Thanks for the further explanations it helps me identify potential bottlenecks and areas I’d not thought to consider. I will read through the information at the Belden site you’ve supplied. Looks like I have some more homework to do.
Thanks, again,
Joe
This review shows how much short length ethernet cables can change sound. My ethernet journey started after reading his cable comparison and reviews and trying cables. https://audiobacon.net/2017/07/09/sotm-iso-cat6-special-edition-the-flavors-of-audiophile-ethernet/amp/
Galen stating we are likely listening to connectors versus cable probably holds weight. Maybe why all the tested and certified cables I bought and tried in listening still sound different. Despite everything else in digital stream is constant as is volume settings and songs when I have done swaps.
Though a lot of members here have agreed at times at which cables we preferred. I found my tastes changed after I made other system changes. But the differences in ethernet cables can be startling at first. Eliminate other sources of digital streaming noise it becomes a refinement tuning exercise. But the primary cables you initially notice and like seem to be keepers when going back and fine tuning. The rest become as appealing as a pile of cold spaghetti that dries out after the steam has escaped and sound hard or instantly were determined to be overcooked and not even close to el dente. Just a bunch of mush though the sauce served up was the same.
Thank you, Vmax. I will check out the link you’ve supplied and your experiences are a good insight.
Hi Galen,
thanks for your infos.
As I life in Germany, it is really hard to find these cables types in a store.
I found this, but I am not sure, if it is comparable with 1700A: Belden 7988R
As I understand, that for shorter cable a larger diameter of the copper is not is not necessarily an advantage AWG24 could maybe be a good choice for my 1m and 10m cable distances.
Will there be a difference in using PE or FEP dielectric of CMP cable for the signal quality?
I dont want to strip the wire down to copper. I just need the single twisted pair, as I mentioned above, that a 2 x 2 cable for 100BaseT with loose twisted pairs without shielding and outer isolation sound best to me - may be imagination or not, but it is how it is.
Regarding my measuring efforts, I just want to know what happens with the signals, when sound changes happen - be it with the cable or also with a LAN isolator/filters (which is not only a galvanic isolator by transormers but also contains common mode choke), which I currently experimenting with.
By only using 2 twisted pairs without outer isolation for 100BaseT the sound is the most detailed and most open for me - but not “unmusical” for me.
I had discussions in other forums with guys who claim that this is EMI interference due to the lack of shielding.
But these guys are the same which DIY shielded CAT cable, and with all the hints of Galen regarding shielded cable design, I find it hard to believe that they have a lot of knowledge about CAT cable.
Regarding shielding vs. unshielded cable with this in mind, I will not trust any so-called audiophile boutique cable that is shielded any more.
The LAN isolators are usually connected to the end device with a very short cable. The effect on the sound is greater, the shorter the cable is. Can it be that the sound change actually comes from a signal deterioration due to reflections?
Most isolators do not use any termination via the center taps a la Bob Smith.
But as LAN Isolators are ethernet magnetics, one have the same situation on the circuit board of the end device, where ethernet magnetics are used to isolation integrated in the jack or as discreet chip. But how are reflections handled there? (Signals from the PHY to magnetics)
Tom,
7988R isn’t an Ethernet cable but a RGB low skew video design.
Belden is sold all over the world for decades now, it should be pretty easy to find…heck, even Internet order!
When we remove the pairs from the jacket of a four pair cable and separate the pairs the IMPEDANCE goes UP and matches the load worst possible situation. RL reflections get way worse. This sounds better?
The ground changes. A single pair in the air is about 115-127 ohms for UTP and STP respectively. So for Ethernet performance, don’t do that.
Find an industrial 2 PR UTP BONDED to try, this will be close to 100-ohms if you must remove the outer jacket as it is re-designed to be 100-ohms with 2PR.
CMR verses CMP will have a slight difference in attenuation and lower dielectric constant but not enough to crow about, especially in short pieces. All patch are oloefin (PP or PE) as the fire safety in plenum spaces isn’t necessary with a patch.
24 AWG is fine for short reach patch type runs. You are correct there.
Try;
7932A
Best,
Galen Gareis
Hi Galen,
many thanks for this!
As a professionell in CAT cable design and mastermind of some of the best Belden CAT cables it must be quiet strange, that there are coming some audiophooles around, hacking CAT cables and claim there is a sonically difference. 
It’s really great that you don’t just hang up, but share your knowledge!
Many thanks!
This is the cable I am talking about, which is unmatched by any other cable regarding sound yet:
For me it is a mystery yet, what the reason for the sonically changes is.
Maybe it is the wrong impedance (115-127 ohms)?
Less NEXT, because greater (but random) distance of the twisted pairs and no outer isolator?
Less irritation and capacity of the two removed, for 100BaseT unused, pairs?
What is going on with the signals? (This I want to analyze with measurements to set the measurements in relation to the sonic changes)
So If I use the twisted pairs of Datatuff 5e in this way, will there also be a change in impedance, as the distance will be greater (but random)?
Should I look for single 100 Ohm unshielded twisted pairs?
I have already done a cable with single 100 Ohm twisted pair cable from Gore:
But sound is not on the same level of the stripped cable - but close.
I just thought to remove the outer isolation and the shield, but then, if I am not wrong, Impedance will change.
Which box of pandore have I opened?
I thought experiments with 4 pairs of wires were quite manageable - now I’m looking for 1Ghz scopes with eye pattern function …
Unfortunately, because I’m an explorer, I can’t go back. . . .
P.S.:
You can tell your former colleagues at Belden that they have to work on sales in Germany.
I cannot find any of the cables you have named in any German online shop, but have to order them in the US - usually with a minimum order quantity of over 100 meters.
Regards,
Tom
Hi Galen,
Just called with the Belden EMEA sales team and they try to send me a short 2m sample of the Datatuff cable, because it is not available in short length via a distributor.
Regards,
Eric
I just replaced one of my 3 ww platinum Ethernet by a starlight, to see if it makes a difference. Oh yes, it sure does! The sound was veiled and muffled. And this is the least critical Ethernet cable, the one from the wall to the router.
So yes, Ethernet cables matter and sound different, at least to my ears and in my system.
Maybe someone here with an answer . . .
When I use a standard CAT6 cable for 100BaseT, what to do with the 2 unused twisted pairs? Background: I have a 100BaseT switch with only connections for 2 pairs.
Not connecting on both sides?
Only one side connecting and terminating with 75Ohms and grounding?
Kind regards,
Tom
in 100 base T days the cat 5 cables all had 4 pairs. just ignore them.
only difference with 100 is you have to use a crossover cable for connecting two devices without a switch in between.
Just thought a bit about the constellation.
Came to conclusions, that best would be to connect the one end to GND like the shield so that EMI interference have a path and let the other end open unconnected.
Regards,
Tom
Hi,
short update regarding my cable journey.
Just ordered Belden Datatuff 7932A ( Thanks to @rower30 for recommendation) in the US with a shipping forwarder, as there is no chance to get this cable per feet in Germany.
Will report outcome.
Regards,
Tom
To be clear, this is a 2 x 2 cable for 100BaseT only. But who needs more for music?
I ordered here: Belden 7932A Cable 24 AWG 2 Pairs Category 5e DataTuff Twisted Pair So
And forward service is myus.com
Tom,
Yes, take advantage of the sample order sales process. If you don’t need too long a cable to do your test…it’s free! Longer test stuff will of course need a 328 footer for TIA accuracy.
Sorry for being late, we were on vacation to Mesa Vera Colorado all week. 3,500 miles of driving around!
Best,
Galen
Hi Galen,
hope you had a good time in Mesa Vera!
There are really other important things in life besides cables - but not much 
Regards,
Tom
I think I’m going crazy.
I was just told that the minimum order quantity for the Belden 7932A is 500 feet.
Unfortunately, I only use it for 50 feet and have to cancel the order.
At the moment I’m a little at a loss as to where to get the cable.
I also haven’t received any feedback from Belden EMEA sales on the samples.
I think I can forget about the Belden 7932A cable.
There is no shop that wants to sell me the cable in humane quantities.
Or does someone have a source where you can buy something like this in low quantities?
I found a place who sells 100 feet for around $130.
https://www.awcwire.com/product/7932a
Thank you waymanchen11.
Also try ALPHA wire and cable. This company specializes in making one-of shorter assemblies and cable lengths.
Belden makes put-ups for large IT networks and to them, shorter reels is inhuman (too many empty reels and waste). It’s all in your end use customer’s needs but off-shoots grow in to the cracks you leave open, thus ALPHA wire and cable.
I’d call them as they may assign their SKU P/N’s to cable stock.
Best,
Galen Gareis
Hi Galen, reading your comment I changed my 0.5M ethernet cable to a 1M. They are the same brand and are connected to a Lan filter.
Wow, the sound has improved in details, soundstage and just sounded fuller. I always thought ethernet cable needs to be shorter. But you are right on the min. length. Thanks!
The Ethernet reference is 90 meters max on the permanent link (what’s inside your wall) and 10 meters total patch length (5 meters on each end max).
The shorter the horizontal the better the Channel (patch plus permanent link) will test as it reduces RL reflections through the connectors which are the worst case components. Always error a little longer on the patch as we use such short cables in home Ethernet systems and with often suspect patch cable quality that are far, far shorter than 90 meters. I suggest 1 meter to keep it reasonable. DO NOT fold or coil it up!! If you have you have already ruined the cable electrical with periodic RL reflections at the bends. Gently unwind a COILED, not folded, cable from the bag and route it so it isn’t bent. Yes, most patch are already FOLDED and BENT from the big box stores.
X= plug and/or jack
X<—5 MTR patch--------->XX<---------90 MTR horizontal------------>XX<----5 MTR patch------->X
The specs assume you have PASSING patch assemblies!!! If you have good patch short lengths SHOULD be fine as they take into account to connector RL properties. Less than 3 feet can cause issues with marginal cable and connectivity but it SHOULD work if it is inside the spec parameters!
The maximum distance between the telecommunications outlet and the horizontal cross connect shall be no more than 90 meters. The maximum length of all patch cords and jumpers in the telecommunications closet shall be no more than 5 meters, and the total length of all patch cords both in the telecommunications closet and at the work area shall be no more than 5 meters.
Best,
Galen
