Ethernet switch

Ahh, glad to see confirmation that you agree the bits are not actually changing.

It wasn’t clear that was generally understood, especially considering some of the apparent dramatic sonic improvements some people are supposedly hearing - vocal clarity, removing brightness, higher highs, lower lows, 20% better etc etc - just from changing Ethernet Switches or Cables :rofl:

Apparently, these audiophiles are able to hear the effects of Ethernet Switch/cable noise, but dCS R&D can’t and supposedly doesn’t know how to measure it nor how to mitigate it to not impact sound quality :man_facepalming:t2:

2 Likes

“Socrates reminds us that an acute awareness of our own ignorance is always the first step toward knowledge.”

Ahh, glad to see you confirmation that you ackowledge that I’ve never ever suggested they were :slight_smile:

Be careful, very careful, lest you conflate two things which shouldn’t be conflated. A switch installed just before the streamer will mitigate (ideally eliminate) noise picked up by any ethernet cable - there may be a separate debate to be had about shielded vs unshielded cables from router to switch (or router-adjacent switch to pre-streamer switch), but all of these things can be heard from installing a switch. I’m not going to get into percentages, I’ll leave the measurements to you… but these are changes definitely worth having.

Apparently so. Doesn’t it suck.

You really must stop implying that anyone anywhere with any system feeding a dCS DAC can’t possibly hear the benefit of installing an ethernet switch before their streamer. You have no experiential basis for this assertion, just a wish!

Have a great weekend,
Nigel :sunflower: :dove:

2 Likes

Question: does positioning of the Roon Nucleus (close to or far from the system) affect sound quality? My Nucleus box is currently far from the system.

You might get more responses as a new thread. However, while you’re here :slight_smile:

What is your playback chain and what are the connections (and lengths of) between the components? And physical distances between those components today?

1 Like

More or less all of us agree to the fact that bits are not distorted when transported over ethernet, either in a switch or in ethernet cables, as long as they are in compliance with the standards, and have been tested as such.

I was wondering what these tests are about, and found the below Application Note:

Physical Layer Compliance Testing for 1000BASE-T Ethernet.pdf (3.7 MB)

To ensure reliable information transmission over a network,
industry standards specify requirements for the network’s
physical layer. The IEEE 802.3 standard defines an array of
compliance tests for 1000BASE-T physical layer. These tests
are performed by placing the device under test in test modes
specified in the standard.

Tested are:

  1. Template, Peak and Droop conformance
  2. Master jitter
  3. Slave jitter
  4. Distortion, Return loss and Common Mode voltage

It is a well written note, and even a rookie can learn a few things from it.

3 Likes

My current digital chain is as follows:

Roon Nuucleus → tp link switch → optical fiber converter → Ethernet → SoTM snh 10g (with SoTM sps 500 power unit) → Vivaldi upsampler (full Vivaldi rack) → external Mutec ref 120 se → dCS Vivaldi dac → preamp (Gryphon Pandora) → amps (Gryphon Essence mono). I’ve added three pink faun lab isolators between some of the devices and the network.

The Roon Nucleus is inside a cabinet where the router also sits and is approximately 20 meters away from the system. All the Vivaldi units sit on the same rack as the SoTM and the Mutec.

1 Like

Thanks for this:

  • Roon Nucleus to TP-Link: CATn? Length?
  • ethernet to SotM - How long? Is this the optical cable?
  • SotM to Vivaldi upsampler: ethernet, how long? 1m or less?

I’m basically keen to understand if most of the 20m is covered by optical cable.

1 Like
  • Roon Nucleus to TP-Link: Cat 6 / 1 meter
  • ethernet to SotM - 1 meter / yes it is the optical cable
  • SotM to Vivaldi upsampler: 1 meter
1 Like

No, it’s the last mile.

If it is helpful, the fiber line should be able to run ~1km without losses (according to MSB literature).

1 Like

This looks optimised to me. You’re covering most of the distance with optical cable which won’t pick up RFI/EMI noise along the way so you’re giving the SotM as little work as possible to do before the signal reaches the Vivaldi.

Relax, and enjoy the mysic!

1 Like

I’m trying to find ways to make sound quality on streaming as good as possible as SACD or analog (my other sources). Is this possible?

I think that’s what we all seek in our streaming!

I do know of some people who have ditched Roon, despite its obvious convenience, because they found the sound quality without it marginally better. I’m sure this depends on streamer, DAC, amp, speakers… and I know many people who are perfectly happy and wouldn’t dream of moving away from Roon.

It’s amazing what can affect sound quality. In preparation for perhaps one day moving away from my Puccini/U-Clock and higher up the dCS ladder with a pure DAC, I bought a well-engineered CD player as a transport. I tried the same CD in the Puccini and from the CDP as tranport into the Puccini, and the Puccini transport beat it by a mile (this is a Puccini DAC with upsampling on all the inputs by the way). No contest. How can “only” a transport make such a difference? No idea. But clearly everything counts.

Best wishes,
Nigel

2 Likes

I want to try to test replacing my current power units in the switch and the optic converter with linear units and also audition the new etherregen which is going to come to market soon.

I love linear power units so would wholly support this move.

Without being competitive, and I do 100% mean that, what is it about the new etherregen which attracts you? As per numerous exchanges, digital is digital and you can ignore any reference to better (more stable/accurate) clocks as they can’t have an impact on sound quality. Do you think this will be lower noise than the well-regarded SotM and the current generation etherregen? Lower noise is really all a switch can offer vs any other.

1 Like

Lower noise and also it’s ability to connect to the Mutec clock.

I can see the attractions of attaching everything to the same master clock but, as discussed elsewhere and earlier in the thread, this won’t/can’t make any difference to the performance of an ethernet switch because of the way ethernet works - frames/packets, buffering, error checking. You should therefore focus only on the noise mitigation performance.

Hope this helps

1 Like

The Ethernet physical layer is so mature that equipment vendors rarely even think about it, they just use off-the-shelf PHY chips from manufacturers like Intel, Qualcomm, Texas Instruments, and many others.

4 Likes

Yes, it works wonderfully for bit-perfect data transfer.

That ignores the question of noise that could affect an audio DAC.

1 Like