EtherREGEN on the way!

[Puma Cat]

Originally Posted by LarsT

Well, I'm convinced. I ordered one last night.

Cool, I think you're going to really like it. Please post your impressions once you get it and get 72 hours or so of burn-in on it.

[Puma Cat]

Originally Posted by Antonmb

Me too, ordered today. Thanks Stephen for all the great info.

You're welcome, Tony. Think you're really gonna like it.

Please post your impressions once you get yours and get it settled, burned-in and "up to speed".

Cheers,
Stephen

[Antonmb]

Originally Posted by Puma Cat

Please post your impressions once you get yours and get it settled, burned-in and "up to speed".



Cheers,

Stephen

Looks like that won’t be until late January.

[Puma Cat]

Originally Posted by Antonmb

Looks like that won’t be until late January.

Yep, sorry to hear that, Tony. The first three runs sold out very quickly. The first run for November sold out in 6 minutes after the product went live for sale, the second allotment for December sold out in less than 50 minutes after going live, and that pushed any subsequent orders later that day, or any day after Oct. 8, into January.

[Antonmb]

Originally Posted by Puma Cat

Yep, sorry to hear that, Tony.

It’s ok, I knew it when I ordered - I have enough to keep keep me busy for a couple of months[emoji106]

[Cohibaman]

I’m using a dedicated, purpose-built pc running Windows 10 and JRiver, fed by my NAS.

NAS > router > switch > PC > Schiit Gungnir (Gen5 USB), all via CAT 6 cable to PC.

How would this help and if it would, why? Not a challenge whatsoever; I want to learn and understand.

[Puma Cat]

Originally Posted by Cohibaman

I’m using a dedicated, purpose-built pc running Windows 10 and JRiver, fed by my NAS.

NAS > router > switch > PC > Schiit Gundnir (Gen5 USB), all via CAT 6 cable to PC.

How would this help and if it would, why? Not a challenge whatsoever; I want to learn and understand.

Hey Mike,
Sure, here is an explanation of why using EtherREGEN in place of your switch would help the audio quality of your digital streaming front-end:

There are usually a number of problems going direct from consumer-grade routers:

1) the clocks are in generic consumer-grade routers are very cheap and induce clock phase noise, which is audible.

2) the flip-flops for these Ethernet switches are also very cheap, and contribute to the noise carried on top of the analog square waves that comprise digital bitstreams.

3) Most routers are powered by switch-mode power supplies, which are insidious with respect to producing low- and high-impedance AC leakage currents, which contribute significantly to clock phase noise and jitter. The impact of high-impedance leakage currents and their role in inducing clock phase noise and jitter is not well-known as this specific class of leakage currents were not even discovered until October, 2017.

4) The ports on routers typically have low-specification isolation tranformers and leakage currents can "jump" from port to port, contaminating more than one component in the chain.

5) There is no distinction between input and output ports on routers, which means that AC leakage currents can move from an "input" device e.g. a music server to an output device, e.g. a network bridge.

The EtherREGEN is a bespoke Ethernet switch product specifically developed for audio applications that mitigates and obviates these specific problems and source of noise as listed below:

1) the clocks in EtherREGEN are ultra-low-jitter/phase-noise Crystek CCHD-575 oscillators. Importantly, the clock distribution system (for the switch chip, the special Ethernet format conversion chips, and the high-speed flip-flops is run differentially throughout.

2) EtherREGEN has two isolated data/power/clock domains (the A side and the B side; separated by a "moat"). Each isolated domain is re-clocked using 10GHz-capable ultra-low-jitter (less than 0.8 picoseconds) differential flip-flops

3) The power supply for the EtherREGEN is specifically designed to prevent the passage of leakage currennts which induce clock phase noise.

4) the RJ45 module utilizes 12 transformer cores in each port (most Ethernet ports have 2~6 cores), and ground their center-taps through capacitors in a way that blocks port-to-port AC leakage currents.

5) ER uses an Active Differential Isolation Moat (“ADIM™”) that separates the A side (typically the input side)—to its ‘B’ side preventing leakage currents from being passed on to streamers, network bridges, and ultimately, DACs.

All of these specific features, functions and design embodiments are why ER provides notably superior audio quality for digital streaming than a generic, consumer-grade router.

Analogously, its no different than using a Shunyata Denali 6000/T V2 for distributing power vs. an el cheapo Monster power strip.

[IM3CPO]

Let me first state that I am not intending to pour water on anyone who claims this type of product makes a difference in their audio experience. If you are happy with whatever gains you think it makes, then great! Having said this, one should keep in mind that no audio signal is being transmitted via the ethernet cable. There are several protocols involved which encapsulate the data. There are even protocols involved which ensure no data is lost in transmission. That is why your music will "pause" when you do in fact lose your signal rather than skip the amount of time the signal was actually down. TCP-IP is even intentionally designed to *NOT* send the data in order. So your device may not even receive the packets in the order they were sent. The claims about jitter, timing, etc., do not make any sense.

The only situation I could see something like this potentially making a difference is if the voltage on the ethernet cable somehow interferes with the functionality of the chips doing the work on the devices through bad shielding/EMI/grounding. For example, if the switch you use adds some extra distortion on the wire (not in the packets, or not bad/consistant enough for the error correction not to work) that somehow interacts with the device processors receiving the transmission due to bad shielding/EMI/grounding which in turn causes something on the device to run out of spec and adds distortion to the output, and this switch does something to fix this, then I could see a gain. Anything other than this simply makes no sense.

[Puma Cat]

Originally Posted by IM3CPO

Let me first state that I am not intending to pour water on anyone who claims this type of product makes a difference in their audio experience. If you are happy with whatever gains you think it makes, then great! Having said this, one should keep in mind that no audio signal is being transmitted via the ethernet cable. There are several protocols involved which encapsulate the data. There are even protocols involved which ensure no data is lost in transmission. That is why your music will "pause" when you do in fact lose your signal rather than skip the amount of time the signal was actually down. TCP-IP is even intentionally designed to *NOT* send the data in order. So your device may not even receive the packets in the order they were sent. The claims about jitter, timing, etc., do not make any sense.

The only situation I could see something like this potentially making a difference is if the voltage on the ethernet cable somehow interferes with the functionality of the chips doing the work on the devices through bad shielding/EMI/grounding. For example, if the switch you use adds some extra distortion on the wire (not in the packets, or not bad/consistant enough for the error correction not to work) that somehow interacts with the device processors receiving the transmission due to bad shielding/EMI/grounding which in turn causes something on the device to run out of spec and adds distortion to the output, and this switch does something to fix this, then I could see a gain. Anything other than this simply makes no sense.

John Swenson, who worked as an Ethernet product R&D scientist for over 30 years for Broadcom and Cisco, has studied this in-depth with respect to audio applications. The reason that these products, e.g the EtherREGEN, SOtM, AQVOX, etc., Ethernet switches and clean linear power supplies provide an audible improvement is due to their suppression and amelioration of "noise components", that, for the lack of a better characterization are "overlaid" onto the analog square wave voltages that comprise digital bitstreams. AC leakage currents, particularly high-impedance leakage currents, and which were only discovered by Swenson in Q4, 2017, impact this significantly, and one of the key impacts they have is to demonstrably increase clock phase noise.

Its against forum rules to post links to other audio forums, so I can't link to it here, but here is a short description by Swenson, discussing the impact of these noise factors that was discussed in the technical forum regarding the 18-month development of EtherREGEN:

"There are two types of "SQ degrading" influences the EtherRegen is designed to radically decrease: leakage, both high impedance and low impedance, and clock phase noise. The clock phase noise travels on the Ethernet signal itself (every edge coming out of any digital device caries the phase noise of the clock used to "clock out" that edge)."

If you look any SOTA digital component, you will note that their engineers do as much as they can to reduce clock phase noise. I'm sure if you spoke to Mike Story of dCS or Xuanqian Wang of Auralic, they could describe in detail the impact of clock phase noise on audio reproduction.

When femtosecond timing differences are audible, increased clock phase noise from crap switches, SMPS, and high-impedance leakage currents all have an audible impact on digital bitstreams. Swenson is working on a white paper describing this in-depth, & he is working on bespoke metrology to provide measurements that demonstrate this as well. Once soon as that white paper is available, I will reference it here.

[IM3CPO]

Originally Posted by Puma Cat

If you look any SOTA digital component, you will note that their engineers do as much as they can to reduce clock phase noise. I'm sure if you spoke to Mike Story of dCS or Xuanqian Wang of Auralic, they could describe in detail the impact of clock phase noise on audio reproduction.

When femtosecond timing differences are audible, increased clock phase noise from crap switches, SMPS, and high-impedance leakage currents all have an audible impact on digital bitstreams. Swenson is working on a white paper describing this in-depth, & he is working on bespoke metrology to provide measurements that demonstrate this as well. Once soon as that white paper is available, I will reference it here.

The phase clock you are referencing only applies to Synchronous Ethernet. Synchronous Ethernet is not used on computer networks; its only used on cellular networks, VoIP phones, IPTV. Ethernet computer networks do not carry any clock synchronization information whatsoever.

If you have a DAC that requires an external clock generator via the ethernet cable, then you are injecting it yourself. No home router, etc., will support doing this..