Ethernet switch box

An underlying question does remain though; if we aren’t able to measure what an Ethernet switch does, how are they being designed (much less rated or evaluated for sonics)? Or is it only the audio effects which can’t be measured? Or did I not understand devg and pumacat’s discussion/disagreement about this earlier? And why couldn’t the switch function be built into the server or streamer?
 
rbbert said:
An underlying question does remain though; if we aren’t able to measure what an Ethernet switch does, how are they being designed (much less rated or evaluated for sonics)? Or is it only the audio effects which can’t be measured? Or did I not understand devg and pumacat’s discussion/disagreement about this earlier? And why couldn’t the switch function be built into the server or streamer?
Click to expand...

The Aurender ACS10 has built in, the necessary Ethernet filtering. The main Aurender (W20SE in my case) then plugs into the ACS10’s filtered Ethernet port.


Sent from my iPhone using Tapatalk Pro
 
rbbert said:
An underlying question does remain though; if we aren’t able to measure what an Ethernet switch does, how are they being designed (much less rated or evaluated for sonics)? Or is it only the audio effects which can’t be measured? Or did I not understand devg and pumacat’s discussion/disagreement about this earlier? And why couldn’t the switch function be built into the server or streamer?
Click to expand...

Companies who design, manufacture, and test switches can certainly measure what an Ethernet switch does. Things get dicey when utility products that have been used for decades by commercial companies get the audiophile makeover. Things start getting gold plated and products that used to be dirt cheap before they received the audiophile makeover now suddenly become as expensive as source components. Then marketing kicks in with a slew of buzz words to describe all of the improvements that have been made to standard utility products that have proven their ability to work in commercial environments under constant use for many years. Switches are one example and Ethernet cables are another example.

You can purchase a standard 3 meter Ethernet cable for $8.04 or you can purchase an Audioquest 3 meter Diamond Ethernet cable for $2899.95.
 
rbbert said:
An underlying question does remain though; if we aren’t able to measure what an Ethernet switch does, how are they being designed (much less rated or evaluated for sonics)? Or is it only the audio effects which can’t be measured? Or did I not understand devg and pumacat’s discussion/disagreement about this earlier? And why couldn’t the switch function be built into the server or streamer?
Click to expand...

We are able to measure the key functionality for a good-sounding Ethernet switch. As for how they are being designed, how the EtherREGN is designed is fully described at UpTone Audio's web page for ER. Have you read it?

As for putting RJ45 connectors, PHYs and flip-flops into a streamer, you could it, but it has to be done correctly. These other manufacturers need to be able to measure high-source leakage impedance current, which is not easy or straightforward. Even the design of the isolation transformer cores is important. Are these manufacturers using grounded center taps for the transformer cores to prevent AC leakage jumping from port to port if here is more than one port? None that I know of.

More to the point, the circuitry between ports on the ‘A’ side decreases phase-noise effects to some degree, but not nearly as much as crossing the ‘A’>’B’ moat. It is the differential isolators used in conjunction with the differential flip-flops that delivers the performance of the EtherREGEN. The differential isolators prevent the data-borne clock signature from getting onto the ground-plane of the PCB, while the differential flip-flops prevent the signature from getting into the flip-flop’s own internal ground network. Are other switch manufacturers doing this? None that I know of.

These other manufacturers need to be able to measure high-source leakage impedance current, and create two isolation zones, as I've described abovet to obtain the benefits ER provides for audio streaming applications; this is not easy or straightforward to do.

I've referenced and linked the white paper by John Swenson not once, but twice. It is worthwhile reading it.
 
Puma Cat said:
... I've referenced and linked the white paper by John Swenson not once, but twice. It is worthwhile reading it.
Click to expand...

White papers are opinion, not fact or evidence. This is at least the second time you have mischaracterized it. I'm not saying you don't have good points, or that you might (or might not) be "right", but it's not appropriate to present promotional material and opinion as fact or evidence.

So far, everything here suggesting that "audiophile" ethernet switches provide an audible advantage over the filtering in a streamer appears to be all subjective, like so much in this hobby. And also like so much in this hobby, subjective opinions can and will differ.
 
rbbert said:
White papers are opinion, not fact or evidence. This is at least the second time you have mischaracterized it. I'm not saying you don't have good points, or that you might (or might not) be "right", but it's not appropriate to present promotional material and opinion as fact or evidence.

So far, everything here suggesting that "audiophile" ethernet switches provide an audible advantage over the filtering in a streamer appears to be all subjective, like so much in this hobby. And also like so much in this hobby, subjective opinions can and will differ.
Click to expand...

Have you read it?

If you had, you'd see that the white paper is an engineering paper based on data, and is therefore not opinion nor promotional material. Thus, it was not being mischaracterized. It was linked, but not posted, and therefore not presented.

And...its not subjective. Just as f=ma or pv=nRT are not subjective.
 
mep said:
You can purchase a standard 3 meter Ethernet cable for $8.04 or you can purchase an Audioquest 3 meter Diamond Ethernet cable for $2899.95.
Click to expand...

@Mike, what do you think ? You have all kinds of high/low profile stuff in your arsenal. Does the Audioquest sounds the same as the one from Amazon ?

In my system, the ones from Amazon actually transmits and receives bit-by-bit data accurately with appropriate latency and there is no TCP re-transmission or timeouts or dropouts, neither there are errors in the PHY or MAC statistics on the receiving end. I even tried a full-duplex line rate at 1.44M packets/sec from an Ixia port and they all work out Ok, though there are some drops in the receiver end which is to be expected at these high data rates and the cables aren't the problem. Having said that, for some reason or the other the $8.04 ethernet cable sounds like a $8.04 ethernet cable, until I tried 4 different ones - JCAT, Sablon, Wireworld and SoTM. This is why I moved to fiber which is dirt cheap as the $8.04 ones and saved a ton of money. If someone is interested in JCAT or Sablon, please PM me (I sold the others).
 
Devg said:
@Mike, what do you think ? You have all kinds of high/low profile stuff in your arsenal. Does the Audioquest sounds the same as the one from Amazon ?

In my system, the ones from Amazon actually transmits and receives bit-by-bit data accurately with appropriate latency and there is no TCP re-transmission or timeouts or dropouts, neither there are errors in the PHY or MAC statistics on the receiving end. I even tried a full-duplex line rate at 1.44M packets/sec from an Ixia port and they all work out Ok, though there are some drops in the receiver end which is to be expected at these high data rates and the cables aren't the problem. Having said that, for some reason or the other the $8.04 ethernet cable sounds like a $8.04 ethernet cable, until I tried 4 different ones - JCAT, Sablon, Wireworld and SoTM. This is why I moved to fiber which is dirt cheap as the $8.04 ones and saved a ton of money. If someone is interested in JCAT or Sablon, please PM me (I sold the others).
Click to expand...

Me? No. And the Shunyata Alpha, Sigma and Omega are my faves.


Sent from my iPhone using Tapatalk Pro
 
mep said:
Things get dicey when utility products that have been used for decades by commercial companies get the audiophile makeover. Things start getting gold plated and products that used to be dirt cheap before they received the audiophile makeover now suddenly become as expensive as source components. Then marketing kicks in with a slew of buzz words to describe all of the improvements that have been made to standard utility products that have proven their ability to work in commercial environments under constant use for many years.
Click to expand...

I despise this as well. There is no reasons to have audiophile makeover for off-the shelf utilitarian items based off great IEEE standards with tight SI and leakages. Today you can get 10-port with 2x10G ethernet switch for $200 but we have to pay $640 for a audiophile 6-port switch which is even limited to 100Mbps on one port. Today's technology can do 400G per port and we can have many in a single unit. I for one would like to save my money if there is a way and this has been my quest as well.
 
rbbert said:
An underlying question does remain though; if we aren’t able to measure what an Ethernet switch does, how are they being designed (much less rated or evaluated for sonics)?
Click to expand...

Good question. You can send an email to the designer/manufacturer and see what they have to say. I don't think anybody here can answer to your satisfaction since none of here has designed them. I would be curios to know as well.

rbbert said:
And why couldn’t the switch function be built into the server or streamer?
Click to expand...

Many reasons but honestly I won't go over it since its extremely difficult to convey anything in this thread :)
 
If you couldn’t measure what an Ethernet switch does, then they couldn’t have been designed, manufactured, and tested so let’s stop saying you can’t measure what an Ethernet switch does.
 
rbbert said:
And why couldn’t the switch function be built into the server or streamer?
Click to expand...
actually multiple products have some kind of filtering inside.
for example the melco servers have even outputs, so they can be used as a switch.

a big part of the improvement comes from putting 2 switches or more in serial. this is for galvanic decoupling as well.
you want your own switch for each unit in best case scenario, however, even if you have a built in unit, you still will benefit from more switches and theyr superior quality.
 
Puma Cat said:
We are able to measure the key functionality for a good-sounding Ethernet switch. As for how they are being designed, how the EtherREGN is designed is fully described at UpTone Audio's web page for ER. Have you read it?

As for putting RJ45 connectors, PHYs and flip-flops into a streamer, you could it, but it has to be done correctly. These other manufacturers need to be able to measure high-source leakage impedance current, which is not easy or straightforward. Even the design of the isolation transformer cores is important. Are these manufacturers using grounded center taps for the transformer cores to prevent AC leakage jumping from port to port if here is more than one port? None that I know of.

More to the point, the circuitry between ports on the ‘A’ side decreases phase-noise effects to some degree, but not nearly as much as crossing the ‘A’>’B’ moat. It is the differential isolators used in conjunction with the differential flip-flops that delivers the performance of the EtherREGEN. The differential isolators prevent the data-borne clock signature from getting onto the ground-plane of the PCB, while the differential flip-flops prevent the signature from getting into the flip-flop’s own internal ground network. Are other switch manufacturers doing this? None that I know of.

These other manufacturers need to be able to measure high-source leakage impedance current, and create two isolation zones, as I've described abovet to obtain the benefits ER provides for audio streaming applications; this is not easy or straightforward to do.

I've referenced and linked the white paper by John Swenson not once, but twice. It is worthwhile reading it.
Click to expand...

interesting!
i d love to see you winning experiance with other manufactors switches as well.
 
mep said:
I do think there could be sonic improvements to be had by replacing a SMPS with a linear power supply, but I’m not sure all digital devices will benefit equally. In my setup, I would like to hear the difference with a linear power supply powering my Roon Nucleus+ vice the wall wart it comes with.
Click to expand...

Where you not scoffing at me some post back. How dare I insinuate your system leaves performance on the table. And here you are saying gains could be had. So which is it.
 
Kingrex said:
Where you not scoffing at me some post back. How dare I insinuate your system leaves performance on the table. And here you are saying gains could be had. So which is it.
Click to expand...

Talking about linear power supplies in this thread was an offshoot that started after I mentioned digital dingleberries. This occurred after you decided to butt in with comments about my digital rig. Your comments about the backbone of my digital rig needing to see a digital chiropractor had nothing to do with linear power supplies.
 
mep said:
Talking about linear power supplies in this thread was an offshoot that started after I mentioned digital dingleberries. This occurred after you decided to butt in with comments about my digital rig. Your comments about the backbone of my digital rig needing to see a digital chiropractor had nothing to do with linear power supplies.
Click to expand...
It has everything to do with it. Your telling people an off the shelf switch fed with cat 8 terminated on premise (I assume) is just as good as a EtherREGEN with say and AQ cinnamon cable. I believe that is hogwash. So I say your leaving performance on the table. You then imply I have no knowledge of how to make it better. Well I do, as do others who have explained all the small parts and pieces that make up a significantly better whole. Your digital system, especially the steaming portion is not optimized. An optimized steaming system will come close to the performance of stored on rhe hard drive playback. On the hard drive will beat it. But it gets good enough that you stop worrying about where the media came from. Hard drive, NAS or internet. It all sounds darn good. Yes it cost more money. Maybe $2k to $3k to do an entry to mid level package correct. After that you could spend another $3k, but that is going to bring a smaller gain than the first outlay. And for how much money people are spending on DAC and servers. Many ponying up $20k to $30k or more, its foolish to not invest the $3k to set the digital backbone right.
 
You must log in or register to reply here.
Back
Top