DIY Ethernet Cable

[TomJ]

Hi Guys,

 

I have since found, on the advice of another forum, that CAT cables ripped, freed of all ballast (isulation, shielding, 2 twisted pair cables not used for 100BaseT) sound best - best resolution, detail and seperation of instruments/voices.

My LAN cables are only 2 twisted pairs from standard solid core cat 6a cables.

 

Now I am looking to improve this approach. 

 

I have been looking around the market for twisted pair cables (I want to use ready made twisted pair cables) that meet the specifications of Ethernet. Thereby I became aware of a new standard "Single pair ethernet". Ethernet is run over a twisted pair cable in full doublex. The cables are shielded and intended for industrial use. https://www.igus.de/info/single-pair-ethernet-spe-kabel
I first thought that the sound improvement comes from the high distance that the two twisted pairs have because of less interference and cross-talk.
But this is not so - my DIY "Double Single Ethernet" from two of these cables (one for TX and one for RX) can not reach the stripped cable.

 

Now I continue to search for cables that I can use for the DIY cable.

I have become aware of Belden Bonded technology. And that there is Plenum Cat cable in the US (I live in Germany) that has Teflon (FEN) insulation on the wires.

Could use two of the twisted pairs for my DIY cable.

 

Has anyone here ever compared Plenum cables with standard Cat cables and seen a difference? 
Likewise with Bonded Technology: are there any audible differences?

 

I've also become aware of cables from Gore that use Gore insulation, which is supposed to be superior to regular PTFE.
The cables are hard to get, as they are actually designed for military use and only sold there.
On Ebay there is a store that sells cables from this: https://www.ebay.com/itm/380833692654
The question here, however, is whether these cables have the required 100Ohm impedance.

 

I'm curious if someone has suggestions or experience on this.

 

Regards,

 

TomJ

 

 

[TomJ]

No one out there with some experience?

[TomJ]

Hi all,

 

thanks for your recommendations.

 

But regarding Ethernet cable and sound I think it is not about data integrity.

I was longtime a believer that if once the data arrives without error caused by jitter or something else, the sound could bot be effected.

But I learned that Ethernet cable can for sure effect the sound, maybe in the same way or  maybe more than analoge cable.

I think it depends on geometry and material. But this is very cloudy and one is groping in the dark. Maybe it has something to do how the cable handles the frequencies and produces itself HF noise which affect then the process aof converting digital to analog in the DAC.

 

I think ethernet engineers cant help here - they will think I'm crazy. 

All I can say is, that the stripped 2x2 TP cable without isolation are the most resolving i heard so far.

If you have an old patchcord, just give it a try - remove the overall isolation and cut the unused wires.

 

As I know the isolation (e.g. PTFE) of the wire in the analoge sector makes a difference, maybe this is also valid for Ethernet cable.

 

 

 

 

[TomJ]

2 minutes ago, lpost said:

Enjoy the experimentation, it's a hobby and since it's low voltage no one dies.

Yes, thats what I will do.

Problem with Belden Plenum cable is, I have to order it in the US with high shipping cost.

Meanwhile I have also an offer for Gore Ethernet cable made for airforce use - but 150€ per meter.....

[TomJ]

22 minutes ago, lpost said:

Be safe out there and don't hack Ethernet cables for no good reason.

Isnt music and sound a good reason?

[TomJ]

Regarding Belden Bonded technology, there is a good demonstration how this improves the impedance stability on youtube: 

 

[TomJ]

3 minutes ago, lpost said:

That is Galen Garis, the man.

Ok, will register on PS Audio and try to contact the master....

[TomJ]

1 minute ago, plissken said:

I just want to make sure that we are talking about something like this:

 

 

Exactly! But with solid core AWG23 wire..... 

[TomJ]

3 minutes ago, plissken said:

 

The potential issue that I have is that this is going to increase the error rate that the switch PHY operates at.

 

I think what is happening is someone started out doing that at 1GBe and found out that it didn't work. Forced the connection down to 100mbps and found out that it did since the tolerance isn't nearly as strict and then also realized you don't need 4 pair, only 2.

As my Squeezebox Touch only supports 100BaseT there are only 2 pair in use, even if 4 are connected.

[TomJ]

3 minutes ago, lpost said:

You'll find him here. He created Iconoclast cables while still working for Belden. It's an interesting story...

 

https://forum.psaudio.com/t/belden-iconoclast-interconnects-and-speaker-cabling/2038

But no ethernet cable from Iconoclast - yet . . . . 

[TomJ]

5 minutes ago, plissken said:

 

So someone can hear the difference between solid core (horizontal run structured cabling) and stranded patch?

I was in the same situation like you one year ago, then I made some tests only to proof my standpoint.

After this test and recognized i was wrong, I went into the basement and started screaming loudly. 

[TomJ]

4 minutes ago, lpost said:

I think this is all that could happen. Perhaps the increase in errors and subsequent retransmission creates a favorable 'sound'. At our normal paltry audio data rates, the link can probably tolerate 90% errors and still work just fine.

I have done tests with software analyzer tools and there is not more package loss, errors or delay with the 2x2 as with standard 4 x 2 cable.

[TomJ]

1 minute ago, lpost said:

No, and I wouldn't expect so. The data transmitted by an Ethernet cable isn't sound, it's TCP or UDP packets. Iconoclast and BAV cables are strictly analog, though some do meet the specs to be used as digital interconnects. Galen makes no claims of SQ improvements only measurements. He invites you to listen to them in your system and decide for yourself. The speaker cables make the most difference in my experience. They offer 3 different copper variants with identical geometry. He knows the geometry matters but what about the copper? It matters or at least makes a difference but it's much more about preference at that point. 

As I have KEF LS50 Wireless (driven by ethernet, not Wifi) i will not be his customer....
But as I said before, one year ago i would have absolutely agreed that bits are bits and no sound influence is possible. 

[TomJ]

1 minute ago, lpost said:

 

Is your Ethernet cable connected directly to your DAC?

No.

For LS these are connected to the KEF LS50W )with integrated streamer and DAC).

For my headphone setup the cable is connected to Squeezebox Touch, which is connected to RME ADI 2 DAC.

[TomJ]

7 hours ago, charlesphoto said:

Just get Blue Jeans Cat 6a (Belden cat 6a) and be one and done. Very neutral, robust build, inexpensive, tested, floating shield, and comes in a variety of colors and to whatever length you need. I went down the lower end 'audiophile' path and except for the Blue Jeans and Ghent (which is a slightly different Belden Cat 6a) the others colored the sound in ways I didn't like. Now I have an entire BJC 'loom' from the coaxial cable in onward. 

Interesting.

 

A quote of Gale Garis at PS Audio about Cat 6A Cable:

 

Oh, don’t use 6A UNLESS you really have 10G. 6A is designed as cat6 internal, not as good as 3600, 4800 or DT600E. The reason is the special and unique designs to reduce EXTERANAL ALIEN NEXT impact the ability to reach optimum internal NEXT figures. Only 6A needs a balance of internal to external crosstalk properties.

6A will NOT be as economical or even as good on 10/100/1000 Ethernet. It isn’t made for that. CAT7, by using ISTP design can manage both tecnologies as external and internal are reduced with the shield over each pair. Overall shields called FTP around the group of 4 pairs are for external noise, but not internal noise shielding. And, outer FTP shield reduce internal NEXT performance as the EM wave are coupled into the pairs more aggressively. But, the external FTP shield reduces EXTERNAL or alien NEXT, called ANEXT. There is a balance of internal to external noise mitigation."

[TomJ]

46 minutes ago, charlesphoto said:

 

No idea what all that gobbleydook means. In my testing BJC Cat 6a sounded slightly and subtle better than the BJC Cat 6. Lots of users over on the Naim forum that also preferred the 6a to the 6. I've tried generic Cat 7 (Teradak?) and it doesn't sound as good. Keep in mind the shield on BJC 6a is untied at one end, so is floating, compared to the metal shrouds on the connectors of Cat 7 that fully connect the shield, which is what we don't want in domestic hifi. 

 

I only posted the quote of Gale Garis, who was like, mentioned before from @lpost , the leading developer at Belden, to show that data integrity / signal transmission and sound of ethernet cable are different pair of shoes.

[TomJ]

9 minutes ago, recepky said:

I get really positive results with 2 pairs awg26 cat5 cable. Without shield without anything fancy I think if you have curiosity just find a cheapest ethernet cable around you cut other 2 pair and jacket and give it a chance. 

So here we are - I am not alone!

 

I had exactly the same experience.
I am now looking to see if a better cable material and dielectric will bring another improvement. 

 

Will get this week twisted pair cable from Gore with silver plated copper and ePTFE dielectric.

[TomJ]

The problem is to match the characteristic impedance of 100 Ohm.

Therefore I ordered single twisted pair with exactly 100 Ohm.

 

I have also produced a DIY cable with solid twisted pair 23AWG copper from a horizontal cat 6 cable with the same result.

Best resolution and details.

 

I'm glad not to be alone with this phenomenon - in most forums, I'm being labeled as crazy 

 

If you combine this kind of cable with Ethernet Isolator you get incredible details, separation and 3D holographic.

[TomJ]

3 minutes ago, recepky said:

where do you find 100 ohm twisted pairs?

which ethernet isolator do you use?

should I place the isolator just before the computer or just after the switch?

Its this cable: https://www.gore.com/resources/shielded-twisted-pair-cables-aircraft-drawing-dxn2602

Its hard to get, because it is for military use.

I used this Isolator: https://www.amazon.de/DeLock-62619-DrNetzteilucker-grau/dp/B00WZ3QHVQ

There are also medical grade isolators on the market which some use, but these are not on the level regarding filtering as the Delock.

 

Meanwhile i built myself an own Isolator out of ethernet magnetics.

 

The Isolator should be placed just before the end device (Streaming bridge or computer).

[TomJ]

9 minutes ago, plissken said:

 

You are getting push back because you aren't willing to admit that expectation bias may be at work. I did a blind test with someone that had a cryo-treated, Telegartner terminated, cable, vs $12 cable from Amazon using their Tidal account. What the cable owner didn't realize is that Tidal will cache entire tracks. He had 300Mbit internet service. He put on his favorite piano track that was ~11 minutes long. I waited 30 seconds and pulled the cable out and showed him that the entire song was stored locally and that you could even pause/play/ff/rr.  Many have no idea how data works.

 

As I said already, its not about the data integrity and jitter of ethernet signal, but about noise.

But this discussion is senseless. 

[TomJ]

11 minutes ago, recepky said:

by the way I have plenty of belden cat 7 cable awg 23 I will make another cable from it

Are the twisted pairs bonded?

[TomJ]

5 minutes ago, plissken said:

 

I have tried a few permutations:

 

EtherRegen vs my Aruba 2530, a $330 WW Starlight 12' Ethernet Cable vs ~300' generic Hypertek Cat5E.  I even posted a view tracks captured to my ADC where I plugged and unplugged the network cable during playback, posted links to the files at AS and no one could pick when any of the changes were made. Not the # of changes or when the change was made

So why are you using a 200$ switch, when a 20$ Netgear will also do the job?

[TomJ]

10 hours ago, recepky said:

no each twisted pair is seperate from each other. 4 pairs. each pair has own shield.

Are the single wires of one pair glued together?

[TomJ]

3 hours ago, recepky said:

This is the cable I have.

 

There are bonded Belden cable and unbonded Belden Cable. So not all Belden cable are use bonded twisted pairs.

[TomJ]

1 hour ago, R1200CL said:

Go fiber, problems solved. Well almost 😀

As I dont have and dont know an endpoint (streamer) with fiber interface, you need a fiber to ethernet converter - here we are again....