advertisement


Audiophile Network Switches for Streaming ... really ?

But how does the noise get into the streamer, are their network socket implementations all so bad as to be useless?
 
At this point surely everyone will have noted the amusing fact that Shannon’s proof of the sampling theorem was contained in a paper entitled …,https://course.ccs.neu.edu/csg250/ShannonNoise.pdf
Gee whiz if only someone could come up with a symbolic form in which to encode data so that it can be reconstructed perfectly despite the presence of noise in the transmission line.
 
Sue Pertwee-Tyr said:
I do care, but perhaps not in the way which you, and others, require me to care. I care because I choose to live in a universe of wonder and possibilities. I don't want to live in a universe of banal certainty. In my universe, the fact that music exists, and that it has the power over us that it does, is a wonderful thing. As is the fact that we can change the way we perceive music. Whether that is because the equipment does something technical to the music, or because our perception is altered by other means, is to some extent a side issue. I'd genuinely love to know what's going on in my head, and what triggers it, but I'm otherwise content just to let it be what it will be. What I'm not content to do is allow somebody else who has no certainty over the cause, to assert that they do.
Click to expand...
As fantasies go, this seems a little tame. Music is indeed wonderful. Ethernet switches are not.
 
tuga said:
This is about the noise travelling along the copper and reaching the D/A chip and clock, not about the data or interference.
Click to expand...
Noise from a connected Ethernet cable will be mainly due to data with imperfect driver and cable balance.
Checking some datasheets, the Ethernet transformers have leakage capacitance of < 40 pF. At that value, audio and just above noise frequencies won't get through. Direct coupling between nearby cables is probably greater
 
davidsrsb said:
At that value, audio and just above noise frequencies won't get through.
Click to expand...

I was under the impression that the noise being discussed is not audio or just above and rather is well above audio frequencies, possibly into the GHz region. Would this therefore pass through the Ethernet transformers?
 
Whatever gave you that impression? I've never seen any reference to a specific frequency band for this mythical alledged noise? What even is the source of this noise.
 
Can you answer my question please? Can high frequency noise pass through the ethernet transformers?
 
Make the noise high enough in frequency and it can pass through 2" lead sheet. You'll have to be a little more specific, you know.

Let's define the source of this 'noise' and the frequency before making some blanket statement around the effectiveness of rfi reflection. That's how specifications work after all.
 
Ok, just as a starter can the transformers pass through say 2.4GHz or 5GHz? Or if you prefer you can state what frequencies they will not pass though. Just trying to get answers or information from you rather than bluster.
 
Normally EMC/EMI is tested as conducted from 150 kHz to 30 MHz
From 30 MHz to 6 GHz is measured as a radiated signal. At these frequencies RF can jump around any barrier component
 
Fourlegs said:
Ok, just as a starter can the transformers pass through say 2.4GHz or 5GHz? Or if you prefer you can state what frequencies they will not pass though. Just trying to get answers or information from you rather than bluster.
Click to expand...
I doubt the issue will ever get settled but I will comment.

Ethernet magnetics are specified to pass differential signals (wanted signals) up to 50 MHz or 125 MHz I think. I am not aware of rejection specifications as asked (one supplier's PDF catalogue is here). However, ISTM the question does not have an answer relevant to the non-technical audio enthusiast.

More broadly, Ethernet receivers have layers of EMI (Electromagnetic Interference) protection. I do not expect anyone will want to read these URLs (and they won't directly answer the question). But perhaps it might be seen that Ethernet receiver designers do know about EMI, and information about how to deal with EMI from Ethernet is widely disseminated. See here and here for example, and some specific comments about magnetics and EMI here.

But IMHO, a better question asks whether worse than normal levels of common-mode and differential-mode EMI on an Ethernet cable cause audible interference from a loudspeaker or potentially audible levels of electrical noise on a DAC output. If they don't then IMHO one can relax.

That question encompasses a more complete part of the chain the signal passes through. It encompasses the EMI rejection you will get at each step in the chain (and side paths) - not just the Ethernet magnetics. Asking about one step in a many-step chain is interesting but the answer won't mean much without asking it about each step/side-path and adding up the results.

I have actually, by listening tests I outlined earlier, checked the whole digital-to-audio chain in my system. I admit I don't know if the network's EMI environment here is benign or bad, but I am satisfied that with a very simple network here nothing unwanted is audible even with the volume turned all the way up. I know this can be done so I am relaxed about the issue.

For some, experiments like adding network switches are part of enjoying the hobby of course. But for others, ISTM that many worries expressed about audio kit are based on the fear that its designers have not done a good job, so the user has to finish the design by buying something extra. Designers do make mistakes occasionally, but my overwhelming experience of interacting with many is of competence.
 
Lolz, better not be putting your dac next to your kitchen microwave...

As a man who sells cables that claim to attenuate rfi I'm surprised you don't know exactly what frequency bands dacs are susceptible to?

Weird.
 
John Phillips said:
<snip>

But IMHO, a better question asks whether worse than normal levels of common-mode and differential-mode EMI on an Ethernet cable cause audible interference from a loudspeaker or potentially audible levels of electrical noise on a DAC output. If they don't then IMHO one can relax.

That question encompasses a more complete part of the chain the signal passes through. It encompasses the EMI rejection you will get at each step in the chain (and side paths) - not just the Ethernet magnetics. Asking about one step in a many-step chain is interesting but the answer won't mean much without asking it about each step/side-path and adding up the results.

I have actually, by listening tests I outlined earlier, checked the whole digital-to-audio chain in my system. I admit I don't know if the network's EMI environment here is benign or bad, but I am satisfied that with a very simple network here nothing unwanted is audible even with the volume turned all the way up. I know this can be done so I am relaxed about the issue.
Click to expand...
I'm not sure of my ground here, but I think focussing only on audible noise (presumably hiss, or something, way down in the noise floor) is to miss a possible adverse effect. I've come to the view that inaudible noise (whether outside the audible spectrum, or within it but down in the noise floor at inaudible levels) may potentially be deleterious. I think of the term 'parasitic'. If it passes through the amplifier, then it gets amplified, which draws power. It might intermodulate with other noise, with unknown side effects. As I say, I'm not sure of my ground but when I listen to a system which hasn't tried to address these noise issues, some familiar factors are a greyness, a graininess, perhaps a subtle 'edginess', or a flattening of dynamics. Or just a sense of 'hash' which isn't audible as noise but perhaps as a blurring of leading edges and a bleeding of sounds into each other. And sometimes you don't even recognise these things until you've heard a system which doesn't do them.
 
Dac, low pass filter on output.

My experience of digital just doesn't mesh with yours at all. I've heard shit dacs, grossly noisy, rolled off FR, dropouts. Not for over a decade though.

The other stuff you mention, hash, edginess, grain, squashed dynamics. Nope. But then I wouldn't consider a dac that doesn't do better than-120db in thd +noise, snr, crosstalk, imd, and doesn't have linearity and FR absolutely nailed on as well as exceeding 120db for dynamic range at full output.
 
davidsrsb said:
Normally EMC/EMI is tested as conducted from 150 kHz to 30 MHz
From 30 MHz to 6 GHz is measured as a radiated signal. At these frequencies RF can jump around any barrier component
Click to expand...

Thank you. That was what I was trying to draw out. And that perhaps leads to a possibility that at the higher frequencies noise might jump around barriers such as ethernet interfaces etc. If so, and I am not saying it is, it might explain why such noise might enter and indeed pass through devices on its way to where it might cause audible distortion.

John Phillips said:
I have actually, by listening tests I outlined earlier, checked the whole digital-to-audio chain in my system. I admit I don't know if the network's EMI environment here is benign or bad, but I am satisfied that with a very simple network here nothing unwanted is audible even with the volume turned all the way up. I know this can be done so I am relaxed about the issue.
Click to expand...

I don't think any of us is saying that the noise is ever audible in that scenario, ie turning the volume up full with no music playing because it is out of audio band noise. Rather we are suggesting that it manifests itself by slightly distorting the music and so is heard by comparing the same music played with and without mitigating measures (which for the purposes of the conversation in this thread might be a switch, audiophile or otherwise but some do appear to be better than others at mitigating the noise transfer). For instance when a friend came over last week he brought his Dela S100 switch and the music sounded different with that compared to my PhoenixNET switch (same ethernet cables with both, same streamer, same dac etc).
 
Fourlegs said:
Thank you. That was what I was trying to draw out. And that perhaps leads to a possibility that at the higher frequencies noise might jump around barriers such as ethernet interfaces etc. If so, and I am not saying it is, it might explain why such noise might enter and indeed pass through devices on its way to where it might cause audible distortion.
Click to expand...
It can also jump straight past a switch near the renderer. The competently designed renderer should be immune to conducted and radiated RF at levels significantly higher than you will face at home.
 
davidsrsb said:
It can also jump straight past a switch near the renderer. The competently designed renderer should be immune to conducted and radiated RF at levels significantly higher than you will face at home.
Click to expand...

Well good luck with finding one! There is only one that has so far been mentioned to me as being immune.
 
You must log in or register to reply here.


advertisement


Back
Top