Network Isolation

[lmitche]

With nearly a month since the last post on this thread, would it be fair to say many who have previously posted in support of using fmcs now see the "direct connection" approach as the better way to go? Fmcs have improved my system but I have not yet tried direct connection.

Yes, a direct connection is best.

[marksas3]

I couldn't get the direct connection to work, at least when using Roon with the microRendu, but I may try it again soon.

[lmitche]

Here is another thread that details the procedure. Best to start from the top.

 

http://www.computeraudiophile.com/showthread.php?t=31110

[ted_b]

I couldn't get the direct connection to work, at least when using Roon with the microRendu, but I may try it again soon.

 

Yes, I have yet to get it to work too. Here is thread

http://www.computeraudiophile.com/f10-music-servers/novel-way-massively-improve-sq-sms-200-and-microrendu-31110/

[lmitche]

I have never been able to get hqplayer to find an NAA from a bridged Windows 10 machine. This is why I got started down the single machine route without an NAA.

 

Nevertheless others have made this work by adding a third, usually wireless, network adapter to the bridge.

 

It's all in the thread referenced above.

[bogi]

I have never been able to get hqplayer to find an NAA from a bridged Windows 10 machine. This is why I got started down the single machine route without an NAA.

 

Nevertheless others have made this work by adding a third, usually wireless, network adapter to the bridge.

 

It's all in the thread referenced above.

 

Larry, I found 2 ways how to find NAA connected to separate NIC. Look at my posts in HQPlayer and HQPlayer NAA threads of the last week.

 

a) Windows NAA can be found in Windows HQP computer when both NAA and HQP computer have enabled IPv6 (IPv4 need not be enabled). Nothing more to configure. Use the latest NAA version. http://www.computeraudiophile.com/f11-software/hq-player-20293/index329.html#post642246

 

b) Linux NAA was found by Windows HQP by using conenction sharing. No bridge necessary. This method could function also without IPv6 (I didn't try). See here: http://www.computeraudiophile.com/f22-networking-networked-audio-and-streaming/hqplayers-network-audio-adapter-13892/index87.html#post643317

[tboooe]

With nearly a month since the last post on this thread, would it be fair to say many who have previously posted in support of using fmcs now see the "direct connection" approach as the better way to go? Fmcs have improved my system but I have not yet tried direct connection.

I actually still use fiber and FMCs in a pseudo direct connection between my HQP server PC and my NAA PC. My server PC has a dual port fiber nic in it so I use 1 FMC between it and the NAA. The sound using fiber and the FMC between the computers is a bit smoother than a direct connection using a cat 6 cable which sounded a bit too detailed and ragged. Dont get me wrong, the differences are subtle. Plus my nerd side likes the idea of having fiber in my system!!!

[tboooe]

Larry, I found 2 ways how to find NAA connected to separate NIC. Look at my posts in HQPlayer and HQPlayer NAA threads of the last week.

 

a) Windows NAA can be found in Windows HQP computer when both NAA and HQP computer have enabled IPv6 (IPv4 need not be enabled). Nothing more to configure. Use the latest NAA version. http://www.computeraudiophile.com/f11-software/hq-player-20293/index329.html#post642246

 

@bogi, thank you for this post. For me using a bridge is required because I find I sometimes have to remote desktop into the NAA PC to restart networkauiod.exe. My system is completely headless so having the ability to use remote desktop is essential. Setup with a bridge was quite easy...give NAA PC fixed IP on same subnet as the rest of my network, create bridge, enjoy music!

[bogi]

I couldn't get the direct connection to work, at least when using Roon with the microRendu, but I may try it again soon.

 

I found interesting Miska's opinion - he considers direct connection to be worse than connecting NAA and HQPlayer computer to a switch. He considers setting a bridge (like recommended in the above mentioned thread) to be non suitable solution ... read here (and some posts around): http://www.computeraudiophile.com/f22-networking-networked-audio-and-streaming/hqplayers-network-audio-adapter-13892/index87.html#post643163

 

On the other side, I found Larry's recommendation to use separate NIC for NAA. AudioPhil recommends that also in his AO setup guide. And people in the above mentioned thread reported improvement when they connected NAA to separate NIC. Are Miska's arguments correct?

[bogi]

@bogi, thank you for this post. For me using a bridge is required because I find I sometimes have to remote desktop into the NAA PC to restart networkauiod.exe. My system is completely headless so having the ability to use remote desktop is essential. Setup with a bridge was quite easy...give NAA PC fixed IP on same subnet as the rest of my network, create bridge, enjoy music!

 

I am doing Remote Desktop via IPv6 addresses. No bridge needed for that ... People are setting bridge to help HQP computer to discover NAA via network broadcast, but that's not needed for Remote Desktop.

[tboooe]

I am doing Remote Desktop via IPv6 addresses. No bridge needed for that ... People are setting bridge to help HQP computer to discover NAA via network broadcast, but that's not needed for Remote Desktop.

Can you please explain? So on the second port of the HQP nic and the NAA nic you set IPv6 and just connect them? Did you set any IP addresses or did you just leave it for both as automatic? You are then able to remote desktop into NAA?

[bogi]

I corrected the 1st link in my post #834

[bogi]

My answers in the quote...

 

So on the second port of the HQP nic and the NAA nic you set IPv6 and just connect them?

YES

 

Did you set any IP addresses or did you just leave it for both as automatic?

AUTOMATIC

 

You are then able to remote desktop into NAA?

YES

 

What you need then is to find the IPv6 address and your interface number on HQP computer ... But if you succeed to discover NAA by HQP then HQP log tells you the NAA IPv6 address. Then you can use that address in Remote Desktop.

[tboooe]

My answers in the quote...

What you need then is to find the IPv6 address and your interface number on HQP computer ... But if you succeed to discover NAA by HQP then HQP log tells you the NAA IPv6 address. Then you can use that address in Remote Desktop.

 

Did you disable IPv4?

 

I assume its ok if I assign a fixed IPv6 address to the NAA nic?

[bogi]

Did you disable IPv4?

 

I assume its ok if I assign a fixed IPv6 address to the NAA nic?

 

On Win10 HQP computer I let IPv4 configured for DHCP, as on other place I am connecting with cable to router.

 

On Win10 NAA I have IPv4 disabled, as I found I don't need it.

 

I didn't try fixed IPv6 addresses.

[bogi]

Hi, did anyone try this LAN isolator, how it works for audio purposes?
http://www.ebayshopkorea.com/main/view?itemID=331057579170

[Solstice380]

14 hours ago, bogi said:

Hi, did anyone try this LAN isolator, how it works for audio purposes?
http://www.ebayshopkorea.com/main/view?itemID=331057579170

I am using a Baaske isolator and it works very well.  I do not know the one that you referred to. 

[asdf1000]

On 2/15/2017 at 5:49 PM, mattjtaylor2809 said:

I also suspect that one or both of the media converters is faulty as I've done the following:

I'll speak to the supplier here in Australia to see if I can get some replacements.

 

 

Hi Matt,

 

Did you end up getting your MC110CS replaced in the end? And was the SQ result good?

 

Cheers, Sean

[mattjtaylor2809]

Hi Em2016,

 

I did replace the media converters but I'm still having issues getting the damn things to work.

I think I'm getting close as I'm finally seeing some activity of the green LED's suggesting I have an active link (TP) but when I connect to the computer I can't get access to my network due to a unassigned IP issue?

 

I'm not sure this is necessary for the link to work in front of my PS Audio DSJnr DAC's bridge II but I'll find out soon enough when I reinstall.

 

Fingers crossed!

 

Cheers

Matt

 

If interested, the latest info can be found on this link:

 

http://www.psaudio.com/forum/directstream-all-about-it/use-of-fiber-optic-to-ethernet-converter-with-bridge-ii/page-3/#p74685 

[mourip]

I have been following a slightly different path than most in that I am using an ethernet to AES converter to feed my DAC. My ethernet connection goes NIC>cat6 >FMC with stock power supply>fiber >FMC with LPS>cat6 > Focusrite Rednet D16 >AES cable to DAC

 

When I first set up the FMCs I was basically blindly following what others had reported here and on Head-Fi as sonic benefits from the fiber isolation with the second FMC run by an LPS. At the time I was also making other changes and so I never did much evaluation. I just left the FMC's in place and moved on.... until today.

 

With my system in pretty much a steady state and a good sense of familiarity with its sound I decided to remove the FMCs and just run straight ethernet. I was very surprised at how much of a positive difference the FMCs made as was evidenced once they were removed.

 

They definitely give my system a more realistic tonal balance, more detail, and a more convincing sound stage. I tested by playing back the same familiar music with and without the FMCs. I mention this just to add a few data points for the use of FMCs for isolation with an LPS to power them.

[Superdad]

On 7/7/2017 at 7:13 PM, mourip said:

With my system in pretty much a steady state and a good sense of familiarity with its sound I decided to remove the FMCs and just run straight ethernet. I was very surprised at how much of a positive difference the FMCs made as was evidenced once they were removed.

 

They definitely give my system a more realistic tonal balance, more detail, and a more convincing sound stage. I tested by playing back the same familiar music with and without the FMCs. I mention this just to add a few data points for the use of FMCs for isolation with an LPS to power them.

 

Thanks for posting your report.  This is further evidence that the signal integrity delivered to the Ethernet PHY matters--similar to the way it matters to USB PHY and input stages.

Also still being researched--but well evidenced by all those who are playing around with better clocking for their upstream devices--is how the phase-noise "fingerprint" of packet data interfaces can possibly survive and impact SQ.

 

I'm sure the "traditional measurement" folks around here would love to ridicule and burn me at the stake for stating the above.  But most in this thread who have been experimenting and listening for a while know that this is far from self-delusion.

 

Yet the mechanisms--and points to directly measure--are not fully identified just yet.  Of course that is no reason to stop experimenting and using your ears to judge the comparisons.  That's not only how great systems are assembled, it is also part of how great audio products are developed.  Ask most any audio engineer whose been at it for a decade or few. 

 

[Sorry if the above seems a bit of a rant.  Guess I'm just a bit weary of the belitlement going on in other threads towards all those who report honestly what they hear.  This thread seems like a safe space where an observation can be made without expecting a barrage of ridicule.]

[mourip]

2 hours ago, Superdad said:

 

Thanks for posting your report.  This is further evidence that the signal integrity delivered to the Ethernet PHY matters--similar to the way it matters to USB PHY and input stages.

Also still being researched--but well evidenced by all those who are playing around with better clocking for their upstream devices--is how the phase-noise "fingerprint" of packet data interfaces can possibly survive and impact SQ.

 

I'm sure the "traditional measurement" folks around here would love to ridicule and burn me at the stake for stating the above.  But most in this thread who have been experimenting and listening for a while know that this is far from self-delusion.

 

Yet the mechanisms--and points to directly measure--are not fully identified just yet.  Of course that is no reason to stop experimenting and using your ears to judge the comparisons.  That's not only how great systems are assembled, it is also part of how great audio products are developed.  Ask most any audio engineer whose been at it for a decade or few. 

 

[Sorry if the above seems a bit of a rant.  Guess I'm just a bit weary of the belitlement going on in other threads towards all those who report honestly what they hear.  This thread seems like a safe space where an observation can be made without expecting a barrage of ridicule.]

 

Thanks and I could not agree more with your last statement.

 

Continuing on...I have been assuming that the principle benefits of using fiber for the ethernet connection was the isolation from grounding issues?

 

Yesterday after my confirmation of the benefits of fiber between my PC and D16 I decided to use a second pair of FMC's to isolate my PC from its LAN connection. I have two ethernet ports in my server, one for the LAN and one dedicated just for the audio connection to the D16. To my surprise once more I felt that I had additional similar positive effects. I am now considering getting a PCIe fiber card for my server to replace the Intel NIC and cut down on the increasing maze of devices and cables.

[Crom]

14 minutes ago, mourip said:

QUOTE: "I am now considering getting a PCIe fiber card for my server to replace the Intel NIC and cut down on the increasing maze of devices and cables."

 

:-) I did exactly the same about 18 months ago. I ended up with fibre directly out of my server going into a fibre switch...unfortunately, I didn't run fibre around the house when I rewired otherwise I would have fibre to the renderer!

[jabbr]

3 hours ago, Superdad said:

 

Yet the mechanisms--and points to directly measure--are not fully identified just yet.  Of course that is no reason to stop experimenting and using your ears to judge the comparisons.  That's not only how great systems are assembled, it is also part of how great audio products are developed.  Ask most any audio engineer whose been at it for a decade or few. 

 

Not everything is understood. Fiberoptic ethernet has always emphasized SI and relatively low power in its design. In particular the power requirements for copper Ethernet rise significantly as bandwidth and length rise. Fiberoptic ethernet is a good green technology comparatively -- doesn't pollute the air either  

[asdf1000]

7 hours ago, Superdad said:

This is further evidence that the signal integrity delivered to the Ethernet PHY matters--similar to the way it matters to USB PHY and input stages.

 

Hi Alex,

 

As someone very new to using FMC's powered by your LPS-1's (I've found this combination to make an incredible improvement) I just wanted to ask, is this more to do with less noise (the littler ethernet transformers don't isolate everything) than signal integrity, or a combination of both?

 

And like you, I hope we can remain safe in this thread to discuss our observations openly  lol

 

Cheers, Sean.