UpTone Audio EtherREGEN Switch Review

[Thomas savage]

$640 worth apparently.

I'm just grateful there's a solution , it's a doggy solution and now I think about it I'm not too sure there was a problem but still you want the transparency right ?

$640 to temporarily make myself feel safe and secure , got to be worth it.

The warmth from knowing your digital chains safe , I feel self-loved even the Mrs hating my guts dose not get me down thanks to uptone.

 

[DonH56]

There is a series of articles (threads) on jitter in the ASR Audio Reference Library sub-forum. I started to look for them to link here but really why bother...

 

[mansr]

if there is X amount of noise on the gnd. plane, then what is the amount of jitter generated at the clock?

That question is too vague to answer. Noise in the power and ground rails causes two kinds of jitter. Firstly, a fluctuating supply voltage to the oscillator can modulate its frequency slightly. With a well designed power delivery system and proper decoupling, this should be a very minor issue. Secondly, noise can shift the point at which rising clock edge crosses the detection threshold of the input circuit of the receiver. The error here can obviously not exceed the rise time, typically 2-3 ns. Realistically, the noise has to be below, say, 10% of the supply voltage or nothing would work. This puts the jitter contribution from this mechanism somewhere in the 100 ps range, at worst.

Then we have to consider the noise sources. The greatest contributions by far are from digital ICs within the device itself. To think that tiny variations in an Ethernet signal would affect this to a detectable, let alone audible, extent is beyond absurd.

 

[pozz]

There is a series of articles (threads) on jitter in the ASR Audio Reference Library sub-forum. I started to look for them to link here but really why bother...

Previously compiled into this list:

Digital Audio Articles

• (6 February 2020) FFT & Windowing Functions
• (27 March 2019) Digital Audio Cable Reflections & DACs
• (13 September 2017) DAC Fundamentals
• (13 September 2017) Sigma-Delta DACs
• (13 September 2017) Jitter Fundamentals, Part 1 & (13 September 2017) Jitter Fundamentals, Part 2
• (13 September 2017) Digital Audio Aliasing & (13 September 2017) Digital Audio Sampling
• (13 September 2017) Composition of a Square Wave
• (18 March 2017) Understanding Upsampling/Interpolation
• (3 April 2016) Understanding FPGAs
• (4 March 2016) Understanding Audio/Video Rates
• (4 March 2016) What is an Operating System? & (5 April 2016) Why Your Computer Crashes
• (14 February 2016) Understanding DSPs
• (14 February 2016) Jitter Attenuation & Clock Extraction
• (14 February 2016) Lossless Audio Compression

Full list here: https://www.audiosciencereview.com/...comers-and-inquisitive-minds.8121/post-325030

 

[jtwrace]

They are running with Wikipedia definitions and inventing the rest.....

Science Is Not A Democracy - Dr. Earl Geddes

 

[cjm2077]

That question is too vague to answer. Noise in the power and ground rails causes two kinds of jitter. Firstly, a fluctuating supply voltage to the oscillator can modulate its frequency slightly. With a well designed power delivery system and proper decoupling, this should be a very minor issue. Secondly, noise can shift the point at which rising clock edge crosses the detection threshold of the input circuit of the receiver. The error here can obviously not exceed the rise time, typically 2-3 ns. Realistically, the noise has to be below, say, 10% of the supply voltage or nothing would work. This puts the jitter contribution from this mechanism somewhere in the 100 ps range, at worst.

Then we have to consider the noise sources. The greatest contributions by far are from digital ICs within the device itself. To think that tiny variations in an Ethernet signal would affect this to a detectable, let alone audible, extent is beyond absurd.

I think the main issue comes from noise coupling from the incoming ground onto the DAC ground. And like you said, the grounds should be well isolated, filtered, and decoupled in a good design. This thing shouldn't make a difference, I agree. But we have seen some "broken" implementations in the past, like that Schiit DAC that Amir tested that coupled tons of noise from the USB. This is like snake oil cables. If they actually make an improvement to the performance of a product, that product is poorly designed or defective, and you should throw it out and get something decent. Spending huge amounts of money to fix a defective product is just throwing good money after bad.

 

[KarVi71]

Today I was listening to some music I decided to open up the webinterface of my Rockstor NAS server.

While playing a quiet song from the NAS i monitored the network activity:

What you see is three small(er) spikes at the start of the song, and then two more (top of the chart is 4.3Mb/s). The song is 2m25s long. Each sample / spike is 1 second long.

At the end of the chart you see the start of the next (more uptempo) tune.

The whole chart is ~ 3 mins long.

These are FLAC files. The song played is 11Mb.

It seems my C658 during the 2m25s long song loads data to its buffer 5 times. The active time for loading is 5 seconds (max) of the entire tune. The rest of the time music is played from the C658's own buffer (RAM).

I highly doubt that Etherregen does any good in this kind of scenario

 

[Thomas savage]

I hear uptone claim to have a vaccine for coronavirus..

 

[thefsb]

I highly doubt that Etherregen does any good in this kind of scenario

In what scenario does it do good?

 

[KarVi71]

In what scenario does it do good?

No scenario (IMHO).

 

[Superdad]

I hear uptone claim to have a vaccine for coronavirus..

You are disgusting and Amir should fire you as "moderator." People are dying and all you can think of is to make up sick bullshit barbs? Crawl back in your hole.

 

[mansr]

You are disgusting and Amir should fire you as "moderator." People are dying and all you can think of is to make up sick bullshit barbs? Crawl back in your hole.

Are you OK? It took you more than an hour to react.

 

[SIY]

You are disgusting and Amir should fire you as "moderator." People are dying and all you can think of is to make up sick bullshit barbs? Crawl back in your hole.

This fellow seems very nice.

Is the disease getting in the way of you fleecing the suckers?

 

[Veri]

I hear uptone claim to have a vaccine for coronavirus..

You are disgusting and Amir should fire you as "moderator." People are dying and all you can think of is to make up sick bullshit barbs? Crawl back in your hole.

This fellow seems very nice.
Is the disease getting in the way of you fleecing the suckers?

None of this is very nice. (try to) stay polite, guys. I'm not one to laugh with a crisis personally.

 

[Thomas savage]

You are disgusting and Amir should fire you as "moderator." People are dying and all you can think of is to make up sick bullshit barbs? Crawl back in your hole.

I'm sorry if it was distasteful, a moment of poor judgement perhaps , maybe a half second. The rest of my day I spent earning a honest living ( as iv done since I was 17) .

I guess we all have moments, I'd not get too self-righteous if I were you. You make a living from a business that by my measure is of dubious merit at best and imo ( and plenty of others ) takes advantage of people and rips them off , that takes a lot more that just a momentary laps of judgement.

As for getting ' sacked ' I'm all for new experiences, I'm just about to start a 12 week enforced holiday from work with no income support and a few dependants .

Still a sense of humour helps in dark times .

 

[Jinjuku]

I think we are all guilty of poor taste and I agree with Alex in this instance. That was a good bit below the belt.

 

[Jinjuku]

Today I was listening to some music I decided to open up the webinterface of my Rockstor NAS server.

While playing a quiet song from the NAS i monitored the network activity:

View attachment 55531

What you see is three small(er) spikes at the start of the song, and then two more (top of the chart is 4.3Mb/s). The song is 2m25s long. Each sample / spike is 1 second long.

At the end of the chart you see the start of the next (more uptempo) tune.

The whole chart is ~ 3 mins long.

These are FLAC files. The song played is 11Mb.

It seems my C658 during the 2m25s long song loads data to its buffer 5 times. The active time for loading is 5 seconds (max) of the entire tune. The rest of the time music is played from the C658's own buffer (RAM).

I highly doubt that Etherregen does any good in this kind of scenario

Yep, this is spot on and the point I'm making. Now with flexible software like JRiver and Audnirvana (so I've been told) you can change the caching behavior to all front loaded. Add to that that I'm now using a mere $210 of switch, 10GBe PCI-e NIC's, SFP+ Fiber modules, Multi-mode LC/LC cabling) I get the benefits of 49 feet of electrical isolation (MM OM3 is good for 900+ feet) and at 332MB/s I cache an entire FLAC 16/44.1 track in fractions of a second.

In your case since you are using copper (?) Alex is going to make a point that at the very least you are still experiencing leakage currents that are messing with the DAC analog out. A way to test this is to play music from local storage and see if anyone can tell when the CAT5/6 cable is plugged or unplugged since the phenomena isn't tied to actual playback but simply any traffic on the wire.

 

[amirm]

Here is the funny thing about EtherRegen. When you use Ethernet (or Wifi) streaming, as shown data is read in big chunks and then played over time. The code that does this executed by a CPU running an operating system. This means the very act of playing any streaming content, will cause fluctuations on your power and ground lines, EMI, etc. These fluctuations are far, far larger than anything a switch can do as they are local to the DAC or streamer driving the DAC.

To use an analogy, the EtherRegen is like waxing your tires so that you get a better ride when you drive on dirt road! It is a completely useless optimization.

The proper solution of course is to isolate the DAC so the streaming subsystem can't impact it. And that is what we see in well designed streamers and DACs. Once this is tackled, then there is no sense in worrying about much lower order disturbances upstream form the streamer.

 

[Jinjuku]

Here is the funny thing about EtherRegen. When you use Ethernet (or Wifi) streaming, as shown data is read in big chunks and then played over time. The code that does this executed by a CPU running an operating system. This means the very act of playing any streaming content, will cause fluctuations on your power and ground lines, EMI, etc. These fluctuations are far, far larger than anything a switch can do as they are local to the DAC or streamer driving the DAC.

To use an analogy, the EtherRegen is like waxing your tires so that you get a better ride when you drive on dirt road! It is a completely useless optimization.

The proper solution of course is to isolate the DAC so the streaming subsystem can't impact it. And that is what we see in well designed streamers and DACs. Once this is tackled, then there is no sense in worrying about much lower order disturbances upstream form the streamer.

Correct. How the EtherRegen can solve the Ethernet PHY power variance through all the other interconnecting circuitry and subsystems just never made sense to me. There are so many other interrupt driven events going on. It's mind boggling.

I keep asking for test setup that can be independently created so I can reproduce this effect.

 

[Wes]

"test setup" - that's a joke, right??

If you can get them to test their units, I'll give you a free set of million dollar cables!