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

[dbastin]

10 hours ago, kostas6a3 said:

All ethernet cables are factory built and I have no intention of opening them up.

 

7 hours ago, Mihaylov said:

But you can at least check with a multimeter whether the rj45 connector enclosures are connected at both ends of the same cable (should not be connected or you may lose some of the benefits of using audiophile switches).

 

If the cables shields are connected to ground at both ends, find a manufacturer of cables who does not do this.  It may matter more in some locations of your network than others.  In one location of my network (way upstream), I replaced a costly audiophile ethernet cable with shields connected at both ends with a well built but low cost UTP Cat 6a and gained an improvement.  However way downstream the UTP Cat 6a was inferior.

[dbastin]

Also, @kostas6a3, I gather Roon Rock is an OS that runs the stock Linux RoonServer.  So it seems you have 2 servers.  I had a quick look thru the Melco manual and it appears there are not instructions to uninstall minimserver or TwonkyServer.  If you could, the N100 would likely have simpler processing and perhaps sound better.

[kostas6a3]

4 hours ago, dbastin said:

Also, @kostas6a3, I gather Roon Rock is an OS that runs the stock Linux RoonServer.  So it seems you have 2 servers.  I had a quick look thru the Melco manual and it appears there are not instructions to uninstall minimserver or TwonkyServer.  If you could, the N100 would likely have simpler processing and perhaps sound better.

I am completely satisfied with the performance of my system so far.

Some Supra CAT8 cables I tried without the external patch cord attached degraded the SQ.

As for Roon Rock and Melco I have not understood what is meant. Roon Rock is the core and Melco is the server.

My original question is still unanswered.

Which optical fiber to use and which SFP module Finisar FTLF1318P3BTL or Finisar FTLX1475D3BTL ?

[Billy_SP]

1 hour ago, kostas6a3 said:

My original question is still unanswered.

Which optical fiber to use and which SFP module Finisar FTLF1318P3BTL or Finisar FTLX1475D3BTL ?

For the FTLX1465D3BTL i have used different versions of  duplex single mode, Mm62,5x125 Lc/lc.  Tried different brands and lengths and did not detect any difference.

What did make a noticeable improvement was upgrading the power supplies of the router and conversion box from a regular switching wall warts to iFi Power X.

[audiobomber]

10 hours ago, dbastin said:

If the cables shields are connected to ground at both ends, find a manufacturer of cables who does not do this.  It may matter more in some locations of your network than others.  In one location of my network (way upstream), I replaced a costly audiophile ethernet cable with shields connected at both ends with a well built but low cost UTP Cat 6a and gained an improvement.  However way downstream the UTP Cat 6a was inferior.

Cables with ground connections on both ends are perfectly useable, as long as the devices on either end are properly grounded.

 

Also, some devices (switches, routers, etc.) have plastic ethernet ports, therefore the connected cables will not be grounded. You can check whether a device's ports are connected to ground. Simply attach two cables that you know have connected shields, and check for continuity with a meter. If no continuity, the ports are not grounded.

[dbastin]

38 minutes ago, audiobomber said:

11 hours ago, dbastin said:

If the cables shields are connected to ground at both ends, find a manufacturer of cables who does not do this.  It may matter more in some locations of your network than others.  In one location of my network (way upstream), I replaced a costly audiophile ethernet cable with shields connected at both ends with a well built but low cost UTP Cat 6a and gained an improvement.  However way downstream the UTP Cat 6a was inferior.

Expand  

Cables with ground connections on both ends are perfectly useable, as long as the devices on either end are properly grounded.

I had previously said "Some cables or devices are preferable when shields are not connected to the ground or headshell at the downstream end" rather than implying we should never use cables with shields connected to ground at both ends 

[Mihaylov]

3 hours ago, audiobomber said:

Cables with ground connections on both ends are perfectly useable, as long as the devices on either end are properly grounded.

It is not always easy to determine whether the device is properly grounded (and not everyone will be able to do this), so a universal way will be the use of ethernet cables that have only one end of the connector body connected to the screen. 😉

[russellbobby]

Is it ok to use a Farad 7 volt LPS without risking damage to the EtherRegen? Looks like the 9 volt may be better suited but have a chance for a nice deal on a used one. 
thanks,

Russ

[Superdad]

1 hour ago, russellbobby said:

Is it ok to use a Farad 7 volt LPS without risking damage to the EtherRegen?

Yes, 7V is fine.  A regulated supply anywhere in the +DC range of 7~12V is fine.

[russellbobby]

Thanks    I know you mention the Stock supply is fine but damn this hobby :)

[barrows]

4 hours ago, Mihaylov said:

It is not always easy to determine whether the device is properly grounded (and not everyone will be able to do this), so a universal way will be the use of ethernet cables that have only one end of the connector body connected to the screen. 😉

Additionally:  Grounding problems (current flowing on ground connections) can occur anytime there are two ground connections to single component.  When there are two ground connections, almost always these two connections will be of different impedances to ground-if these two connections are of different impedances to ground, current will flow, and can induce noise into the system.  The way to avoid this is to have only a single ground connection to each component.  Now consider than many/most components with a metal chassis will have a ground connection through the AC connector/cable.  If we now add an additional ground connection through an attached Ethernet cable via a shield grounded at both ends, we have a potential problem/source of noise.  Best practice to avoid these problems is to not use Ethernet cables which have a shield connected to the shell at each end.

Now, I am not guaranteeing that a shield connected at each end of the cable will cause a problem in every situation, but I am guaranteeing if there is no shield connected at both ends of an Ethernet cable, the cable will not be introducing ground based noise problems on its own.  This is a simple step to take to avoid problems.

[GryphonGuy]

 

14 hours ago, kostas6a3 said:

I am completely satisfied with the performance of my system so far.

Some Supra CAT8 cables I tried without the external patch cord attached degraded the SQ.

As for Roon Rock and Melco I have not understood what is meant. Roon Rock is the core and Melco is the server.

My original question is still unanswered.

Which optical fiber to use and which SFP module Finisar FTLF1318P3BTL or Finisar FTLX1475D3BTL ?

 

FTLF1318P3BTL is a 1gb/s transceiver

FTLX1475D3BTL is a 10gb/s transceiver

 

If the switch ports at both ends of your fibre cable are at the same speed it doesn't matter which one you buy. If one switch port is SFP (1gb/s) ER is this speed, and the other is 10gb/s then I would opt for the FTLF1318P3BTL 1gb/s model to save handshaking/connection issues with both ends being different default speeds.

 

I purchased the 10gb/s model and thought it was quite coloured in the sounds transported by it so I placed them where they least influenced the sound with their colour, feeding audio out from my NAS (the start of the audio transmission).

 

I use the Fiberstore (FS.COM) single-mode transceivers SFP-10GLRM-31 where it matters inside the ER SFP port. They seem very transparent (uncoloured) to my ears. As always YMMV.

[dbastin]

7 hours ago, Mihaylov said:

It is not always easy to determine whether the device is properly grounded (and not everyone will be able to do this), so a universal way will be the use of ethernet cables that have only one end of the connector body connected to the screen. 😉

It is a puzzle.  Trial and error (by listening) is needed.  Many audiophile cables are ground the shields at both ends.  I have Shunyata Sigma and JCAT Signature and both Shunyata and JCAT advise they connect shield at both ends because it sounds best - which leaves me wondering "sounds best when connected to what?".

 

This is a compelling explanation, many thanks to the author for illuminating some of the mystery ...

 

2 hours ago, barrows said:

Additionally:  Grounding problems (current flowing on ground connections) can occur anytime there are two ground connections to single component.  When there are two ground connections, almost always these two connections will be of different impedances to ground-if these two connections are of different impedances to ground, current will flow, and can induce noise into the system.  The way to avoid this is to have only a single ground connection to each component.  Now consider than many/most components with a metal chassis will have a ground connection through the AC connector/cable.  If we now add an additional ground connection through an attached Ethernet cable via a shield grounded at both ends, we have a potential problem/source of noise.  Best practice to avoid these problems is to not use Ethernet cables which have a shield connected to the shell at each end.

Now, I am not guaranteeing that a shield connected at each end of the cable will cause a problem in every situation, but I am guaranteeing if there is no shield connected at both ends of an Ethernet cable, the cable will not be introducing ground based noise problems on its own.  This is a simple step to take to avoid problems.

 

Here are examples of this puzzle in practice.

 

I have a national network termination device that receives fibre and outputs RJ45.  It has multi rail SMPS plugged into wall outlet.  This is connected to an ER with a LPS plugged into the same outlet - the ER send fibre downstream.  I am not sure if the resistance on the ground of the SMPS and LPS are different.  I tried a number of ethernet cables between these two devices, including low cost Cat 5e UTP and Cat 6 UTP, Afterdark cable with shield connected at one end, Wireworld Platinum Cat 8 cable with shield connected at both ends and he one I felt was most accurate is a Cat 6a UTP that seems a lot more well built than the low cost ones, even through the Wireworld Platinum is an excellent cable in different locations.

 

I use Mikrotik CRS 305 as router that received fibre from the ER mentioned above.  It is powered by a W4S LPS via Synergistic Research (SR) Power cable and powercell, and both the router and LPS are grounded via SR HD ground wire to SR Active Ground Block SE.  The only wired connection from this router is to an Antipodes EX server also grounded via SR HD ground wire - I also note the EX chassis and PCB are grounded to AC earth by its design.  I tried the AfterDark (shield connected at one end), JCAT (shield connected at both ends) and Synergistic Research Atmosphere X Reference between the router and EX, and I consider the SR cable is most accurate there - the SR cable shield is connected only at the sending end, and was further improved by using the cable ground connection via HD ground wire to the same ground block mentioned above.

 

The JCAT and SR cables are both extremely good but with different 'flavours'.

 

Finally I have a Mikrotik mAP lite (WAP) connected to the EX - the mAP is powered by a lithium USB powerbank battery and not grounded.  The last time I tried, the Shunyata Sigma cable works best for this connection, compared to the Cat 6a UTP, AfterDark, JCAT and SR - the Sigma shield is connected at both ends.  Nevertheless, the mAP and its battery still benefited when I recently added an Akiko USB stick to the battery which reduced noise even further.

 

The mAP sends wifi to the endpoint (Devialet Pro), and despite this wifi connection and the fibre connection mentioned above, the differences in the cables I tried (above) are still not difficult to hear.

 

So in summary, three cables in different locations/situations with different shield arrangements.

 

The lessons for me from all this are:

• listen to the network gurus
• don't assume too much
• consider the possibilities
• be patient with trial and error

I hope this is helpful and thought provoking.

[dbastin]

18 hours ago, kostas6a3 said:

As for Roon Rock and Melco I have not understood what is meant. Roon Rock is the core and Melco is the server.

 

N100 is the player/renderer/streamer (these are generally interchangable terms to name the thing that prepares and sends data to a DAC).

 

N100 is also a server if the minim server or twonky software has been installed.  Roon Rock is also a server., Maybe you could benefit form watching this ...

https://antipodes.audio/support/wp-content/uploads/2021/07/anatomy1.mp4?iframe=true

 

With 2 servers there could be extra processing and traffic etc that may be detrimental and unnecessary.

 

With 1 server and 1 separate player, the processing workload of the player is reduced and not diminished by any server functions/processing.

 

I am not sure of you can or want to roll back the N100 to factory settings to uninstall the server software.

 

I have heard SQ improvements on many occasions when roon tweaks its code in various updates.  These are small compared to separating the server from the player, but it might depend on the hardware. 

[R1200CL]

18 hours ago, kostas6a3 said:

My original question is still unanswered.

Which optical fiber to use and which SFP module Finisar FTLF1318P3BTL or Finisar FTLX1475D3BTL ?

I suggest you use FTLF1421P1BCL instead of 1318. Also probably no need spending money on the 1475. 
 

You shall look for a yellow cable with blue connectors. (Single mode fiber). 

[Mihaylov]

13 hours ago, dbastin said:

So in summary, three cables in different locations/situations with different shield arrangements.

 

It seems to me that your comparisons are not entirely correct precisely because of the use of three different cables - to identify the effect of connecting the screen on the sound quality in the system, you need to use the same cable, changing the way the screen is connected to the rj45 connector. Otherwise, you hear not only the influence of the way the screen is connected, but also the influence of the cable itself of different brands (i.e. its material, design, etc.).
So often it is not possible to change the way the cable connectors are connected to the screen on the same cable without interfering with the cable itself, it causes difficulties in determining which cable (with which way to connect the screen) to use. That is why a cable with a screen connection from one end will be universal. And in the process of listening, you can turn the cable over to choose the direction of the best sound . (the direction of the cable is also possibly caused by the peculiarities of the cable material, but changes after the cable burned up).

[barrows]

11 hours ago, R1200CL said:

I suggest you use FTLF1421P1BCL instead of 1318. Also probably no need spending money on the 1475. 
 

You shall look for a yellow cable with blue connectors. (Single mode fiber). 

Is it possible that you might be able to describe the subjective qualitative sound differences which you hear between these two SFPs?  I am curious.

[kostas6a3]

The cable was found here, last piece :

https://uk.farnell.com/phoenix-contact/1115636/fibre-cable-lc-duplex-lc-duplex/dp/3462889

[GryphonGuy]

21 hours ago, R1200CL said:

I suggest you use FTLF1421P1BCL instead of 1318. Also probably no need spending money on the 1475. 
 

You shall look for a yellow cable with blue connectors. (Single mode fiber). 

 

That model FTLF1421P1BCL is flagged as obsolete in some websites.

[GryphonGuy]

I have a newish EtherREGEN that I unboxed last night for a special project.

 

I run a dante audio system here for recording and general audio distribution. In simplistic terms dante is just ethernet with timing protocols and targetted broadcast messages on top to make sure data arrives in time to who its meant to be sent to.

 

I was having trouble getting the PC to communicate well over this new network extension. My existing network is mostly fibre but in this instance I had to use copper-only. Anyway in the traditional sense of input or network side being ER side "A" and the PC on "B", the timing protocol was reporting 46-48 millisecond latency (response time) when it should be a maximum of 6 milliseconds. Swapped the cable on the "A" side between the copper ports on the "A" side and nada. No difference. Still 46-48 milliseconds.

Then I had an idea and swapped so that the input or network side was side "B" and PC was now on the "A" side copper ports. 2-3 millisecond response time and now well within the 6 millisecond limit for good audio.

 

So who knew that "B" to "A" seems to have a faster path than "A" to "B"?

 

Just sayin'. No complaints.

 

Regards

GG

[dbastin]

10 hours ago, Mihaylov said:

It seems to me that your comparisons are not entirely correct precisely because of the use of three different cables - to identify the effect of connecting the screen on the sound quality in the system, you need to use the same cable, changing the way the screen is connected to the rj45 connector. Otherwise, you hear not only the influence of the way the screen is connected, but also the influence of the cable itself of different brands (i.e. its material, design, etc.).
So often it is not possible to change the way the cable connectors are connected to the screen on the same cable without interfering with the cable itself, it causes difficulties in determining which cable (with which way to connect the screen) to use. That is why a cable with a screen connection from one end will be universal. And in the process of listening, you can turn the cable over to choose the direction of the best sound . (the direction of the cable is also possibly caused by the peculiarities of the cable material, but changes after the cable burned up).

It was not scientific, just observations and some presumptions/speculation about the influence of the shields in each situation.  Really what I was hoping to convey is:

• a given cable can have a different effect when used in different locations in the network
• be careful not to assume that if a cable provides an outcome in one situation that it will have the same outcome in every situation
• it can help to pay attention to how each cable design deals with the shield

Let's say we have 3 cables with shield connected at both ends (Group 1), and 3 cables with shields connected at one end (Group 2). 

• When used to connect devices A and B, it could be that Group 2 cables give the best sound and have more similarities than differences because devices A and B give the best sound with shield connected at one end. 
• But when used to connect devices A and C, it could be that Group 1 cables give the best sound and have more similarities than differences because devices A and C give the best sound with shield connected at both ends. 

I suppose the way to work out what sounds best between any two devices is to directly compare say 3 cables that are identical (including same manufacturer, dielectric etc) except:

• 1 has no shield (UTP)
• 1 has shield connected and one end
• 1 has shield connected at both ends.

I suspect it would be necessary to make them, or have them made, and would be general rather than 'audiophile' cable materials/construction.

 

This would not take into account how some 'audiophile' cables include dielectrics or sheaths etc that could attenuate or absorb noise from the shield.  For instance, It could be possible that the Shunyata CMODE Modules actually attenuate noise not only from the twisted pairs and cable shield but also from the device it is connected to (like Akiko tuning sticks do), and this could be why the designers decided it sounds best with shield connected at both ends.

 

Shield connected at one end is a safe bet, but there are so many other possibilities.  There are too many variables to methodically test, and the expense would be mind boggling.  We can only patiently trial and error the cable we can get our hands on. 

[R1200CL]

13 hours ago, barrows said:

Is it possible that you might be able to describe the subjective qualitative sound differences which you hear between these two SFPs?  I am curious.

Absolutely no 😂

However the technically differences is my reason to suggest this. As I’ve been writing about before. Just search with my name and 1421. (Or white rabbit).

 

[R1200CL]

11 hours ago, kostas6a3 said:

The cable was found here, last piece :

https://uk.farnell.com/phoenix-contact/1115636/fibre-cable-lc-duplex-lc-duplex/dp/3462889

These cables is very reasonable priced on eBay from China. 
 

https://www.ebay.com/itm/392274190204

[R1200CL]

5 hours ago, GryphonGuy said:

 

That model FTLF1421P1BCL is flagged as obsolete in some websites.

Have a look the official website. Several 1421:

https://ii-vi.com/product-category/products/optical-communications/optical-transceivers/?product-form-factor=sfp&product-protocol=ethernet&product-wavelength=1310nm-band&product-media-type=duplex-smf

FTLF1421P1xCL

FTLF1421P1xTL
FTLF1421P1BTL-3G
 

 

[Mihaylov]

8 hours ago, dbastin said:

directly compare say 3 cables that are identical (including same manufacturer, dielectric etc) except:

• 1 has no shield (UTP)
• 1 has shield connected and one end
• 1 has shield connected at both ends.

My guess is that the same manufacturer connects the screen the same way for all of its erhernet cables. At least Audio Quest is the name of it – all of their cables have a connection to the screen from only one end of the cable.