NDS and Ethernet: does distance make a difference?

Cat5e is good for 100m assuming:

Its installed correctly

Not running with powerlines

No kinks or twists

Terminated correctly at either end to a faceplate/patch panel

Exposed runs are proper off the shelf jobs

Typically if its not done correctly you might still get a full functioning connection, but done correctly you can walk about with a smug smile on your face.

A 1 metre cable between boxes will not make it work faster or better.

Chris,theoretically the shorther the patch lead the more propensity for interference. Inpractice this is probably not noticeable. There are some 'green' switches out there that sense the patch lead distance and reduce the switching voltage accordingly, hence the propensity for interference.. And if nothing else reduces your electricity bill ... Slightly..

SiS - Ferrite chokes to prevent the antenna effect of a longer ethernet cable?

Congrats Chris -- hope you enjoy your NDS for many years to come!

ATB.

Hook

Thanks for the feedback, I want to be clear my question is in regards to sound quality.  Is there an audible difference between a long run vs short run of Cat6?

Hook, how long did your NDS take to warm up/ break-in?  

Hi Chris -

I don't think the length of your Ethernet cable will make any difference at all. Mine is connected with a 10' run to a switch (shared by an Apple AE).  The switch, in turn, uses a 50' run to a second switch in a remote closet (where my router, cable modem, NAS and AVR are).  Prior to adding the AE and the local switch, the 50' run went directly from the NDS to the remote switch.  I am using all Cat 5e, and heard no difference between the two setups.

My only concern about the Ethernet cable is the same I have for all the other cables entering the NDS, and that is RF interference from them acting like antennas.  As Simon has pointed out on several occasions, this is easily solved by using ferrite chokes, close to where the cable enters the NDS.  I noticed that the two chokes I added to my Cat 5e cable seemed to place a bit of stress on the NDS's Ethernet port, so I inserted a small piece of foam between the chokes and the Fraim glass to give them a bit of support.

My NDS sounded good after only a couple of hours of running.  There was no obvious harshness like there was with the DAC. But having previously heard a fully burned-in unit, I knew there was more to come. I decided to stream my Isotek burn-in CD 7x24 for the first couple of weeks, so by week 3, it was sounding...superb.  Haven't noticed much change at all since then, but honestly, I haven't felt the need for "critical listening" lately.  Come to think of it, that's probably a very good sign!

ATB.

Hook

Hook, yes to a certain extent but IME the real benefit of ferrite chokes is to reduce common mode RF currents flowing along the cable through to the end device. It's the same principles you use in RF patch leads.

As you say as long as the cable is 100m or less you shouldn't notice much if anything over the length. But if you are less than a few metres in length using one of these green switches could provide additional benefit in reducing the switching voltages on the Ethernet patch lead, it certainly is not going to do any harm...

I didn't find the burn in too onerous for the DAC though it's smoothness and fluidity improved over a few weeks.. It was quite easy to leave it playing 24/7.. In my recent expierience it's DR PSUs and preamps that seem to change quimar significantly after burn in.

Hook, I am interested to hear whether you feel chokes still have an effect on the NDS.. To the same extent as on NDX?.

Simon

My stored music travels 50 metres and internet radio 100 metres on Cat 6 cable. Its terminated at both ends into proper wall plates and I have a switch at each end. I don't use chokes, don't get drop outs and can't hear any degradation in the sound quality at all, in fact it sounds superb.

Mine 20m Cat6 with no degredation - sounds marginally better than the HDX USB on Hi-Res files I think.  G

Yep, there should be no or minimal data loss upto 100m. Chokes have nothing at all to do with dropouts or data loss, they are there to impede common mode RF currents.. 

All,

I am going to slightly go off topic here, so apologies in advance.

I have my Qnap NAS in the study upstairs and it will take a lot of demolition to install an Ethernet connection. I also did not want to use the build in wifi in the NDS for the obvious reasons. So I added a wifi adapter well away from the hifi rack and powered it from a completely separate power socket. The Qnap is connected to the wifi  router in the study and the wifi adapter in the living room picks the wifi signal and connects to the NDS via a 7m cat6 cable. 

I find this works for me just fine. But I don't know if this is as good as a direct Ethernet connection. I know the thought will bug me so I plan to bring down the NAS to the living room and test to see if there is any difference.

ASenna04

The difference will be the general performance of the nas, especialy upload download speeds.

might not make a difference for streaming but copying files to and from it is probably painful.

i have a proliant server running nas4free and get around 50Mb/s up and 85Mb/s down.

but i do use it for a lot of things.

I've been into Naim Audio for a long time and it feels like streaming goes against the conventional wisdom of cabling.  Perhaps I am have a hard time believing that inexpensive Cat6 cable is good enough for high end audio.  Is ethernet cable directional?  Is Naim planning on introducing their own branded cable?  Naca-E or Naca-Cat?  

I just discovered my Ethernet switch is "green"... is this an issue?  

I might suggest that this is because you are drawing a comparison between carrying an analog signal via a current, as opposed to transmitting a series of 1's & 0's, that as long as they arrive on time & in the proper order, with no drop-outs, for the conversion process, should have no bearing on the SQ.

Digital transmission is not carrying what you will hear, per se...you couldn't amplify them.  But analog signals are exactly that which you will hear.

Hi Simon -

I had all the ferrite chokes in place before the NDS arrived, so have only ever listened to it with them attached.  It's too late to try this evening (I was away for the weekend, and only got home a short while ago), but when I have some daylight, I'll pop them off, have a listen, and report back.  Right now, I think I have two each on the Ethernet, HiLine and 3.5mm RCA cables.  

Should I still hear a difference, do you think it would do any good to also add a couple of larger chokes to the two Burndy cables (coming from the 555PS)? Or perhaps the right way to ask the question is:  do all cables of all types have the potential for delivering common mode RF?

The chokes definitely had a clarifying effect on the NDX->DAC, so it will be interesting to see if they still make a difference with the NDS.  But even if the improvement is vanishingly small, given their low cost, and given there is no downside, I will likely leave them in place.

ATB.

Hook

Chris, having a 'green' switch is  really a positive thing. At the physical Ethernet level, really the signals are sent as differential analogue signals, ie they are voltages modulated at RF speeds. There is not such a thing as 1 or 0s apart from in conceptual thinking. A green switch usually reduces the gain and therefore the amplitude of these signals it transmits into the patch lead if the switch detects the length of the patch lead is short. This means the power received the other end of the patch lead is reduced. This saves money and also the amount of any RFI energy produced. As a by product.. So is a good thing..

Simon

Hook, it will be interesting to hear from you whether you detect a difference. The NDS has increased screening over the NDX internally so perhaps the RFI effects on the digital clocks and analogue circuits from the Ethernet patch lead is reduced. As far as the Burndy leads I can't see the PSUs themselves producing RF current.. But it isn't something I have tried.. I would be intrigued  if it makes a difference.. But not surprised..I

When I demoed the NDS I left the chokes in place.. 

Please explain more - I thought that was the whole concept of digital...I used to work on Nortel OC-192 Long Haul optical equipment (SONET here of course, SDH where you are), and the smallest tributary I ever dealt with was OC-3...sometimes DS3.  VoIP was just coming in as I left & ethernet was local stuff only, so I am interested to know more.

Mark, you are showing your age... So lets use WAN bearer technology, then we are therefore  at the physical level.. Which according to the OSI seven layer model, a layer of  continous (not discrete or digital) physical signals such as electric voltages or light that are modulated to provide digital signals for the link layer. These days the physical layers use of lots of continous encoding techniques such as in DSL, wifi, VSAT etc. to support our ATM or Ethernet links.

In our humble home LAN we are using, usually, 4 differential twisted pairs to convey the send and receive voltages which when combined and reconstructed in each direction represent a binary signal for the encapsulated link layer.

Now so called Green or Smart switches detect shorter cables and reduce the current and therefore the voltage on the send wires... These therefore save energy. This is possible as Ethernet UTP is specced (upto 1Gbps) for 100 Meters, and so the line voltage has got pass through the losses of the cable. If the cable is shorter and there is therefore less loss and so the line voltage can be reduced.

Starting to see this appear more and more in data centres  as well as in the home/home office  etc, where energy savings are very topical.

Sometimes these switches are marketed as 'Smart Switches' and detect if the cable length <10m to reduce line current.

Hopefully that's rung a bell for you.

Simon

Simon - if it were ONLY network topology where my age was showing!

Funny, because with long haul lightwave networks, the problem wasn't if you would attenuate the signal because of the shorter distance, but how far could you push it before you had to run it through an optical amplifier, and then how much further before you had to use an ADM and rebuild the signal. But Long Haul and home/business networks are almost nothing alike - except perhaps for the ring structure concepts in some of them.

I can still remember going into the MCI/WorldCom lab and there would be giant spools of fiber optic cable - 500+ miles of it, to do testing between units...and how many different DWDM wavelengths could they cram onto one stand, with each signal only nanometers difference from the others: 32 discrete signals on one hair thin strand of glass...

Thanks for the explanation!