'Audiophile' Ethernet Cables.

[dnoyeB]

Keeping noise out of the audio electronics is the job of the electronics itself. The cable is not going to help with that beyond the fundamentals of its design. All Ethernet cable is fundamentally the same design (4-twisted pair with or without shielding) or it wouldn't be Ethernet. I suppose I did somewhat contradict myself. But in fairness, the benefit of an isolated connection will only be to the audio equipment manufacturer. With an isolated interconnection technology, they can be much less careful about their connection. e.g. no electronics designer worries about ground loop noise if they have a TOSLINK connection.

 

To 'realhifi' point:

Can noise affect the data of a lesser made cable? Sure. But once it crosses a certain threshold it will no longer qualify as CAT5 for example. So long as the cable satisfies the requirements of CAT5, and you are not in any special environment requiring special handling, a higher quality cable will provide absolutely no benefit..

 

You can:

1. use a shielded cable

2. use a CAT 6/6E cable.

3. use a dedicated network path.

4. Choose a slower sync rate (100Mb instead of 1000Mb)

 

I wouldn't do any of these things unless I had evidence of packet loss. I personally have experienced packet loss from bad cabling. Cabling I crimped the connectors on. Worked great for years until I tried to use it on Gb network. Then I got 5-10% packet loss. Figured out my crimps were not good enough for Gb. My cables would have never qualified under CAT5. I now purchase most of my cables...

[junker]

Maybe someone would want to buy one of these and give us an objective review:

 

[Paul R]

There is more to network transmissions than packet loss guys - my favorite buggaboo is packet queue delay, a form of latency that drives me nuts sometimes. Each packet transfers at lightning speed, but it lollygags waiting to be transmitted for who knows how many milliseconds? That can make even the fastest network slower than a dog not wanting to go out in the rain...

 

More, that kind of latency affects all kinds of things up in the application layers.

 

So, yeah, while I basically agree with a great deal of what you saying, there is more to the story. A lot more actually. Just trying running a hgh end player off a NAS device, and see what perfect packet transfers buy you. Especially as the device gets more and more used.

 

-Paul

 

 

Keeping noise out of the audio electronics is the job of the electronics itself. The cable is not going to help with that beyond the fundamentals of its design. All Ethernet cable is fundamentally the same design (4-twisted pair with or without shielding) or it wouldn't be Ethernet. I suppose I did somewhat contradict myself. But in fairness, the benefit of an isolated connection will only be to the audio equipment manufacturer. With an isolated interconnection technology, they can be much less careful about their connection. e.g. no electronics designer worries about ground loop noise if they have a TOSLINK connection.

 

To 'realhifi' point:

Can noise affect the data of a lesser made cable? Sure. But once it crosses a certain threshold it will no longer qualify as CAT5 for example. So long as the cable satisfies the requirements of CAT5, and you are not in any special environment requiring special handling, a higher quality cable will provide absolutely no benefit..

 

You can:

1. use a shielded cable

2. use a CAT 6/6E cable.

3. use a dedicated network path.

4. Choose a slower sync rate (100Mb instead of 1000Mb)

 

I wouldn't do any of these things unless I had evidence of packet loss. I personally have experienced packet loss from bad cabling. Cabling I crimped the connectors on. Worked great for years until I tried to use it on Gb network. Then I got 5-10% packet loss. Figured out my crimps were not good enough for Gb. My cables would have never qualified under CAT5. I now purchase most of my cables...

[Jud]

Snake oil. Advantages are less resent packets on long runs, or in extreme RF environments (i.e. server farms, etc).

 

Anyhing work it's salt will re-clock the digital signal that is pulled from a buffer with jitter dependent of the re-clocking device (i.e. ESS Sabre, etc.)

 

Creative terminology! :-)

 

I'm perfectly willing to listen and learn, that's one of the primary reasons I'm here. But I can't learn if you don't provide information. So please provide the following information, if you kindly would:

 

- Explain why barrows is incorrect in his explanation of how low-noise power supplies are important to the DAC he built using the ESS SABRE chip:

 

OK, examples with my DAC: The ESS 9018 DAC chip uses a local masterclock, and 4 local, separate power supplies: two supplies which power the analog outputs of the chip, and two supplies which power the digital sections of the chip, additionally, a fifth local supply is needed to power the local masterclock. Most designs will implement these supplies with local, IC chip based voltage regulators. These regulators are good, but not great. The Twisted Pear Audio Buffalo DAC uses five individual, discrete, shunt regulators for these supplies. Shunt regulators are lower in noise, and much lower in output impedance (across a broader bandwidth) than ordinary IC regulators. Using the shunt regulators results in much improved sonic performance, this is probably mostly due to their low output impedance, which stops the individual supplies from being modulated by each other (better isolation from one supply to another).

In my DAC I then go one step further-the DAC board (with its 5 onboard shunt regulators) are fed with another shunt regulator, acting as a pre-reg. This pre-reg is fed from its own transformer winding to isolate it entirely from the other power supplies (which are three other shunts, one for the USB receiver, and two for the +/- 15 VDC analog output stgae supplies). TPA also provides a very, very good masterclock on the Buffalo DAC (Crystek CCHD Series), this clock is much better than the clocks used on most commercial DACs.

Additionally, The Twisted Pear Audio offers two output (I/V) stages designed specifically for the Buffalo DAC. I use the Legato II (now there is a III version) which is a discrete bipolar input MOSFET output design with no global feedback. This output stage presents a less than 1 ohm load to the ESS 9018, and this low loading allows the ESS output to operate in current mode, resulting in the best possible performance.

 

The above illustrates that there is a lot more to making a DAC perform really well than chip selection. I am not an engineer, and I am sure if any of them are reading this they probably find my examples overly simplistic, and that they could probably point out many other areas of implementation which matter to the performance.

 

- Explain how the rest of the system to which an Ethernet cable is electrically connected cannot be affected by noise involving the Ethernet cable, but all other digital, analog and power cables connected to the system can affect it.

 

As I said, I'm ready to listen, but if you want to call BS you'll have to back it up. If you can't, that doesn't necessarily mean "audiophile" Ethernet cables aren't BS, just that you haven't conclusively demonstrated it.

[4est]

Incidentally, there is one real advantage of fiber over Ethernet. No conducted or radiated noise is picked up with the laser. I think most of you probably have experienced this during the TOSLINK vs. RCA digital data days...

 

Remember that at the end of the day, the fiber needs to be turned back into electricity and that requires a power supply and what ever noise and other issues it may have.

[junker]

Look you guys have had a great technical overview primer of the entire TCP/IP stack and it's packet-based error control provision at a number of layers.

 

With regards to another point made about latency - that is not due to the cable. If there is another HW issue than that is something else entirely.

 

I think that at this point no amount of data or technical information is going to convince anyone one way or another. We may just have to agree to disagree.

 

And even for arguments sake let's say there is some difference when specialized cables that (hopefully) still meet the Cat 5e or 6 spec (and again Cat 6 has extra shielding) then you have to ask what is the value proposition and could that money be used elsewhere more effectively.

 

Cat 6 Cable Info

 

For example, for $5k I would rather have an Berkeley Alpha 2 DAC than a fancy 5m run of silver network cable. It just seems absurd. Where does the madness end?

[4est]

For example, for $5k I would rather have an Berkeley Alpha 2 DAC than a fancy 5m run of silver network cable. It just seems absurd. Where does the madness end?

 

Was there someone recommending $1000/m cable? I must have missed that...

[junker]

Precisely. This is my slippery-slope ad absurdum argument that at what criteria do you consider your cabling sufficient?

 

Is it $1, $10, $5000? And for what performance gain?

 

Where can even $100 (8m of the cheapest AQ Forest) be used better? Oh, let's say an entire DAC from Schiit.

 

So, you have to admit that the value proposition fall apart? For me it's anything above quality Cat 6 cable.

 

What's the saying? I've got a bridge in New York that I want to sell you. But hey, if there are enough customers willing to pay then I don't hold it against them to make a buck. I do, however, think that it is a waste of money...

[dnoyeB]

...

- Explain how the rest of the system to which an Ethernet cable is electrically connected cannot be affected by noise involving the Ethernet cable, but all other digital, analog and power cables connected to the system can affect it.

 

...

 

I never said what I think you are suggesting I said. Ethernet cabling is standards based, other cabling is not. Once you have surpassed the requirements of CAT5, any additional noise shielding that is added to the cable is negligible. That is, unless you have an inordinately high noise environment. Remove the CAT qualification, and all bets are off.

 

Exactly what problem are 'audiophile' Ethernet cables trying to address?

[4est]

Precisely. This is my slippery-slope ad absurdum argument that at what criteria do you consider your cabling sufficient?

 

Is it $1, $10, $5000? And for what performance gain?

 

Where can even $100 (8m of the cheapest AQ Forest) be used better? Oh, let's say an entire DAC from Schiit.

 

So, you have to admit that the value proposition fall apart? For me it's anything above quality Cat 6 cable.

 

What's the saying? I've got a bridge in New York that I want to sell you. But hey, if there are enough customers willing to pay then I don't hold it against them to make a buck. I do, however, think that it is a waste of money...

 

Well, I do not have a horse in this race, but $100 for 8m seems much more reasonable than someone dumping 5K for 5m. Considering the net value of some systems, even if they do not notice a difference between Cat 6 and AQ Forest it is not much. I would assume anyone using these would using a streaming or UPnP DAC and the Modi is not applicable.

[Jud]

I never said what I think you are suggesting I said. Ethernet cabling is standards based, other cabling is not. Once you have surpassed the requirements of CAT5, any additional noise shielding that is added to the cable is negligible. That is, unless you have an inordinately high noise environment. Remove the CAT qualification, and all bets are off.

 

Exactly what problem are 'audiophile' Ethernet cables trying to address?

 

That was in response to Junker's posts, nothing you said.

 

BTW, before going further - I agree with Junker generally about the value proposition. Maybe I'd get one of the AQ Forest line for under $30 just out of curiosity, but I agree sums larger than that are better spent elsewhere, at least in the part of the market I'm playing in.

 

One last thought - when you said it was the job of the electronics to deal with noise, that may be the only time I've heard someone recommend a pound of cure (dealing with noise at the components) rather than an ounce of prevention (keeping noise from getting to the components in the first place). Or to put it another way, noise is something perhaps best attacked in the context of the system as a whole.

[junker]

I will concede that you have both Spectral and Vandersteens so I know you are careful and thoughtful with your purchases. Respect!

[Jud]

I will concede that you have both Spectral and Vandersteens so I know you are careful and thoughtful with your purchases. Respect!

 

Noticed your Vandys & A+ as well.

[realhifi]

Run, run away from "audiophile" ethernet cables, or any companies that promote their products are such. These types of charlatans are dead to me.

 

Sounds a bit harsh to me but hey, it's tough out there!

[realhifi]

Once you have surpassed the requirements of CAT5, any additional noise shielding that is added to the cable is negligible. That is, unless you have an inordinately high noise environment.

 

You mean like Led Zeppelin at concert levels? Or an orchestra at full throat? Not all music is quiet and tinkly.

[dnoyeB]

deleted

[Audio_ELF]

You mean like Led Zeppelin at concert levels? Or an orchestra at full throat? Not all music is quiet and tinkly.

Sorry... You mean some music IS quiet and tinkly??

 

Eloise

[Jud]

Sorry... You mean some music IS quiet and tinkly??

 

Eloise

 

I'm sure Kylie had some quiet tinkly bits in there for all the little girls.

 

/me runs and hides!

[realhifi]

Sorry... You mean some music IS quiet and tinkly??

 

Eloise

 

*s* Just my lousy description of some of the "audiophile" music that is (has) been favored for manufactures to show off products.

Speaking mainly about breathy female vocals recorded up close, soft acoustic guitars; that sort of thing.

[junker]

Think he was referring to RF/EMI noise where shorter runs, twisting, and shielding would help.

[VandyMan]

My personal opinion is that if an audio-cable company wants to sell expensive Ethernet cable, they better have the equipment to prove that it handles RF/EMI better than the less expensive cables used in intensive settings like data centers and they also need to make detailed specs available. They always say: After X years of prototyping and doing listening tests, this wondrous cable has been created. Maybe this is ok for analog, but I don't accept this approach for any digital, but especially packet based.

[Jud]

My personal opinion is that if an audio-cable company wants to sell expensive Ethernet cable, they better have the equipment to prove that it handles RF/EMI better than the less expensive cables used in intensive settings like data centers and they also need to make detailed specs available. They always say: After X years of prototyping and doing listening tests, this wondrous cable has been created. Maybe this is ok for analog, but I don't accept this approach for any digital, but especially packet based.

 

My hat's off to anyone with Vandersteens, let alone 5a, but I've got to mention that by focusing on digital transmission we keep losing sight of something very important, which is that electrical noise through any cable, analog or digital, has effects on the system irrespective of the content the cable is carrying.

 

Let's do a little thought experiment about this. Let's say no data is being sent through the Ethernet cable at the moment. And let's say there's noise running through the cable (from ground, EMI, RFI, whatever). It's not enough to disrupt the very robust packet-based Ethernet transmissions when they do occur, but right now they're not occurring, just the noise. And say that noise makes it through ground to the analog side of the system. Just because this cable can carry Ethernet packet transmissions doesn't negate the fact that as an electrical connection it can also carry noise into the system. When the Ethernet data transmissions resume, they won't eliminate the noise, they'll both simply be transmitted through the same cable.

 

So that's what I think we need to remain conscious of. Just because the intended purpose of an electrical connection is to carry packet-based data doesn't magically confer immunity from various sources of noise that any electrical connection may be subject to. The fact that the noise isn't enough to disrupt the data transmissions (and remember, as many commenters have mentioned, packet-based data transmissions are robust enough that they will make it through even relatively noisy connections) doesn't mean the rest of your audio system magically becomes immune to noise. If present, noise can affect the system in a variety of ways having nothing at all to do with whether the Ethernet data transmission is successful.

[dnoyeB]

But Jud...

 

I can't think of a noise scenario where there is noise but it is not affecting the signal carried by the cable yet that same noise can be mitigated by improving the cable itself. e.g. ground loop noise is not going away no matter how fancy your cable.

[VandyMan]

Jud - I agree except that I think that unless the cable has a major design flaw, it is up to the DAC/receiver to handle any electrical noise from the connection. Additionally, I think that RF/EMI noise, like jitter, is being tossed about in the industry to justify high prices with very little proof that the cable actually reduce noise or that the noise effects the performance of well designed DACs/receivers.

 

Hardware engineers correct me if I'm off base, but I'm under the impression that it is far easier to remove/isolate electrical noise from a digital signal without impacting the data than it is from an analog signal. I have $800 speaker cables, but unless I learn something new about digital transmission that changes my mind, I would not consider spending anything like that on a digital cable.

 

One more thing... with packet based audio, I'm fairly sure you need some kind of CPU or DSP in the DAC to convert the packets and handle the buffering. Any CPU or DSP will likely generate far more noise than comes in on a shielded digital cable.

 

BTW, I still have my old 2ce Sigs in my bedroom system. Great speakers.

[VandyMan]

I can't seem to edit the above, but I probably should have said "controller" instead of CPU or DSP. As in an Ethernet or USB controller chip, which is basically a special purpose CPU.