The EtherREGEN thread for various network, cable, power experiences and experiments

[skatbelt]

On 7/3/2020 at 10:21 PM, sandervdp said:

After enjoying the EtherRegen for more than 6 months now, have decided to go ahead and order a second one

The EtherRegen has so dramatically increased the enjoyment I get out of my set music, that it's worth seeing what another one upstream might do

 

See below for my system setup. The EtherRegen will replace the Netgear in this setup. Any ideas on how the moat should be crossed? Cable modem on the b-side, everything else (oppo, mac mini, apple tv, downstream etherregen) on the a-side? That should isolate it from the rest of the network most, right? And the second etherregen will create another moat to isolate the streamer from the server/etc. Is that correct?

And is it worth experimenting with fiber in between the etherregens? 

Also happy to hear any other suggestions for experiments

 

My suggestion: keep it 'as is' and go fiber to the etherREGEN by placing a Sonore opticalModule between the Netgear and the EtherREGEN. You could power the oM with the JS-2 or go with a separate LPS. So, no 2nd EtherREGEN!

[skatbelt]

Normally an early adopter but due to busy times at work now probably part of the late majority; I just wanted to mention that going from copper to fiber has brought me quite a substantial step up in sound quality. Thanks for al the advice in this forum. I now have:

 

ISP modem/(wifi)router -> Telegärtner CAT6 U/UTP (0,5m) -> Sonore OpticalModule fed by Uptone LPS1.2 -> Planet MGB-TLX SFP -> Telegärtner 9/125 µ Singlemode OS2 (10m) -> Planet MGB-TLX SFP -> EtherREGEN

[skatbelt]

My setup: ISP router ->opticalModule -> EtherREGEN ->streamer/bridge -> DAC

 

I now have my Roon Server connected to EtherREGEN A-side. Has anyone tested a comparable setup vs. placing the Roon Server (mac mini, nucleus etc.) upstream the opticalModule. So, a very electrically clean use of the EtherREGEN with only fiber in on the A-side but as a sacrifice the server in the more noisy environment of a generic router or switch. What sounded best?

 

I know it is best to experiment with this myself but a fair A/B would be quite a hassle because of a joint LPS use in my setup at the moment.

[skatbelt]

25 minutes ago, HumanMedia said:

I have received a 3m cable and single mode SFP transceivers to optically connect a Sonore opticalModule to the etherRegen.

The SFP are the single mode PlanetTech (20KM range), no attenuators. All devices have quality linear power supplies.

 

All functions fine and I have had it running for a week now.

 

Only issue is that this optical connection sounds worse than a CAT6a ethernet connection from the same router. Slightly pushed, unnatural high frequencies. I've swapped back and forth a few times and optical in my system is consistently worse.

 

I know others have tried this exact setup without any issues, with a noticeable improvement over wired ethernet.

 

Has anyone else experienced this? or have suggestion to what is wrong and how this could be fixed?

 

The only thing I can think of is that I need to add attenuators to the cable, maybe 5db to 10db. however multiple others have found that this is not an issue with the PlanetTechs.

 

I am at my wits end! Help me brains-trust!

 

I have 10m single mode combined with (I believe) the same Planet SFP's without attenuators and absolutely prefer this to copper.  I cannot imagine that 3m vs. 10m makes a difference. Which optical cable do you use? I use this (OS2) one: https://www.conrad.nl/p/telegartner-l00875a0034-glasvezel-aansluitkabel-1x-lc-stekker-1x-lc-stekker-9125-singlemode-os2-1000-m-1542091

 

Edit: the opticalModule needs some time to burn-in so this could also be a reason.

[skatbelt]

1 hour ago, Heckyman said:

I didn't try your exact set up, but I did have a try with fibre and didn't like it. 

The fibre was between Edgerouter-x SFP and Etherregen (Ubiquiti SFPs).

Perhaps the background was "blacker" and my initial impression was quite good. But over a period of days I noted less engagement with the music. There was a kind of mechanical quality whereas returning to copper felt more natural and alive even if there might have been a touch more "noise". 

Normally I have a small go with something and if the improvement is in the right direction I optimise further. Regarding fibre networking, I decided not to pursue it further as it struck me as something that cuts both ways (i.e benefits of opto-isolation vs drawbacks of extra conversion/processing). 

 

In summary, trust your own ears in your own system!

 

My experience is quite the opposite but only after adding a Sonore opticalModule. My guess is that your Edgerouter is a weak link when going fiber for audio.

[skatbelt]

43 minutes ago, HumanMedia said:

 

I think I remember reading in one of the OpticalRendu threads, that John said something about implementing improvements to the way SFP noise was handled (in the OpticalRendu). Its a vague memory and it might have even been before the etherRegen was released so it may already be in the etherRegen.

 

In rereading some of the older posts about optical connections there were quite a few people saying that the sound was a little harsh until they added optical attenuators. However people also reported that the PlanetTech SFP didn't need them.

 

I have been talking to an optical networking guy, who does solely optical networking for a living and he had this to say. He said that these days to make things simpler parts wise they only do single mode installations even for short runs, and use attenuators.  He uses a meter to take readings then calculates which value attenuator to use. For 20km strength and greater they always use attenuators. In the absence of a signal strength reading he can also do calculations from the spec sheet to give a really rough approximate of what attenuator to use. For the PlanetTech he said that you probably wouldnt get into a situation where you could burn the receivers out but still recommended a 5db attenuator as a starting point.

 

Maybe this is it? I know someone - was it you kennyb123?, mentioned that attenuators on the PlanetTech dulled the sound. Do you still feel this way? have you revisited attenuators? And could dulling actually be the removal of artificial 'sharpness'?

 

 

 

 

 

Not @kennyb123 but I am happy to experiment with attenuators on the PlanetTech's. Just out of curiosity about the results.

 

To throw in a few other perspectives: when going from copper to fiber with opticalModule + PlanetTech's in my system and to my ears SQ became less harsh and edgy and more natural. To a marginal extent of course. It was the first time I felt no need for the HF filter active on my Chord DAVE and the first time that I couldn't distinguish between streaming audio and locally stored files within Roon. My Roon server + DAS is directly connected to the EtherREGEN a-side, so by-passing the fiber path for playback.

[skatbelt]

57 minutes ago, GMG said:

Hope it's OK to mix in some JCat here

 

JCat has a new network card coming out next month

https://jcat.eu/product/net-card-xe-high-end-hifi-network-card/

 

Has anyone given some thought to weather such a card would make a difference when using an ER?

Computer-->JCAT new card-->ER A side--> ER B side--> network player

 

 

 

My guess is that it will potentially eliminate noise coming from your computer (and the network card itself), despite of the isolation measures already taken on the JCAT card (transformers and external PSU). Assuming you can meet the moat related setup requirements of the ER and (preferably) power it with a high quality power supply.

[skatbelt]

11 hours ago, Rsbrsvp said:

I have a mutec Ref10 SE-120 connected to my Etherregen and a superb Power supply.  It does help a lot.

 

Still, playing downloaded music from my SSD is better than streaming direct.

 

Why?  And is that download quality being effected by the switch, PS, and Mutec clock?

 

To bring even more variety to the discussion. For me, adding a Ref-10 as an external clock to the etherREGEN did not improve things. This surprised me because of the positive comments here of adding an external clock. I suspect this has to do with the up- and downstream components relative to ER in my system. Downstream: dCS Network Bridge and the Chord DAVE DAC. The dCS NB re-clocks the incoming signal again and it is also buffering. Upstream: Sonore opticalModule fed by LPS1.2. This unit also re-clocks to my knowledge.

 

Another difference from your observation: the difference between streaming and local file playback is negligible in my system since etherREGEN / opticalModule.

 

Question for others: did anyone hear improvements with a better than Ref-10 external clock in a system that already had etherREGEN / opticalModule? If so, which clock and what is downstream etherREGEN?

[skatbelt]

16 hours ago, Clockmeister said:

 

One question though, are you using the DSP in the Lyndorf for room correction?

 

Thank you for your extensive reply and suggestions. Yes, I do use Lyngdorf's roomperfect/DRC. It does a lot in my listening environment. It takes it to a really high level. However - and to rule out too much variables - the first critical assessments and comparisons are always with my headphone system.

 

I don't see the direct benefits of adding a MC3+ usb between dCS NB and DAVE. To output USB? I now use AES/EBU between dCS and DAVE and feel blessed that I have been able to drop USB as a protocol and connection type. Too brittle, too sensitive, too many dependencies in my experience.

[skatbelt]

8 hours ago, One and a half said:

Thanks @skatbelt, could you describe what you heard when connecting the REF-10.  In particular,  if the image deteriorated , distortions or just no difference worthwhile. 

 

Pretty sure the ER BNC shell is floating,  so any ground would be referenced back to the clock. If the clock is also ungrounded, there could be a problem there.

 

Hi Gary, just no difference noteworthy. To give an indication: 'rolling' SFP's made a bigger difference. I borrowed the REF-10 from a good friend and in his system it makes a positive difference. But he has the ER B-side directly connected to the ethernet input of his Lyngdorf TDAI-3400. 

 

 

[skatbelt]

8 hours ago, LowMidHigh said:

 

It takes a few good days for the Ref10 to stabilize.

 

I had it in my system for more than a week.

 

My provisional thought is that there is only so much that can be gained by cleaning up the incoming signal to the dCS Network Bridge. I can imagine as per @Clockmeister suggestions that there is further improvement possible between the dCS NB and DAVE but I don't see myself stacking two MC3+'s and a REF-10. I like to keep things simple. But I can imagine future system where one master clock references a switch, streamer and DAC.

 

If I remember correctly, 'the Uptones' also indicated that there would be little to gain with an external clock for the ER. Or did @Superdad and @JohnSwenson come back from that?

[skatbelt]

12 hours ago, Superdad said:

 

No, our stance has always just been that an upgraded clock and power supply are not required in order to enjoy the majority of the benefits of the EtherREGEN.

But producing the device with provision for those to be added allowed us to offer the basic (yet still VERY advanced!) switch at reasonable cost, and for those who wish to carry it further.

We could build a big EtherREGEN with state-of-the-art AC>DC power supply and top-end OCXO built in (plus other enhancements), but the price would then be several thousand. Not going there anytime soon... 

 

To check if my memory had failed me I read back and during the launch of the ER you literally said that it would take a very high-end master clock to gain benefits, is there were are any.

 

Have you come back from this thought? And have you actually experienced yourself that an external clock delivers significant improvements? If so, which types did you experiment with?

 

[skatbelt]

1 hour ago, Mops911 said:

I think the value was -130dB at 10Hz

 

I just wonder more generic. If lets say the eR improves sound by +1 (whatever that is), how much improvements do peripherals do?

 

- Clock? +0.2?

- good LPS, +0.25?

- cables and others, +0.1?

 

I get the diminishing returns, but if a decent clock (AfterDark Prince) with its own LPS would only add 20% of SQ but 100% in price, its likely not worth it....

 

Any thoughts?

 

Exactly what I was thinking. In my setup an EUR 3000 Mutec Ref-10 didn't improve anything. A good LPS did improve a lot. Going from copper to fiber was the icing on the cake.

 

May be the EUR 5000 Ref-10 120 or a higher line Afterdark clock will squeeze out a bit. Quite a few members are stating this. But this money is probably much, much better spent buying a better streamer, DAC etc.

 

What I find difficult to filter in all this is that Uptone and Afterdark are commercially linked. I know this is a sponsored forum but still.

[skatbelt]

2 hours ago, Superdad said:

No, the only link between UpTone and AfterDark is that AfterDark is a dealer for UpTone in Hong Kong/China region. And not even exclusive dealer there--we have two others as well.

AfterDark is creating their own products and they enjoy promoting them for use with the hugely successful EtherREGEN.  They also enjoy selling our JS-2, UltraCap LPS-1.2, and ISO REGEN.

 

Alex, seriously? I think statements like this will make even the greatest Uptone Audio adepts frown. And I count myself among your fan base. I think it is dubious that you let a dealer who has multiple interests do so in a forum where members are mainly looking for reliable information. A dealer who also makes handy use of your reputation in his sales techniques and almost sells a kind of feeling of 'Uptone-certified' for his clock products.

 

I would appreciate if you answered my earlier question: have you actually experienced yourself that an external clock delivers significant improvements? If so, which types did you experiment with?

[skatbelt]

13 hours ago, LowMidHigh said:

 

Router -> DX  cable + DX -> Unshielded generic Cat6 -> DX + DX cable -> ER side A

So the generic Cat6 is the setup above "breaks" the shield and prevents ground loops.

 

Does it make sense to assemble 2 short unshielded generic Cat6 to replace the DX cables?

 

 

 

The metal housings of the two ethernet ports in a DX are not connected internally. So no need to worry about ground loop issues.

 

I just installed dual DX's between ISP router and opticalModule. Certainly not night and day in my system. Slightly cleaner / a lower noise floor is my first perception. But also a tad leaner. I am not yet sure whether it benefits musicality. My system was already very transparent and neutral. Let's see how things evolve, giving it some more time.

 

Anyway, thanks Gary @One and a half for pointing to this cheap tweak!

[skatbelt]

29 minutes ago, agladstone said:

Does anyone know how or if these DX Filters are different from the Baaske Medical Ethernet isolators? They also use the same Halo RJ45 plugs for the in and out. They seem very similar to me, but for all I know they could be doing something completely different and just physically look similar? 
I’ll have to go back a hundred responses or so to see what @JohnSwensongave for the explanation of what the medical Ethernet isolators do? 
I have a few of these Baaske isolators now just collecting dust since it seemed like John Swenson thought it wasn’t likely to do to much upstream the A side of the EtherRegen and I wanted to eliminate all the additional Ethernet cables and connections using them added. 
I was only using one in the middle of each connection as follows:

router -> Baaske-> eR , so maybe I should experiment using two of them like how the DX filters are being used and see if that gives me any benefit on the A side prior to buying and trying the DX filters if they’re essentially doing the same thing ? 

 

My thinking is the DX, Baaske and Emo isolators all share the same principles. The only real difference I can see is the the recommended use in pairs. I have a Emo EN-70HD between my Mac mini and the EtherREGEN A-side and this certainly makes quite a positive difference. I guess the etherREGEN benefits from the rejected noise by this isolator. It does more in my system than the DX-pair between ISP router and opticalModule I have upstream the etherREGEN since a few hours.

 

To my memory, the statements JS made were related to the use of isolators on the B-side.

[skatbelt]

3 hours ago, One and a half said:

I had a look at the Emo unit which has a high voltage isolation component, and uses the isolation properties to block noise, whereas the dx filter is based on noise filtration only. Not possible to see the data sheet on the Emo, some weblink faff, would suggest the DX units after the EMO and in front of the EtherRegen Port A, since you then have safety/isolation plus filtration. By accounts from others the DX units are effective just before the EtherRegen. 

 

If there is not much change in the SQ with the DX in this position it means the EMO is doing its job. Isolation transformers (like in the EMO) can still allow some noise to pass, so I think there will be a benefit. 

 

Yes, the B side should be clear to the Renderer/DAC/Endpoint, suspect there is a feedback mechanism which may be compromised when a filter or isolater are used. Just a guess there.

 

Find attached the datasheet of the Emo EN-70 series.

 

I thought the most logical place for the DX's is where I put them now. Please note that my (Uptone DC modded) Mac mini is only serving as a Roon server. I can easily determine it's (negative) effect in electrical terms by temporarily switching it off and comparing the sound quality with the very good player software of the dCS Network Bridge (Mosaic), which also supports Qobuz. The Emo EN-70HD brings Mac mini/Roon on par with Mosaic.

 

So if I omit the mac mini my current chain is:

 

ISP router > DX > short Telegartner CAT6 UTP > DX > OpticalModule (fed by LPS-1.2) > 10m fiber with Finisar SFP's > EtherREGEN (fed by JS-2) > Ghent Belden CAT6A (JSSG360) > dCS Network Bridge > ..

 

(DX's for the time being connected with with the supplied short cables, ISP router with factory SMPS and with 2 other (noisy) ethernet devices connected; a Synolgy NAS and a Orbi mesh system)

 

 

Datasheet_EN-70-V40-1591535.pdf

[skatbelt]

4 hours ago, MartinT said:

 

Nope. For the second time I've taken them out of circuit. Some of the life was gone from my music. Not exactly dynamics but the DX filters just make it sound flatter. Removing them has restored its sense of life and vibrancy.

 

I'm afraid these are just not working for me. They're staying out.

 

Exactly my findings. I removed them and the musicality and emotion is back. May be I am going to try them in another position later on.

 

@PYP, replacing the supplied shorts cables with short CAT6 UTP cables did make a difference, but not satisfying as well.

[skatbelt]

9 hours ago, roman410 said:

I fully agree with you. But I want add one more word"right".

 

My music enjoyment with DX filters it is similar to " late night music listening experience" when music for some reasons sound so"right".

 

 

 

At night our ears are probably extra isolated by alcohol... 😇

[skatbelt]

5 hours ago, roman410 said:

Yes, that it is one of the theories,  but actually I was experiencing this effect early morning weekend listening sessions! I know my self if I will drink night before I will have different problem (hangover)😁.

 

Some people believes the main grid is cleaner late nights. But after installing DX filters I was experiencing this effect saturday afternoon, when grid was polluted from opening all local restaurants and everybody used A/C because of heat wave here in the northeast US.

 

 

 

I know the stories about a cleaner mains grid at certain times of the day and have experienced this myself. However, since I use an isolation transformer on a dedicated power line, the difference has become negligible.

 

I would like to clarify something about my experience with the DX's. I also strive for a transparent, neutral, and yes, even naked sound, but what I experience with the DX's is not an improvement of these criteria but a kind of unrest or imbalance. Especially in micro details and in the upper frequency range. High registers of a violin become too sharp, for example. The impression that jitter has increased instead of a reduction. As an analogy: like new glasses that correct the eyes just a bit too much and the brain cannot get used to. Even after a longer listening period.

[skatbelt]

8 hours ago, Superdad said:

Okay folks, can we give the DX filter debate a rest for a bit?  I think it is clear that some people enjoy the extra transformer cores on the line while others do not.  Cheap enough for anyone to try.

 

 

I would like to go on a little longer on the subject. After all, this is the experimentation thread and it is related to how the ER works... 😉

 

Last night I did another experiment. I replaced the EMO EN-70HD between the Mac Mini and the ER (A side) with the DX pair. At this place I do hear a small improvement without negative effects. I am curious if this is due to the 'in series' of two transformers. In other words, whether two EMOs in series would have the same effect.

 

So for now my conclusions are:
- DX pair does not work positively between ISP Router and opticalModule
- DX pair works positively between Mac mini and close to ER (A-side; all short 1-foot cat6 UTP cables)
- DX pair works better than single EMO EN-70HD between Mac mini and ER (A-side)
- No transformer at all between Mac mini and ER A-side is a big step back in SQ <- @Superdad disputed this way, way back

 

Maybe I should have listened to @One and a half right away ...

[skatbelt]

46 minutes ago, roman410 said:

Long time I was rank dedicated power line  for best tweek in this hobby. That way when a bought my new condo, with luxury to have acces to the atik I have now 4 dedicated lines in my listening room. Was also experimented with isolation transformer, but found not for me,not just was buzzing so loudly also rob the dynamics. After experimenting with different brands of power conditioners and filters, I settled up with very good PI audio UberBUSS power filter.

 

Your experince with unrest and imbalance indicated to me something can get wrong with your ring longer time ago, but was masked with the noise, which when now it is reduced everything fleshed up.

 

When I talk about the noise, now after installed 4 pairs of DX filters I start hearing noise floor of the music. Noise what was embedded to the music during the recording, but that is very pleasurable noise😉.

 

 

If you follow the JS 'guidelines', isolation transformers are very, very effective. Especially in a balanced configuration. And if the capacity of the IT is correctly dimensioned, there is certainly no question of reduced dynamics. On the contrary. It is often power conditioners that limit the dynamics, especially with power amps. Mine is placed in an adjacent room, so I don't suffer from the humming.

 

For me the key to success was to have the whole system (streamer/dac/amp) behind one isolation transformer and to have all other incoming/outgoing signals electrically isolated. 'Moated' so to speak.

 

So, my system is also very revealing and has a very low noise floor. I can hear every noise and all imperfections in recordings, including tape hiss. And still it is able to bring across all the emotion and feeling that is in the music.

 

But I don't want to make a yes-no discussion of it. Every system is different and everyone's preferences are different. My DX-filters have found their place and in the end I am happy I bought them.

 

[skatbelt]

5 hours ago, R1200CL said:

Definitely NOT. 
 

Use of the EtherREGEN’s Ground terminal screw:
We have measured high-source-impedance leakage on Ethernet cables coming from various other common network gear and computers. Leakage is common-mode AC current traveling over DC connections. It is pernicious stuff—passing easily along both analog and digital connections in our audio systems—and it is desirable to shunt it away. We address leakage in the design of the EtherREGEN in several ways, including our use of RJ45 magnetics having 12 tiny transformer
cores per port. And their center-taps are wired to ground and to the Ethernet switch chip through capacitors to maximally block port- to-port leakage. This works best if the EtherREGEN is grounded—either through the power supply or via the ground screw.
The DC output 0-volt/-VE (‘ground’/shell) of the UpTone-branded AC>DC power supply included with the EtherREGEN is common to AC mains ground. When our supply is used to power the EtherREGEN it is NOT necessary to use the Ground screw. [Do not assume that 3rd party linear power supplies are like this—some are and some are not. Some multi-rail supplies have separate diode bridges and transformer secondaries and thus each output ground is "floated.” This is easy to test with a continuity/multimeter (be sure power is disconnected.)]
If you power the EtherREGEN with a "floating" supply—one whose 0- volt/-VE (‘ground’/shell) is not common to AC mains ground—then ideally the Ground screw of the EtherREGEN should be used. You can do so by running an insulated wire safely to some AC mains ground or to the chassis of a known-grounded component in your audio system. [By the way, our UltraCap LPS-1.2 is a completely "floated" and isolated supply, so if using it with EtherREGEN the ground screw should be used.]
Please keep in mind that all of the above regarding grounding concerns only the blocking of leakage from one RJ45 jack on the ‘A’ side to another RJ45 jack also on the 'A' side. Regardless of grounding, zero AC leakage from upstream Ethernet devices ever gets past the full active digital isolation and reclocking ADIMTM—to the EtherREGEN’s 'B'-side port.
    

What makes you think that ?

 

Yes. Why not purchase a second JS-2.😀

If you have a LPS-1.2, you can power that with JS-2. 

 

Actually the output of the JS-2 is isolated from safety ground on the AC power cord so yes the use of the ground screw is advised. The 'our supply' @R1200CL is referring to is the Mean Well SMPS.

[skatbelt]

I posted this already in another thread but maybe interesting as a datapoint for some here as well: recently I switched from the Uptone JS-2 to a Ferrum Hypsos powering the EtherREGEN. Against my expectation, no small change! I have a very 'clean' EtherREGEN setup with only fiber on the A-side (coming from a Sonore opticalModule/LPS1.2 combo). The B-side is connected to a dCS Network Bridge. The Hypos is powered by a dedicated power line (as was the case with the JS-2). DC cable: Ferrum's own. With the JS-2 I was using a very short Ghent Gotham GAC-4/JSSG360 cable.

[skatbelt]

2 hours ago, sgb said:

Just curious, what voltage did you set your Hypsos to for the etherRegen. I set mine to 10 volt.

 

I use the Hypsos etherREGEN preset: 9V.