity and use of the laptop for other purposes).

 

https://www.head-fi.org/threads/chord-electronics-dave.766517/page-736#post-14206142

Quote

Yes I have the camera adapter and will try with my iPad.

 

https://www.head-fi.org/threads/chord-electronics-dave.766517/page-737#post-14207030

Quote

iPad on USB: the iPad connected to USB came second but very very close to the optical and completely smashed the USB out of the Mac Mini via SOtM chain!

 

Granted it's gotta be DAC-specific because Chord DACs are supposed to be jitter-agnostic, though in specific cases we might wanna question the value of investing in certain products from SOtM if we're indeed happy to live with Toslink.

 

So far it's quite consistent with the findings of @romaz who praised the Toslink inputs of both Hugo 2 and DAVE when they're fed by a modified Oppo UDP-205.

 

That wasn't even end of the story, here's yet another example with PS Audio DirectStream fed by the Toslink output of this $50 USB DDC

 

https://item.taobao.com/item.htm?id=20827883149

https://www.head-fi.org/threads/chord-electronics-dave.766517/page-733#post-14202845

Quote

I wish I could agree. I spend two nights, say 4-5 hours each night, listening to my new USB-to-TOSLINK converter. And compared to my current USB with 15 ferrites (and tested with/without an Intona and Battery inserter on the +5V), its no comparison. The TOSLINK converter with the new XMOS U8 is just remarkable ...kills the USB. I hear things, beautiful things like subtle nuances in instrumentation and background vocals. OK, this is to a PSAudio DirectStream while I wait for my Qutest to arrive ...but still. Given that these DACs are supposed to be immune to SPDIF jitter ...is this the miracle of zero RF noise impacting the DAC? I am a single sample, but now my Yoga laptop thru this $68 converter is the best I have heard. I can live with a max of 24/96 to get this quality of sound for so little cost.

 

Both 22.5792 MHz and 24.576 MHz clocks (their "flagship" TCXO would cost 90 RMB so that's like just under 15 bucks a piece) should be absolutely nothing to write home about, not to mention the fact that USB DDC was actually bus powered to boot

 

https://szjyec.taobao.com/category-639720104.htm

 

Finally this is definitely the craziest example by far since we're talking about freebie Toslink cable here, we could turn off the upscaling part of M-Scaler inside Blu Mk. 2 and then convert a single coaxial output into Toslink. That just happened to beat the daylight outta dual coaxial outputs that were blessed by the power of 1 million taps

 

https://www.head-fi.org/threads/chord-electronics-dave.766517/page-735#post-14204373

Quote

I have a spare LPS-1 (after upgrading to the LPS 1.2) which I use to power it - 5v/1A

 

The Toslink cable supplied with the Dave goes from the convertor to the Dave. All upsampling is turned off, so it’s just plain vanilla Red Book into the Dave. It sounds glorious and much better than BNC Blu 2 to Dave. I’m afraid the interface between the RFI generated by the BLU 2 and Dave’s ground plane is a major issue for me.

 

So yeah, those 1 million taps inside Blu Mk. 2 made a huge difference for sure but we're also paying the price of a major "RF noise generator" just like how Rob Watts would slap a label on quite a few products

 

https://www.head-fi.org/threads/chord-electronics-blu-mk-2-the-official-thread.831343/page-196#post-14103674

Quote

He described my Clarity Cables Natural USB cable as an "RF noise generator" and he thought it sounded the worst of the lot.

Quote

Once again, Rob labeled the Zenith SE an "RF noise generator."

 

It's just incredible for me to realize what's the deal with both USB and coaxial inputs of Chord DACs, both of them seemed to be super sensitive to RFI while Toslink actually saved the day for those of us who don't need anything above 192kHz.

[seeteeyou]

1 hour ago, auricgoldfinger said:

Will any generic toslink cable achieve this performance, or do we need a certain quality cable?

 

Once again it's gotta be DAC-specific so YMMV if you weren't trying Toslink with DACs from Chord or PS Audio etc. Basically we don't have to spend that much on Toslink cables, and in @romaz's case the Toslink output of his modified Oppo UDP-205 was able to achieve the same level of performance as his ZENith SE + tX-USBultra + Paul Hynes SR7 + Mutec REF 10 combo (NOT quite as good as his megabuck custom server with tons of bells and whistles but close enough?) with this Mapleshade Clearlink optical cable

 

http://shop.mapleshadestore.com/Optical-Audio-Cables/products/338/

 

Seriously the megabuck part was such an understatement because we're talking about well over 10K but that's actually working so darn well after @romaz specifically tailored everything to match his entire system

 

https://www.head-fi.org/threads/review-comparison-of-5-high-end-digital-music-servers-aurender-n10-cad-cat-server-totaldac-d1-server-auralic-aries-audiophile-vortex-box.787020/page-69#post-13801299

Quote

 

I'll get a few things out of the way. Both of these units are excellent and the consensus was unanimous that in my system as stated, either of these units is a large gap better than a Windows 10 laptop on batteries running Roon. It was also a unanimous consensus that my current server build is better than either of these units but in different ways. For those not aware of the details of my current server build, here it is:

 

https://www.computeraudiophile.com/forums/topic/30376-a-novel-way-to-massively-improve-the-sq-of-computer-audio-streaming/?page=153&tab=comments#comment-724987

 

To be fair, what I have now was built and fine tuned for my specific system and to my preferences and so not surprisingly, using gambling nomenclature, it has a "house advantage." For those not wishing to build their own servers, either of these would be considered state of the art and I would have no reservations recommending either of them depending on your priorities.

 

 

BTW, keeping clock cables as short as possible might be a challenge for some of us who have multiple devices with 10MHz reference clock inputs.

 

More importantly, pretty much all owners of Chord DACs who were quite happy about Toslink seemed to have fairly lousy PSUs / clocks for those relatively cheap source components. USB DDC was bus powered. Macbook Pro was battery powered. Oppo UDP-205 was powered by (cheap?) linear PSU with multiple rails. The only noteworthy one turned out to be LPS-1 for Lindy 70411. No fancy clocks whatsoever and even that 27MHz SBtron OCXO inside Oppo should be meh at best.

 

In other words, nobody even got halfway serious about optimizing those sources in the first place but the results for those jitter-agnostic DACs from Chord / PS Audio should already speak for themselves. That $50 USB DDC turned out to have DIL14 clock sockets, we could take advantage of some clocks with fantabulous phase noise (such as Pulsar Clock) without much degradation when both clocks could sit right on top on the clock sockets

 

http://www.pulsarclock.com/ENHome.html

[seeteeyou]

1 hour ago, austinpop said:

All that said, if Chord come out with an "affordable" Mscaler and DAC (TT2?) combo, I'd be interested to see how the Toslink idea pans out. I suppose one can continue to use USB for the DSD and DXD formats.

 

Well, the thing is we'll definitely have to "settle with" either USB or dual coaxial in order to utilize a million taps while Toslink could only go up to 192kHz. It kinda sucks when the upscaling part of M-Scaler and the advantage of Toslink should be mutually exclusive.

 

We've got USB 3.0 AOC (Active Optical Cables) but they aren't backward compatible with USB 2.0 unless we're adding that cheap-looking adapter

 

https://www.lindy-international.com/USB-Fibre-Optic-Extenders.htm?websale8=ld0101.ld020102&ci=500303

 

This one from Corning should be somewhat flaky since it might not even last that long

 

https://www.corning.com/optical-cables-by-corning/worldwide/en/products/usb-optical-cables.html

 

These guys might look interesting but once again we've gotta know more about backward compatibility

 

http://spec.go4fiber.com/networking/gusb-xx-xxxx_20151120.pdf

https://www.go4fiber.com/usb-3-0-to-lc-active-optical-cable-length-1m

https://www.go4fiber.com/usb-3-0-to-usb-3-0-active-optical-cable-length-1m

 

Could this one from VIA be good for $239 by any chance?

 

https://www.viaembeddedstore.com/shop/accessories/usb-3-hybrid-aoc/

https://j3fau2wsnso1qkiarz7keon5-wpengine.netdna-ssl.com/wp-content/uploads/2018/02/VIA_USB_3.0_Hybrid_AOC_datasheet_v170926.pdf

Quote

Compatible with a full range of USB standards, the cables also deliver superior signal quality and high EMI and RFI resistance.

 

Even replacing the system clocks inside Adnaco-S3B with sCLK-EX seemed to be no good so @romaz didn't keep that.

 

BTW, did @romaz ever give you a clear answer offline?

 

https://www.head-fi.org/threads/review-comparison-of-5-high-end-digital-music-servers-aurender-n10-cad-cat-server-totaldac-d1-server-auralic-aries-audiophile-vortex-box.787020/page-70#post-13805109

Quote

Fascinating findings! But just to be clear, after all the dust has settled: your custom server, with end to end clocking and SR7 PSUs, is still sounding superior to the Zenith SE, correct?

 

[seeteeyou]

1 hour ago, tgb said:

I've been using this Corning for 3years now : very good.

The only point is that you need to cut the power from the computer & power it with better PS ; I use a power bank, even better SQ than using a modded LPS (and far cheaper, so why bother...).

 

Thanks a lot and it's good to know that maybe some batches might be better than the others, though it's still hard to tell why there were so many horror stories reported by other owners before

 

https://www.computeraudiophile.com/forums/topic/22816-corning-optical-usb-cable-experiences/

https://www.computeraudiophile.com/forums/topic/24759-how-i-resurrected-a-corning-usb-cable/

 

For $109 maybe it's worth the gamble?

 

https://www.amazon.com/Optical-Cables-Corning-Cable-Meter/dp/B00JOJRF6K

https://www.bhphotovideo.com/c/product/1057862-REG/optical_cables_by_corning_aoc_acs2cva010m20_type_a_plug_to.html

 

Some PCI Express cards have connector(s) for external power so that could be the way to go for USB 3.0 cards.

 

BTW, there's something similar from Monoprice for $98.99 and it's backward compatible with USB 2.0 by default

 

https://www.monoprice.com/product?p_id=16377

 

[seeteeyou]

Oh well, Corning seemed to be no good for galvanic isolation

 

https://www.computeraudiophile.com/forums/topic/22816-corning-optical-usb-cable-experiences/?page=10&tab=comments#comment-416520

On 4/28/2015 at 2:55 PM, JohnSwenson said:

I received a corning cable today. I have not had time to try it out, but I did do some continuity checks:

the shield and pin 4 (USB GND) are connected at both ends AND between ends. There is a low resistance path between the GND pins on both ends, thus this cable is NOT galvanicaly isolated. The resistance is 3 ohms, much higher than most cables, but not "isolated" by any means.

 

This is what I expected given the copper wires inside designed to power the device end.

 

Tomorrow I'll hook it up try some listening with it.

 

John S.

 

https://www.computeraudiophile.com/forums/topic/22816-corning-optical-usb-cable-experiences/?page=10&tab=comments#comment-416676

On 4/29/2015 at 4:02 AM, JohnSwenson said:

This morning I checked some "normal" USB cables I had and they all wound up with around 0.3 ohm for the GND (pin 4) connection. So the corning has about 10 times the ground resistance of normal copper cables.

 

The 3 ohm of the corning is higher but by no means can be called galvanically isolated, that would take at least a mega ohm. The 10 times higher resistance CAN have an affect on ground loops in some circumstances, and not in others. For example if there is another ground path in parallel to the USB cable that is around 0.5 ohms, the ground current will be shared between them. Replacing the USB cable with the corning will cause most of the ground current to go through the other path, significantly decreasing ground current through the DAC. But lets say the ground source impedance is 75 ohms and the USB cable is the only path for a particular loop segment, the change from 0.3 to 3 will make almost no difference.

 

My personal take is that most of the sound improvement of the corning is that it is generating a new USB signal right at the DAC, sort of like what the regen is doing, but not as good.

 

Technical stuff below:

 

As an interesting tidbit, the corning cable cannot be sending straight USB over the optical path. USB is a differential bidirectional bus (the same physical wires are used to send data in both directions). The corning contains two fibers, one for data going in one direction and the other for data in the opposite direction. Thus it has two unidirectional single ended connections. In USB there are some conditions where the two wires are not differential (both hi or both low), this cannot be directly represented on one single ended connection. There are two ways to deal with this: use a different protocol that can handle these extra states, or add one or more states to the analog signaling, for example instead of just having the light on or off add an additional part way on state or two to represent the non differential states.

 

My guess is that corning did the different protocol approach, which also has the big advantage that they might be able to get around the delay issues by putting their own protocol in place between the endpoints.

 

I'll try and get some listening in tonight.

 

John S.

 

[seeteeyou]

1 hour ago, tapatrick said:

Thanks for pointing this out @seeteeyou !

 

You're very welcomed.

 

1 hour ago, tapatrick said:

The common understanding 5 years or so ago was that Optical was inherently jitter prone. However this conversation on HeadFi does raise some puzzling and interesting questions.

 

Maybe the mitigation of jitter back then wasn't as good as what we've got now? It's certainly a viable option to go for Toslink these days if DXD and DSD128 (or above) weren't such a big deal.

 

1 hour ago, tapatrick said:

I also tried the Optical input via a cheap DDC and it also sounded very good.

 

That could very well be the "strangest" part if you will, there are so many great things out there while they don't always cost that much to begin with. LT3045 from Alexey (0.5A and 1A) and Michael (3A and 5A) are fantastic as long as we don't need anything more than 15V / 5A.

 

Recently Bob also discovered the transformational power of passive PoE with quite a few shielded CAT 7 short cables, they almost cost next to nothing and we're good to go as long as the current draw ain't too high.

 

Then we could also combine both LT3045 and passive PoE to feed an extremely cheap USB hub plus a card reader, the only one that's "somewhat" expensive should be this 4GB SLC card for $65 or so

 

https://ciiva.com/part/rp-smsc04da1-20938590

[seeteeyou]

On 4/30/2018 at 1:59 AM, austinpop said:

All that said, if Chord come out with an "affordable" Mscaler and DAC (TT2?) combo, I'd be interested to see how the Toslink idea pans out. I suppose one can continue to use USB for the DSD and DXD formats.

 

No standalone M-Scaler yet but here's TT 2 for 4,000 quids with balanced outputs (a must-have because single-ended outputs of LAu kinda suck) for your Liquid Gold, featuring dual Toslink inputs and accepting 12V DC power

 

https://www.whathifi.com/news/chord-unveils-hugo-tt-2-dacheadphone-amp-high-end-2018

http://chordelectronics.co.uk/product/hugott2/

Quote

1.5m 12v Switching power supply

 

Not sure why but it didn't really look like 12V to me

 

[seeteeyou]

3 hours ago, TheAttorney said:

Interesting. A few weeks ago I asked Adrian for the option of dropping

 

I know it ain't the nicest thing to say (especially in public) but I just couldn't help myself because this is way too funny

 

 

So, where did I get the inspiration?

 

[seeteeyou]

On 5/21/2018 at 9:38 AM, charlesphoto said:

I would be interested if anybody’s compared the sound of the TLS switch to a Cisco 2960? These can be picked up used for about $30 and the sound quality is night and day compared to a Netgear, even with an LPS and JSGT. 

 

I could find two clocks from the pictures below, most likely that should be quite a bargain once we replace both clocks?

 

http://youtu.be/HJLqF7Gg5WQ

http://youtu.be/VyvxMbIFJGo

 

 

Once we've figured out what the voltage of 341-0208-01 really is, it should be fairly easy to power that with something better than the default SMPS

 

Cisco Power Supply 341-0208-01 for WS-C2960-8TC-L
https://www.ebay.com/itm/Cisco-Power-Supply-341-0208-01-for-WS-C2960-8TC-L-/361750014670

[seeteeyou]

I looked at Table A-7 Catalyst 2960-8TC-L, 2960G-8TC-L, 2960-8TC-S, and 2960PD-8TT-L Switch Specifications below, all 8-port models seemed to share VERY low current draw that's under 1A 

 

https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst2960/hardware/installation/guide/2960_hg/higspec.html

Quote

48 VDC, 0.3 A (Catalyst 2960PD-8TT-L)

 

In other words, it's fair safe to assume that all 8-port models should require 48V and obviously it wouldn't be any good for stuff like LPS-1.2 or LT3045 etc.

 

A single piece of LT3045 could only output up to 15V @ 0.5A so we'll need at least 4 of them (12V) in series

 

https://www.ebay.com/itm/LT3045-S-Ultralow-noise-0-8-Vrms-LDO-linear-regulator-0V-15V-500mA-Fixed-out/253397942111

 

So we're talking about 23 × 4 = $92 for a few LT3045 in series and then roughly 30 bucks for Cisco Catalyst 2960PD-8TT-L respectively.

 

$55 for an OCXO @ 25MHz that's the same as we're getting from TLS, plus the costs of clock mods while adding another piece of LT3045

 

https://www.digikey.com/product-detail/en/connor-winfield/OH4610LF-025.0M/CW899-ND/5641634

 

Finally there's one more clock to mod and most likely that's meant for the CPU itself? It's hard to tell what the frequency is since those pictures weren't exactly clear enough, though either Neutrino 2 or Neutron Star 2 might be good to go

 

http://www.newclassd.com/index.php?page=24
http://www.newclassd.com/index.php?page=200

 

Since it only costs $30 or so, maybe it's actually a better idea to find out what kinda regulators did Cisco put inside their 2960 series since they're probably dropping all the way from 48V to a much lower voltage. End-of-Support Date turned out to be the end of July 2018 and just wondering if they're still willing to answer such an "odd" question like that?

 

https://www.cisco.com/c/en/us/support/switches/catalyst-2960-8tc-l-compact-switch/model.html

https://www.cisco.com/c/en/us/support/switches/catalyst-2960-8tc-s-compact-switch/model.html

[seeteeyou]

On 5/24/2018 at 2:20 PM, austinpop said:

Details of the Habst cable group buy are here.

 

Not sure if Habst were solid core or stranded, and we also couldn't tell the gauge since that wasn't listed.

 

Anyways, maybe it's gonna be worth the effort for those of us who only need 2 outta 4 taps from sCLK-EX. Let's say Habst could actually make something small enough to fit the tiny little U.FL connectors while matching the 50Ω impedance, that's gotta be great for replacing those cheap 50Ω clock cables inside sCLK-EX itself.

 

One U.FL for signal, and then another U.FL for ground. That should be able to mitigate the weakest link of sCLK-EX by reducing the degradation.

[seeteeyou]

An old school BIOS like this seemed to have even more options

 

 

AMD motherboard from MSI

 

https://www.tweaktown.com/reviews/8175/msi-x370-gaming-pro-carbon-motherboard-review/index5.html

[seeteeyou]

On 5/27/2018 at 7:10 AM, tboooe said:

I wonder how the 2960 compares to the Cisco SG300 series which I have.  

 

Cisco SG??? Series should be meant for the small business market, OTOH they're targeting their Catalyst Series (namely 2960 and 3560) at the enterprise sector.

 

So far the latter one seemed to have the "magic sauce" baked right in while nobody actually confirmed if that were the same deal for the former one

 

http://forums.naimaudio.com/topic/cisco-switch?page=8

Quote

I have a Cisco SG110 24HP switch sitting awaiting installation. How, in your view, does this product compare to the (widely recommended) 2960?

 

Quote

Hi, I honestly don’t know... I haven’t evaluated it... I suspect however it may not bring the same benefits as the catalyst models... as it is probably built differently to a lower specification as it is designed and marketed to a low cost market ... but I am speculating... certainly in my experience not all Cisco devices sound as good as the 2960/3560 series Catalyst switches

 

FYI - the "magic sauce" might be explained according to the same gentleman who happened to be a pro in a very specific field
 

Quote

 

I am also a network design engineer with a background in system and RF engineering, and in my line of work it makes total sense... by the way you are right it nots specific to the 2960 including the new models, but it does seem specific to the various Catalyst switches I have tried including layer 3 Catalyst switches... The EM response to physical differential line driver layouts and the PHY clock stability and its lack of intermodulation products would appear to be helping. The effects or otherwise are taking place before (or after depending on your perspective) of the ASIC(s)

 

What I have found in my experience is that it  is nothing or very little to do with the devices powersupply. But being an engineer I observe (and listen) and then look to evaluate why one experiences an outcome. I believe it was myself who first discovered this ‘benefit’ of using, as it was then, a 2960 Fast Ethernet switch device a few years back, when I discovered it by accident when I was was narrowing down an issue of Ethernet intermodulation PHY layer clock frequency products from a particular home network switch product affecting stereo decoding birdies on my FM tuner... and I investigated further.

 

i have also found and analyzed sound differences between different media servers based in inter Ethernet frame timing consistency... I had shared my findings with Naim... and well what I can say is that the network  streaming module design of the newer devices has changed.... its good to have an open mind and the where with all and tools to investigate further ... you never know little discoveries have the possibility of improving our Hi-Fi products further...

 

 

In other words, that's definitely a great baseline to start with and the price is certainly right. Now let's take a quick look at the specs and we could learn more about the DC input voltages for RPS 675 (Table A-4 Specifications for the Catalyst 3560-24TS-S Switch)

 

https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst3560/hardware/installation/guide/3560hig/hgspecs.pdf#page=3

Quote

12 V @ 5 A

 

In other words, maybe we could find a way to feed that with 10 piece of LT3045

 

https://www.mpaudio.net/product-page/hc-hpuln

 

Granted the ones with 24 ports ain't fanless to begin with, while their fanless 8-port model would come with a different PSU

 

https://www.ebay.com/itm/CISCO-341-0207-02-Power-Supply-for-Cisco-WS-C3560-8-/232627232528

https://www.surplustechmart.com/lab-test-equipment/power-supplies-loads/cisco-341-0207-02-power-supply-for-cisco-ws-c3560-8.html

 

Most likely there will be one clock @ 25 MHz plus another one for the ASIC itself. Let's say we're replacing both of them with something like OCXO, there's a chance that we could beat most audiophile offerings (until we're getting something new from UpTone Audio) without having to spend that much in the first place.

[seeteeyou]

Does it sound like what we've been discussing on this thread or what?

 

http://thelinearsolution.com/one_server.html

Quote

Ditch those spaghetti setup, This is THE ONE Server you ever needed.

 

[seeteeyou]

17 minutes ago, Confused said:

Has it got a clock input??

 

Actually they should be trying to avoid the degradation of the connection between music server and external clock, IMHO that's exactly why they're kinda making fun of SOtM etc. with the very first sentence quoted above.

 

In other words, it could very well be a proof of concept that direct connection to the motherboard itself (i.e. without cheap 50Ω cables that are terminated with tiny little U.FL connectors) is cheaper and yet potentially superior to all that Pasta Parade IMHO. Keeping the distance between the clock(s) and the associated components to a minimum so that degradation should help to solve the traveling salesperson problem.