Jump to content
IGNORED

Article: CAPS v4 Maroubra and Bundoran


Recommended Posts

Chris, the SotM USB card you used; is it the tX-USBhubIn mentioned on SotM's website?

Eloise

---

...in my opinion / experience...

While I agree "Everything may matter" working out what actually affects the sound is a trickier thing.

And I agree "Trust your ears" but equally don't allow them to fool you - trust them with a bit of skepticism.

keep your mind open... But mind your brain doesn't fall out.

Link to comment

Been curious about this one for over a year so I finally ordered one today.

 

 

USB audio power supply, galvanical isolated, audiophile USB Power supply, isolated USB low noise transformer, USB DAC power supply, audiophile power supply

 

If you want a lower cost outboard USB power supply, try this one. Another advantage is that it requires only one USB cable as the power injector is a USB-B female to USB- B male end.

 

I currently use it with my Bel Canto U-Link with excellent results.

Link to comment
Chris, the SotM USB card you used; is it the tX-USBhubIn mentioned on SotM's website?

 

That is interesting. Is that some sort of the SOtM PCIe USB card replacement ?

Adam

 

PC: custom Roon server with Pink Faun Ultra OCXO USB card

Digital: Lampizator Horizon DAC

Amp: Dan D'Agostino Momentum Stereo

Speakers: Magcio M3

Link to comment
That is interesting. Is that some sort of the SOtM PCIe USB card replacement ?

I'm not sure ... I thought (but may be wrong) that the SotM tX-USBhubIn is more akin to something like the iFi iUSB - more a power filter for USB.

 

I'm slightly guessing that this is what Chris has used (he hasn't so far responded to my question) but the connectors look similar to those shown in the manual on the SotM site (from the link above).

Eloise

---

...in my opinion / experience...

While I agree "Everything may matter" working out what actually affects the sound is a trickier thing.

And I agree "Trust your ears" but equally don't allow them to fool you - trust them with a bit of skepticism.

keep your mind open... But mind your brain doesn't fall out.

Link to comment
Chris, the SotM USB card you used; is it the tX-USBhubIn mentioned on SotM's website?

 

Yes. The tX-USBhubIn was commissioned from SOtM specifically for this project. Also, this week Andrew received the first batch of the boards and he now has all the components to build the Maroubra server.

 

CAPS-v4-Maroubra-CAPS-Maroubra

 

Jesus R

Link to comment
Actually yes you can. Here is a link to the work around. :)

 

Intel NUC D54250WYKH: Installing LAN Driver on Windows Hyper-V Server 2012 R2 | JDeployment

 

Yes I have seen a few people get this working. I have also seem some people who could not get it working. Basically Intel is trying very hard to keep it from working so future updates of the driver or Windows 2012 might break this hack.

 

Don't seem like a good idea to go against Intel.

agillis

Small Green Computer

http://www.smallgreencomputer.com/

Link to comment
I'd like to revisit a section of the CAPS v3 build's write up as it pertains to this build:

 

mSATA drives are much more like computer memory in size and appearance. These drives are solid state and fit directly into the motherboard without any cables. Even though the DN2800MT board has mSATA capability the Carbon design doesn't use this slot. The server is still very easy to build but absolute simplicity was outweighed by the desire for a lower power SSD that requires internal power and SATA cables.

 

I guess my question is, why not use a mSata drive in v3 when it's the only way to go in v4? Is the technology better now? Is it because it's a means to an end in v4? Is it a power thing? Just curious.

 

The CAPS v3 spec came out a while ago. At the time (2012) mSATA SSD drives were not as good as regular SSD drives. They were slower and used less reliable technology. Now (2015) that mSATA SSD drives are as good as SSD drives it makes sense to use them for any build.

agillis

Small Green Computer

http://www.smallgreencomputer.com/

Link to comment
Yes I have seen a few people get this working. I have also seem some people who could not get it working. Basically Intel is trying very hard to keep it from working so future updates of the driver or Windows 2012 might break this hack.

 

Don't seem like a good idea to go against Intel.

 

It is basically the driver signature, I would not call this a hack. There is never a reason to update a network driver once it is working.

 

This is not going against Intel, they just do not see a market to have a server driver for non server devices, but that is not any reason to make it seem like doing this is wrong.

 

It is actually an excellent idea IMO to enable one to run a better sounding OS on one's system if one want.

 

regards

Bob

Link to comment
  • 3 weeks later...
  • 1 month later...

I just spoke with the supplier of the ML320 case. They told me they have no more and will not be manufacturing or sourcing anymore, as the Haswell NUC for which they were designed are no longer in production, replaced they say by the Broadwell chip models. Any recommendations of alternative sources of the cabinet?

Link to comment

I am glad I found this series. I have been looking into the latest NUC. I have a Theta Gen VIII series three which has every type of SPDIF input, including the Single mode laser (or really ELED). The direction I am going to with extract the audio from the HDMI and convert it to SPDIF. ST Micro amongst others have an IC for this. This will allow transmission up to 192 kHz which the Theta can resolve. Next I plan on taking this signal and reclock with an WM8805 or DIV9001 then into a TTL to PECL converter and finally into a DFB laser module with about 1GHz bandwidth and <.1 nS rise time. I had worked with theta, developed their ST Laser interface. Originally we used and demonstrated a DFB laser at CES. The difference was night an day. The cost for the device was $2k. Impractical for even High End audio. So we went to an ELED which performed well. BTW, DFB lasers have come down in price dramatically. Another issue is the short fiber length and with TOSLIK, high index of refraction for the finer and back reflection at the interfaced and also inside the fiber. During this time I manufactured fiber optic cables for Straightwire and Audiquest. These use a higher bandwidth 50 um graded index fiber with angle polishes. If done right the optical interface can outperform all others. Unfortunately, these high performance lasers (these days up to 40 GHz) are in wavelengths of either 1300 or 1550 nm. Not compatible with TOSLINK's 628 nm.

 

I am photonics packaging engineer by trade and device cooling is one of the main hurdles. Lasers have a high thermal flux than u-processors. on top of this they have to typically be kept to a tenth of a degree of the set point over say the military temperature range (-40 to 85 C). A lot different than making sure an IC doesn't exceed 125 C. So to remove the fan I will go with a thermal electric cooler (TEC). With this instead of the processor being some delta above ambient, I can have it at 10 to 20 C below. A simple supply is needed if there is no set temperature. the Heat sinks will work harder, but that's where you want it to happen. This will still use reflex condensers (heat pipes)

 

Modifications to the NC will be, drill holes for the laser interface and PSU, additions of a heat sink to top or bottom or sides, and separate liner supplies for the HIDM to SPDIF conversion board, the NUC itself and the thermal management system.

 

At first glance it appears all this can fit into the NUC case.

Link to comment
I am glad I found this series. I have been looking into the latest NUC. I have a Theta Gen VIII series three which has every type of SPDIF input, including the Single mode laser (or really ELED). The direction I am going to with extract the audio from the HDMI and convert it to SPDIF. ST Micro amongst others have an IC for this. This will allow transmission up to 192 kHz which the Theta can resolve. Next I plan on taking this signal and reclock with an WM8805 or DIV9001 then into a TTL to PECL converter and finally into a DFB laser module with about 1GHz bandwidth and <.1 nS rise time. I had worked with theta, developed their ST Laser interface. Originally we used and demonstrated a DFB laser at CES. The difference was night an day. The cost for the device was $2k. Impractical for even High End audio. So we went to an ELED which performed well. BTW, DFB lasers have come down in price dramatically. Another issue is the short fiber length and with TOSLIK, high index of refraction for the finer and back reflection at the interfaced and also inside the fiber. During this time I manufactured fiber optic cables for Straightwire and Audiquest. These use a higher bandwidth 50 um graded index fiber with angle polishes. If done right the optical interface can outperform all others. Unfortunately, these high performance lasers (these days up to 40 GHz) are in wavelengths of either 1300 or 1550 nm. Not compatible with TOSLINK's 628 nm.

 

I am photonics packaging engineer by trade and device cooling is one of the main hurdles. Lasers have a high thermal flux than u-processors. on top of this they have to typically be kept to a tenth of a degree of the set point over say the military temperature range (-40 to 85 C). A lot different than making sure an IC doesn't exceed 125 C. So to remove the fan I will go with a thermal electric cooler (TEC). With this instead of the processor being some delta above ambient, I can have it at 10 to 20 C below. A simple supply is needed if there is no set temperature. the Heat sinks will work harder, but that's where you want it to happen. This will still use reflex condensers (heat pipes)

 

Modifications to the NC will be, drill holes for the laser interface and PSU, additions of a heat sink to top or bottom or sides, and separate liner supplies for the HIDM to SPDIF conversion board, the NUC itself and the thermal management system.

 

At first glance it appears all this can fit into the NUC case.

 

Some great info and an interesting project ! Keep us posted.

 

BTW - are you saying that TosLink cables have similar issues with reflection as regular cables ? Can you minimise that by using longer TosLink cable ? What would be the minimum receomended lenght ?

Adam

 

PC: custom Roon server with Pink Faun Ultra OCXO USB card

Digital: Lampizator Horizon DAC

Amp: Dan D'Agostino Momentum Stereo

Speakers: Magcio M3

Link to comment

Hi canyoncruz,

 

Welcome to CA! The more experience we have we USB, the more we realize that it has it's has serious flaws. We await with great interest your explorations on optical connection to DAC's

 

May I humbly suggest that you include handling data rates needed for DSD 512 in your studies.

 

Quite a few of us here are upsampling Redbook and Qobuz / Tidal 16/44 streams to DSD 128 and DSD 256 using HQ Player

Sound Test, Monaco

Consultant to Sound Galleries Monaco, and Taiko Audio Holland

e-mail [email protected]

Link to comment
  • 6 months later...
Hi canyoncruz,

 

Welcome to CA! The more experience we have we USB, the more we realize that it has it's has serious flaws. We await with great interest your explorations on optical connection to DAC's

 

May I humbly suggest that you include handling data rates needed for DSD 512 in your studies.

 

Quite a few of us here are upsampling Redbook and Qobuz / Tidal 16/44 streams to DSD 128 and DSD 256 using HQ Player

 

Toslink has a few problems. The cables" the plastic ones have a very high index of refraction and are not AR (Antireflection) coated and hence a large amount of light entering and leaving the fiber will be reflected, Upward of 10%. With Fused Silica (Quartz) it is about 4%. Though Toslink uses LEDs which are broadband, the back relection sets up an optical cavity causing the LED to favor some discrete wave length with in the their broad line witdh. This can cause modulation of the source and modulation withing the fiber due to Etalon effects. Next the band width of the fiber does not really support the crazy low jitter (ps) in SPDIF transmission. Glass toslink uses bundles wgich is slightly better in bandwidth but now you have multiple signals which can modulate (interefometric effects) at the detector. The core size of these fibers still do not support large bandwidths with the toslink sources.

 

With the so-called AT&T systems, they use single mode optical fiber though some where multimode. Single mode is best. But lasers used in them are sensitive to back reflection. There is a type of laser, the DFB (Distributed Feed Back) that have internal grating to stabilize and reduce reflection effects but they are expensive. With the laser signle mode fiber systems back reflection to the laser are the biggest contributor to jittie. Single mode optical fiber can easily transmit 100 GHz as is the knew state of the are these days. A 2.5 Ghz laser and detector (transmitter and receiver) are commodity devises these days.

 

For audio, the fiber runs are extremely short so back reflections withing the fiber are high. There are solutions: For the fiber purchase one with an "APC" polish (And angled conical polish) with AR coating. for 1.3 um. Here's a 5 meter patch cable with APC polished FC type connectors for $10. AR coating will probably add another $20 to $40.

 

FIS Simplex 3mm SM FC/APC-FC/APC Ultra Fiber 5 Mtr - Fiber Instrument Sales

 

For the Transmitter and Receiver: I would go with the 1.25 GHz 1300 nm (1.3um) FB (Fabry Perot) Laser because this one gives you the choice of PECL or TTL and DC or AC coupling.

 

http://appointech.com/TRX/Tx-Rx/SM%201.25G%20Transmitter-Receiver%20_May%2014_.pdf

 

These are similar to what Theta Digital offered but at a way lower Bandwitdh.

Link to comment
Some great info and an interesting project ! Keep us posted.

 

BTW - are you saying that TosLink cables have similar issues with reflection as regular cables ? Can you minimise that by using longer TosLink cable ? What would be the minimum receomended lenght ?

 

You best bet would be to polish with a 6 - 8 Deg angle. What this does is push the back reflection away from the LED by use of its lens and also increase the angle of back reflection within the fiber such that is no longer contained by TIR (total Internal Reflection) and escapes the fiber - but you still have the band width problem as far as low jitter is concerned

Link to comment
Is i3 based NUC is good for a server,which model ? I got PPA V2 card + Terdack LPSU

 

Would like to use the same cabinate to save money ?

 

That should be sufficient for music. If you want to stream 4K video, you might want to look at the high end models. For music, you might also want to look at the Pentium quad core models. As far as keeping the same case, I would suggest the PPA be placed above the cooling fan. and wired down the side to the internal USB 2 connector. The lid is plastic so you could 3D print a taller one which would house the PPA. Intel has the mechanical models available for the case. If you need help, I could CAD up a lid and you could shop it out for fabrication.

Link to comment
  • 7 months later...

Is Small green Computer still selling the completed versions of CAPS? I'm specifically interested in purchasing Maroumba unless there's a better solution out there for an entry level music server with the SOtM USB.

ARC Ref 5SE, ARC Ref Phono 2, Ayre MX-R Twenty monoblocks, VPI Scoutmaster, Clearaudio Concerto, BAT VK- D5SE, Vandersteen 5a, Ayre Codex, Senore microRendu with HD-Plex Linear power supply

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...