Ethernet Cables and Sound

Good offer and I´m interested in results too,no matter how it ends up going. Maybe people could send some other cables too for testing? That of course if Galen is interested in such nonsense :slight_smile:

2 Likes

Vmax,

Easy, have Blue Jeans test it for PATCH compliance. Be glad to talk to Kurt Denke about it. Was your cable compliance tested? All Blue Jeans cable is tested for Ethernet. They are not tested to all NIC card I/O and RF rejection properties in “audio” applications. That block belongs to the equipment vendor, not the cable. We would be looking at the CABLE, not the I/O variations in equipment so there would be no “finding” about cable doing that. We already know the cable will move zero BER over such short distances. What part of the analog I/O block is RF sensitive and why? Is noise lift REALLY measurable at the signal ground point?

We’d test your assembly just to prove a point…and no, we can’t test everyone’s cable for free. As a customer you can adopt anything you want, we, as a manufacturer’s won’t.

The quality of a Ethernet cable starts at the singles and Inakustik doesn’t have a million plus dollar singles extruder good enough to make Ethernet quality singles. Sure, they can have the cheaper braiders and other machines but a world class Ethernet singles line is expensive and can’t be paid for too easily making specialty products as it has to run LLOONNGG runs ALL THE TIME to stay in spec range. Turn it off and it is even DAYS to get it adjusted again. Either way, the cable needs to be UL verified for Ethernet compliance or tested after assembly if a patch cable.

Best,
Galen Gareis

1 Like

Galen,
Could I trouble you please for a specific link for the specific cable you use?

Been looking and having some difficulty finding one.
Thanks. :slight_smile:

The BJC cable came with data. Yes it showed and how it measured and pass. Certicable is another provider I tried and provided test data. The Inakustik cable provide a hand sign document certifying it was inspected and tested and met all requirements but did not provide specific parameters but all the compliance seals on the box.

Thw whole point of starting this thread was based on sound comparisons. I too thought in the beginning a well measured and documented cable would be the holy grail and it could get no better. But low and behold at short lengths the same manufacturer certified cables sound the same per manufacturer samples i purchased. But swap them between manufacturers is where the differences in sound are noticed. even when all certify performance.

I respect you can’t bite the hand that feeds you or be expected to like someone’s else’s cooking better in a listening test that is a cable you did not design. I totally respect that. I was just wanting to prove they sound different more than just send to have measured. If I designed and built cars I’d drive my own and like it best.

That said listening is subjective to us all. My one BJC cabe connects my P20 to internet the other sits unused. I had wanted to like it better for music streaming. Maybe when I change equipment again the preference will change.

Thanks for helping us understand the design variables and what they might mean.

Worth noting…!

https://www.bluejeanscable.com/store/data-cables/index.htm

Try the CAT6 BONDED. See the Length and price calculator down the page.

Best,
Galen

1 Like

Thank You. :+1:

I have a bunch of these BJC CAT6 cables. Awesome service, product, and support.

Order placed in assorted lengths! Thanks for all the information.

Well I just ordered a full loom of them too. :grin:

Galen

I like to try and understand your posts, I think I got close to understanding this one.

My digital audio only uses ethernet cables (no RCA, XLR, usb etc.) and a couple of years ago I bought three of the BJC6a bonded. I still use them. I replaced the main cable installed from modem to server, put in by my broadband supplier, that was CAT5e running over a roof. I bought Audioquest Pearl CAT6a simply because I needed a 25m run the next day and there is a local supplier near me in London, UK who can do the terminations properly and test them.

I am relocating my audio in the next few months, rebuilding half my house, and the long and the short of it is that the BJC6a bonded is all that I will ever need? I am running a 15m cable this time and have some shorter ones to replace as well.

Is this correct?

My grounding is assisted by a small number of components and a Shunyata Hydra Alpha and Shunyata cables. I was also thinking of running some Belden 19364 mains cable back to the consumer unit, which is only about 5m away. I understand that cable has a drain cable that returns noise to the ground.

Is that a worthwhile exercise?

I use Sky Q, one of the three main broadband suppliers in the UK (the others being BT and Virgin - the latter have their own fibre network). They supply a standard dual band modem. I thought it was fine, but this week bought off eBay a Netgear N7000 modem/router, mainly because I want to install the Netgear Nighthawk Mesh system. The Netgear N7000 loads videos from Netflix about 3 or 4 times faster than the supplied Sky modem/router. I was just amazed.

I will be anxious if you like them as well as the Inakustik’s or have similar differences that I experienced to see if there are other factors in the streaming chain. For grins I will put my BJC Ethernet back in and listen this weekend and describe my results. Both the streamer and rendering links. I don’t have a EtherRegen like you do.

Hi Steven,

A “DRAIN WIRE” on a properly shielded cable is a little off from what it is doing. It is really a “neutral” wire. If we have a perfect ground, the resistance is the EXACT same everywhere it is supposed to be at ground. We can’t have current flowing from A to B if A and B are the “same” resistance, right! That’s what we’re after in a ground, no differences to start a too big ground loop, and it is a current looping per Kirkhoff’s current law in the “ground” that isn’t really at ground.

Bad grounds are tiny, or sometimes not so tiny, little circuits. Even Ethernet can’t remove a bad ground between two points. It CAN remove ingress on both pairs. It isn’t magic. A ground issue with a shielded cable may be worse than using UTP. Shields attenuate the noise but a bad ground can inductively couples MORE noise.

Few systems need shielded Ethernet cable, even systems with good grounds. Many a large CAT7/8 system is hung-up with ground noise issues that need to be solved first. CAT7/8 is more about the BANDWIDTH capability than the a “shield” per say. We get the bandwidth with two things; less attenuation of the signal level and pair-to-pair shielding is 40 dB better on average. Both increase the ACR. If you need the bandwidth (10G) CAT7/8 has been capable as the noise isolation is INTERNAL (NEXT) and EXTERNAL (ANEXT) both with individual shields.

The talk about the sound is more a DA and AD noise filtering issue than what the cable is responsible for. With a system with poor grounding, we may see more noise in the ground that can modulate the signal that is referenced to a point that is not supposed to be moving above ZERO volts, ever. If it have a bad ground loop we can change what we hear.

ALL systems aren’t perfect as all wire has DCR. We do indeed have some ground that is NOT at the lowest ground potential. This factual argument can be used to spread FUD on ADDING any unwanted additional current to the ground (galvanic isolation). Good systems are well below where this is going to be heard, though, but that’s the basic premise of what may be happening that is based on at least real stuff we can then mitigate properly.

No system can be the same as all our grounds aren’t the same. Every cable will act differently based on how ground noise is managed. UTP cable prevents ground noise between the NIC cards. It doesn’t use a ground with the balanced signal, it FLOATS the ground between two differential points like I explained earlier. This is a nice feature IF, IF, IF we can keep the ACR in spec for a worst case LENGTH. Most modern Ethernet cable is able to keep NEXT (near end coupling) and ELFEXT (far end coupling) between pairs well in check.

I would use UTP first as few home systems are over 100 meter long (90 meter peramaent link and 10 meter total patch length added on). Shielded solutions are really a pro’s only thing as so much has to be maintained to not get WORSE than UTP! Lift a ground somewhere and we inject more noise into the system. Not good. Use the wrong shield termination method and we radiate RF noise at the connector end right at the NIC card, not good. You need a 360 degree shield ground at higher frequencies. The ground impedance keeps going UPO as frequency goes UP, so the ground resistance needs to be as low as you can get it (lots of contact surface area).

Use good CAT6 cable and you’re set for everything up to 1.0 Gig Ethernet. More than fast enough to stream DSD1024.

I have information on all this if you are interested in how all this works and why. It isn’t too weird at all, just not common knowledge.


On your speakers, did you keep your QUAD electrostatic panels, or go dynamic driver? I use CLX and really like them. I like them so much that and my second set of speakers are an electrostatic panel (CWT 1000-40 T+A design) hybrid design that sound great. I love the sound of electrostatic panel midrange and treble. The dynamic drivers add the upper bass/ lower midrange energy smaller panels aren’t good at. Electrostatic speakers big enough to mitigate the upper bass “drive” won’t fit in my room.

Best,
Galen

1 Like

Galen, thanks for your reply and I will study the substantive part later in the day. On the cables, I will place an order. I’ve not got past 24/192. DSD does not rock my boat, and I’ve been reverting increasingly to analogue. Last week the UK government said there was a greater chance of dying of dementia than COVID, so I’ve spent the last week worrying about the increasingly aberrant processing ability of my brain, if there ever was any.

So this is what Quad does to me. The last audio show I went to, 4 years ago, I mainly went to see a photo exhibition and talk by Ross Halfin, and then check out the latest gear. I went to a couple of rooms, then went into Quad. Spent 2 hours listening to 2912’s and left. That was it.

It was pointless asking the wife, so when the kids had left home and a pair of ESL63 came up, having had a full factory rebuild 6 months earlier, I bought them. The wife went mad - nothing was said, I just got a deadly stare. A month later the wife posted an eviction notice and I old them to a lovely guy called Raphael Todes. He’s a professional violinist (The Allegri Quartet, before that the Schidlof Quartet) and does reviews for HiFi Plus. He seems to be using them as one of his reference speakers. I don’t blame him.

Harbeth are the best value for money speakers I know and had then for 6 years. I don’t like their looks and my wife, who hated the sight of them, put up with them as one of marriage’s compromises. Alan Shaw of Harbeth has said he considers one of his design objectives is to get as close as possible to Quad ESL as is possible from a box at a sensible price.

The big difference is that Quad ESL image tremendously well, whereas Harbeth do not. That is not a flaw, it’s a design objective, Alan Shaw thinks of sound as a wave, and as I listen a lot at opera and ballet where the orchestra is hidden, I would agree - but you need good imaging for jazz.

When it came to the new room (a music/reading room with no TV) I just let her chose and she chose Wilson Sabrina in Dark Titanium. What a girl! They have a Harbeth midrange and a bit more at both ends, but not took much, and imaging almost on a par with the Quad ESL.

I thought one day I might be allowed a pair of 2805 (considered perhaps the best all round), but it’s not going to happen and I’e got over it.

1 Like

Yes, we have the IMAGE or not to IMAGE question. Large stuff clearly doesn’t image where smaller stuff does. Since I can’t really get SUPERTRAMP into my basement, I chose to IMAGE and get Peter Paul and Mary sounding good. Panel type speakers do imaging REAL well, I agree. But newer dynamic driver speakers are also getting pretty good.

You also have the sweet spot. Panel type speakers shrink that up some over dynamic drivers. But, I listen on my own so no problem for me.

My B&W 801’s, circa 1979 (I bought them after the Chicago CES show as demo speakers), are still playing in my mom’s house! Other than the tweeter spider wires corrosion (moisture in the air permeates the TPO coating over them) killing the tweeters and getting new ones the Brits build good stuff for audio, no complaints there.

COVID is weird. One half on one percent fatality isn’t going to kill the human race but that low percentage fatality number seems higher if it is you are a close one to you.

Best,
Galen

1 Like

The last time I listened to Supertramp was with 20,000 other people at the O2 Arena. No imaging but a lot of fun.

I use other products for off-axis listening, and may install a system that has not yet been announced - but will be very shortly.

There is quite a lot of good audio from the UK, also plenty from Denmark, Germany and many other places. Cable seems to be a very USA thing. I can only think of Chord and QED over here. QED have been going forever, 50+ years. They have published two reports on cable design here:
https://www.qed.co.uk/qed-story

2 Likes

Wait, what’s your name again? :thinking:

You know if you’ve got dementia if you can remember small details from decades ago but not what you had for breakfast. So if you remember the details of a long gone reel-to-reel tape machine but can’t remember the last time you saw a streamer, you’re in real trouble.

As for my name, it escapes me.

2 Likes

Hi Galen, I also have a pair of ML CLX speakers. I’m sure you are using Iconoclast speaker cable I’m curious of what conductor are you using? Also what are you using for amplification? Thanks, Paul

1 Like

Paul,

I used the SPTPC or TPC with them, usually the SPTPC stayed hooked up with a MOON W-8 amplifier. Dead stable into the CLX.

I used two BF212 subs but DID NOT use the built-in filter slope downloaded per USB, or the DSP! I set the subs to MANUAL and set the level to 6.0 and the 25 Hz to -4dB and the cross-over to 55 Hz in my room.

Use the Perfect bass kit to look at your room. Mine is 39 feet long and I have a 25-30 Hz bass peak of 8 dB. I CUT the 25 Hz 4-6 dB as the 25 Hz lump of bass kills the speakers midrange.

Best,
Galen

2 Likes