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

[GryphonGuy]

 

First of all, thanks to those of you who offered advice and recommendations during my tests. This was both educational and then fun to hear the differences as I changed just about every connection at one time or another. I said I was going to go off and buy dedicated single-mode transceivers but I ended up buying more single-mode fibre too.

 

So first of all let me address the loss of bass, accentuating non-bass sounds and the flattening of the soundstage using dual-mode 1310nm wavelength signals over single-mode fibre in my last post. Replacing the dual-mode SFP+ transceivers with dedicated single-mode SFP+ transceivers brought back the bass but the soundstage was still flat (no depth). I perceived the image to be at the tip of my nose using headphones.

 

So I reverted to the dual-mode 1310nm over multi-mode fibre. This allowed the tonal balance to return and a fairly deep soundstage. Whew!

 

So I looked at the fibre connection to the server. I changed it from 850nm multi-mode to intel compatible SFP+ 1310nm single-mode transceivers. This provided nice sonic purity uplift and little to no depth removal. Nice!

 

So I put the EtherREGENs back to single-mode SFP+. Wow. The sound was quite nice but still a bit “off”. Slightly better sonic purity than before and a depth about the same.

 

So by looking at the schematic diagram below, I knew that my PC was dragging the music over its normal network connection before sending out the dedicated Audio VLAN (Audio VLAN was now all single-mode SFP+ 1310nm wavelength). That normal PC connection was still at 850nm multi-mode fibre. So I switched that out for SFP+ single-mode 1310nm fibre as well and … Holy Crap Batman! The transient attacks of drums and percussion instruments was palpable. The depth immense (I perceived some sounds to be about 3 metres or more behind the singer, for example). So it was my PC’s normal connection dragging down the soundscape all this time. Interesting that the server and my PC have the exact same model intel fibre network card in them but the server only made a small change to the sound versus my PC a huge change.

 

But after listening for an album or two of familiar music, the best way to describe the sound was that it was shiny. Slightly too much emphasis on the high frequencies for my liking.

 

Just in time, the Finisar FTLX1475D3BTL SFP+ transceivers (made in Malaysia about 1 hour from where I am staying, by-the-way but cannot buy here!!!!) arrived from the USA. So I replaced the EtherREGEN SFP+ modules with the Finisars and in the switch as well. OUCH! Flat soundstage again and overly “fat” sounding with the space between the instruments and voices disappearing and were replaced with wall-to-wall sound. Mid-bass was bloated. “Let them run-in” my brain said, “they are electrical components after all”. So I left them running for about 3 hours and returned. Bummer! No real perceivable difference.

 

A lot of experimentation on positioning these expensive SFP+ modules then occurred.

 

The final iteration of Finisar placement was at the rear of the server only, one Finisar there added a nice touch of finesse to the sound without too much degradation. The switch was an FS SFP+ module. There was degradation and colouring but only marginally, and like a tube presentation, the colouration is “nice”. Ultimately the unlikable shine had been removed but a bit of colour added.

 

After days of settling, this combination (see diagram below) is simply amazing. Kick-drums sound very similar to real kick-drums in that the transient of the striker is fast with the after resonance very low and slow. Well I can feel that and I think that is what is making the sound seem so real. The clarity and positioning in the background of, say, a triangle struck just once can be very clearly heard along with its decay a long way back in the soundstage like it would if you heard it live.

 

What bugs me about my installation is that now my music goes through 4 conversions from light to electrical and 4 conversions from electrical to light. But somehow manages to sound as close to real audio as I think I am going to get out of electronic reproduction.

 

Of course, I am not stopping here. I still have to get customs to tell me how much for the taxes on the SR4T for my 2^nd EtherREGEN, and I am sure in the next few months I will have enough to buy a 10MHz external clock for both the EtherREGENs to see what more I can ring out of this system.

 

So to those of you who urged me to look into single-mode SFP+, I thank you. It has been an incredible journey of discovery and non-belief in what I was hearing at times. This hobby can be so much fun at times like these.

 

All the best

GG

 

PS. By way of explanation and for the keen-eyed, you’ll notice my PC has two network connections. I use Dante hardware and software to send audio throughout my setup and I keep that separated from my main network because of the constant network traffic synchronising clocks in the Dante devices. It is point-to-point protocols (non broadcast) but never-the-less, I chose to keep it separated.

 

[GMG]

17 hours ago, Avalfa said:

Depends on where you are on the globe.

 

It's 'just' €2750.- where I am

 

Over here ER is around €800 lps1,2 €450

Clock€800-5000 depending on  which one another LPS 450-700 and than cable etc. 

 

So it's depending on so many factors, and than again I think few people have heard the Innuos yet, I haven't. So if it's any better? I Don't know. 

I just think it's interesting

 

Though the description is a bit confusing, maybe the bellow @AfterDark. product is a worthy comparison to the phoenix,

It seems to be an OCXO controlled switch, with SFP option and powered by an internal LPS + has an OCXO out,

So, it's an all in one like the phoenix but with more functionality? (SFP, OCXO clock out)

 

Not affiliated with Afterdark - just my thoughts here

 

https://www.adark.co/collections/project-giesemann-ocxo/products/afterdark-project-clayx-giesemann-network-switch-x-ocxo-10m-master-clock-audiophilestyle-edition?variant=32208640540790

[Rsbrsvp]

1 hour ago, GMG said:

 

Though the description is a bit confusing, maybe the bellow @AfterDark. product is a worthy comparison to the phoenix,

It seems to be an OCXO controlled switch, with SFP option and powered by an internal LPS + has an OCXO out,

So, it's an all in one like the phoenix but with more functionality? (SFP, OCXO clock out)

 

Not affiliated with Afterdark - just my thoughts here

 

https://www.adark.co/collections/project-giesemann-ocxo/products/afterdark-project-clayx-giesemann-network-switch-x-ocxo-10m-master-clock-audiophilestyle-edition?variant=32208640540790

How could it be that inexpensive if it has a built in super oxco and LPS?

 

If it is an all in one- why would anyone buy an ER ever again when for less than double you get the LPS and OXCO built  in?

[LowMidHigh]

54 minutes ago, Rsbrsvp said:

How could it be that inexpensive if it has a built in super oxco and LPS?

 

If it is an all in one- why would anyone buy an ER ever again when for less than double you get the LPS and OXCO built  in?

 

It seems that you judge equipment by cost alone.

I don't believe many people share your view. 

 

[MasterWarzombie]

I have a remark.
Many people praise the addition of this or that device to the etherregen but no one talks about the impact of the room and its acoustic treatment, nor on the behavior of the speaker in that room.
If you have an acoustically rotten room, racking up tweaks won't get better.
A lot of acoustic work, passive or not, should be considered before investing.

[GMG]

14 minutes ago, MasterWarzombie said:

I have a remark.
Many people praise the addition of this or that device to the etherregen but no one talks about the impact of the room and its acoustic treatment, nor on the behavior of the speaker in that room.
If you have an acoustically rotten room, racking up tweaks won't get better.
A lot of acoustic work, passive or not, should be considered before investing.

I don't think anyone would argue with you that room acoustics are one of the most important ingredients in the quest, but it is not the subject of this thread of the forum :-)

 

I use the ER to feed a MiniDSP SHD studio which runs DRC (digital room correction) which takes care of some of the acoustical issues in my room.

[Rsbrsvp]

31 minutes ago, LowMidHigh said:

 

It seems that you judge equipment by cost alone.

I don't believe many people share your view. 

 

O.K.  I did not think of my comment as a judgment.  It was a question?  How can a ER plus LPS plus OXCO sell for $1,100?

 

My assumption behind the question is that perhaps I do not understand properly what the product is...

 

I am asking for clarification to help me understand; I am not judging anything.

 

Perhaps you have misjudged my intentions due to my lack of clarity in stating my question? 

[GMG]

No one suggested that the network switch in this box is equivalent to the ER,

It is a switch (of some sort and quality) + OXCO + SFP + LPS,

 

[Rsbrsvp]

Thanks.  That is what I was looking for; a clarification...

[charlesphoto]

2 hours ago, MasterWarzombie said:

I have a remark.
Many people praise the addition of this or that device to the etherregen but no one talks about the impact of the room and its acoustic treatment, nor on the behavior of the speaker in that room.
If you have an acoustically rotten room, racking up tweaks won't get better.
A lot of acoustic work, passive or not, should be considered before investing.

Or just use headphones. Sorry, poor argument against source tweaking (and yes, the room is an important factor in the overall sound)

[Rsbrsvp]

Is there any conclusion about using attenuators on the SFP input of ER or is it system/ear dependent?

[MarkusBarkus]

Was there an objective for the addition of attenuators to the chain or just an experiment? Sorry if I missed that in the thread. FWIW: I have been trying to remove gadgets, where possible, unless there is a problem to address...

[kennyb123]

16 hours ago, Rsbrsvp said:

Is there any conclusion about using attenuators on the SFP input of ER or is it system/ear dependent?

Some SFPs will oversaturate the receiver and this can harm the sound.  An attenuator will reduce this harm.  But to my ears attenuators do harm.  Better to use SFPs that don’t need them.  

[MarkusBarkus]

...wouldn't that be multi-mode fiber, at typical home-audio distances of fiber runs?

[agladstone]

On 5/3/2021 at 1:11 PM, Rsbrsvp said:

Please forgive me for the question if this has been covered:

 

Does the ER need to be on 24/7 to sound its best or is a one or two hour warm up enough?

 

It gets so hot, I want to extend it's life so I prefer to turn it on only when necessary.

 

I thought at least to turn it on in the early afternoon and off at night before bed.  I listen to music in the evening....  That would allow a 4-5 hour warm-up time.

I’m curious what voltage and amps the power supply is at that you’re using to power eR ? 
I keep reading from most how hot the eR gets, however, I’ve been powering mine from day one with a 12V / 3A power supply (Vinne Rossi DC4EVER super cap) and I’ve never turned off my eR and it has never gotten too hot. 
I have a suspicion that 7v or 9V is creating more heat (just a hunch, I’ve never experimented myself and I don’t know enough about electric from an engineers perspective)? 

[Mike Rubin]

4 minutes ago, agladstone said:

I’m curious what voltage and amps the power supply is at that you’re using to power eR ? 
I keep reading from most how hot the eR gets, however, I’ve been powering mine from day one with a 12V / 3A power supply (Vinne Rossi DC4EVER super cap) and I’ve never turned off my eR and it has never gotten too hot. 
I have a suspicion that 7v or 9V is creating more heat (just a hunch, I’ve never experimented myself and I don’t know enough about electric from an engineers perspective)? 

I use an LPS 1.2 set to 12v and the eR definitely gets hot, even when placed in a stand that acts as a bit of a heat sink.

[agladstone]

2 minutes ago, Mike Rubin said:

I use an LPS 1.2 set to 12v and the eR definitely gets hot, even when placed in a stand that acts as a bit of a heat sink.

The LPS 1.2 is about 1A at 12V, whereas I’m using 12V / 3A , not sure if that would effect heat or not, but if the LPS 1.2 is getting near its max Amp, it may ? 

[JohnSwenson]

The ER is a "constant power" device. It dissipates the same amount of power no matter what voltage you feed it (from 7 to 12 volts).

 

About 70% of the power is fixed, it is always there no matter what you are doing. The rest is determined by how much data is flowing through it. The lowest power will be one "input" (on an A side RJ45 jack) and an output on the B side RJ45 jack. Just doing 44.1 PCM will be the lowest. 384 or DSD512 will take a little more power. Using the SFP port takes a little more power.

 

The major increase is caused by using the other A side ports with lots of data flowing through the ports. Say running 4, 4K TVs off the A ports as well as your audio will get the ER power draw up to its maximum.

 

I don't HAVE 4 4K TVs to actually test what that is, but if anybody wants to donate that to me I'll be glad to run the test!

 

John S.

[agladstone]

2 minutes ago, JohnSwenson said:

The ER is a "constant power" device. It dissipates the same amount of power no matter what voltage you feed it (from 7 to 12 volts).

 

About 70% of the power is fixed, it is always there no matter what you are doing. The rest is determined by how much data is flowing through it. The lowest power will be one "input" (on an A side RJ45 jack) and an output on the B side RJ45 jack. Just doing 44.1 PCM will be the lowest. 384 or DSD512 will take a little more power. Using the SFP port takes a little more power.

 

The major increase is caused by using the other A side ports with lots of data flowing through the ports. Say running 4, 4K TVs off the A ports as well as your audio will get the ER power draw up to its maximum.

 

I don't HAVE 4 4K TVs to actually test what that is, but if anybody wants to donate that to me I'll be glad to run the test!

 

John S.

Thanks for the actual engineer explanation! 
I only have one Ethernet in on A side, and one Ethernet out on B side. 
Also, I use the EtherRegen with my Aurender streamer / server, which loads everything onto an internal SSD first, prior to playback, so perhaps this is why my never gets too hot like others have claimed ? 

[Pokey77]

23 minutes ago, agladstone said:

Thanks for the actual engineer explanation! 
I only have one Ethernet in on A side, and one Ethernet out on B side. 
Also, I use the EtherRegen with my Aurender streamer / server, which loads everything onto an internal SSD first, prior to playback, so perhaps this is why my never gets too hot like others have claimed ? 

I have the same situation as you. ER sits under a desk with plenty of airflow but is very warm. Had it in the system over a year now and never a hiccup.

[PYP]

3 hours ago, JohnSwenson said:

About 70% of the power is fixed, it is always there no matter what you are doing. The rest is determined by how much data is flowing through it. The lowest power will be one "input" (on an A side RJ45 jack) and an output on the B side RJ45 jack. Just doing 44.1 PCM will be the lowest. 384 or DSD512 will take a little more power. Using the SFP port takes a little more power.

Interesting.  I have two eRs (both with one input).  The first is A > B, fiber in and copper out.  The second is A > B, copper in and out (to the DAC).  The second one uses an external clock.  The second one runs much hotter than the first, but I'm assuming that results from differences in air flow within my cabinet.   

[Rsbrsvp]

 

 

 

My entire system is in a closed cabinet.  Of course, I make sure there is airspace around all my components within the cabinet, but still- in a cabinet which is often closed and in a hot climate.

 

So, I am always a bit concerned about heat when leaving everything on 24/7.  I have received confirmation from every component manufacturer that it is O.K...  so- I hope it is.

 

I generally shut off my entire system (except the Mutec Ref10 which runs pretty cool) once a week for around two hours just because of my heat paranoia...  Perhaps this is unnecessary.

 

[PYP]

16 minutes ago, Rsbrsvp said:

 

 

 

My entire system is in a closed cabinet.  Of course, I make sure there is airspace around all my components within the cabinet, but still- in a cabinet which is often closed and in a hot climate.

 

So, I am always a bit concerned about heat when leaving everything on 24/7.  I have received confirmation from every component manufacturer that it is O.K...  so- I hope it is.

 

I generally shut off my entire system (except the Mutec Ref10 which runs pretty cool) once a week for around two hours just because of my heat paranoia...  Perhaps this is unnecessary.

 

During the summer, a 5" X 5" quiet fan is on all the time (plugged in AC outlet).  The difference in the external temperature of the eRs and LPS 1.2 is dramatic.  From very, very hot to warm.   An inexpensive ($19) experiment.  

In the US:  https://smile.amazon.com/gp/product/B01MZ6T8VF/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1     

[JohnSwenson]

I have previously posted here that the ER was designed to run at an optimal temperature of 50C, which is quite warm. So by being obsessive about keeping it cool you can actually hurt performance. I don't quite understand the mind-set of spending thousands on power supplies and external clocks and then run it too cool. Running an ER at around 50C (do not obsess about tenths of a degree here, a 5 degree spread on either side is fine) is an optimal point where conflicting effects are at a minimum. An infrared thermometer on the black case works great for measuring the temperature. In the design of the ER we went to huge lengths to make sure that all parts are at almost the  same temperature after thermal equilibrium has been achieved, thus you don't need to worry about some parts being hotter than others.

 

50C is way below the maximum operating temperature of every part in the design. Running at 50C instead of say 30C will decrease the lifespan of the unit by something like 5 days. I made the assumption that the increased performance was going to be more important to most of the people here.

 

I can't tell you what temperature to run yours at, but I can say it was designed for best performance at 50C.

 

The only time I would worry about using heatsinks is if the temperature is above 60C.

 

John S.

[Rsbrsvp]

13 minutes ago, JohnSwenson said:

I have previously posted here that the ER was designed to run at an optimal temperature of 50C, which is quite warm. So by being obsessive about keeping it cool you can actually hurt performance. I don't quite understand the mind-set of spending thousands on power supplies and external clocks and then run it too cool. Running an ER at around 50C (do not obsess about tenths of a degree here, a 5 degree spread on either side is fine) is an optimal point where conflicting effects are at a minimum. An infrared thermometer on the black case works great for measuring the temperature. In the design of the ER we went to huge lengths to make sure that all parts are at almost the  same temperature after thermal equilibrium has been achieved, thus you don't need to worry about some parts being hotter than others.

 

50C is way below the maximum operating temperature of every part in the design. Running at 50C instead of say 30C will decrease the lifespan of the unit by something like 5 days. I made the assumption that the increased performance was going to be more important to most of the people here.

 

I can't tell you what temperature to run yours at, but I can say it was designed for best performance at 50C.

 

The only time I would worry about using heatsinks is if the temperature is above 60C.

 

John S.

Once again,

 

these posts by John are VERY meaningful.