Audiphile Ethernet cables - advice needed

Hi THX,  these have been around for a while & I suspect a have a smidgen of "sponsorship" from  AQ.

I would like to see something a bit more independant & comparing price banded cables. 

Ahh sorry Mike didn't realise the 'sponsorship' element... not worth anything then really. I did search I im interested in purchasing a 'final link' cable (Switch to Naim) myself.

I have an update on my quandary of yesterday I'll add later 

THX,  I'm also interested in a final link cable as due to furniture rearranging, my Supra Cat7+ is now only just long enough,  so an opportunity to try something different.  

No way am I going down the Sarum price rout,  but maybe a C-Stream,  or do I stick with Supra. 

My problem is I'm an electrical engineer & objectivist & "I know" ethernet cable cannot change sound.The data being sent is based upon an IP packet, with header & frame, its not a true analogue audio signal where cable quality can be affected by signal loss & corruption. With IP networking the packet is received & verified as intact, or re-transmitted.

My problem is I have heard differences, I’m suspecting the low spec cables have defects and/or they have crosstalk, losses & bandwidth issues, whereas higher spec cables are not so affected. Whatever, to my ears the differences are minute at best. Problem is you can spend a lot of money seeking audio nirvana with Ethernet cables & end up with very little bang for a lot of bucks.  

Mike - remember its not to do with the digital encodings in the Ethernet frames, but the high frequency  analogue voltages (RF) present associated with the transmission and how they couple with the connected equipment.

The modern Ethernet cable is effectively designed and implemented as four analogue transmission lines - but the connectors and cross talk between transmission lines causes some side effects and emissions.

In the data world this is usually transparent unless you are working in a RF necessarily quiet, very RF noisy or Wifi sensitive environment - but is different in the analogue world.

A certain satellite uplink and down link station I am fairly familiar with would not permit electrical Ethernet in certain areas because of RF emissions (irrespective of Cat ) from the cables and connected equipment like an IP handset.

Simon

I guess my amateurish wording is saying much the same thing Simon.  I understand in the digital domain that the cable does not effect or change anything.   But what goes on in analogue, my ramblings are saying that seems to me is where the SQ variable could exist, if in fact it does.  Its the main reason I prefer to use correctly grounded STP over UTP & hopefully avoid crosstalk & noise emissions..

And to my point;  there is no unbiased cable testing/reviews that I can find. What's printed seems to be either setting out to "prove" that a cable changes nothing - & uses well known testing equipment to prove their point,  which it always will unless the cable is defective. Or on the other hand the reviews (not tests) set out to wax lyrical about the wonders of a specific cable make & smell strongly of "sponsorship".

A USA test has shown that many cables do not conform to the claimed Cat specification,  but they do not touch on SQ variations,  other than implying that it does not.

I get the feeling that the hifi comic journalists choose to not get involved, maybe too controversial.

I suspect you're right Mike. I don't think the companies do themselves any favours by the term audiophile cable (whether Ethernet or USB) in the marketing blurb - the cables themselves have no sound signature as in conventional analogue signal cables but can have an audible effect further down the chain in the connected audio device. 

If these effects and emissions are well-understood then shouldn't it be fairly straight forward to design cables to minimise them - at a reasonable price?

Presumably the galvanic isolation that many high-end hi-fi products have on their ethernet ports helps but doesn't stop all the unwanted 'noise' hence the reason why many people use chokes as well.

Mike, Simon,

I'm in the same boat.

Superficial hypothesis:

Ethernet works, resends or fails.  In each case the data are transferred or lost.  The data are the sound encoded, so I shouldn't hear a difference (or I should hear nothing in the case of failure)

But I do hear a difference...  Aaagh!

OK, What now...

Reform the hypothesis to include

Analogue effects carried down the digital cable, including RFI etc

Possible interference or system loading from increased processing to handle resends etc.

Effect of processing a WiFi connection close to sensitive analogue electronics.

Ah, phew! normality restored, theory once more corresponds to practice.

One should not place too much faith in theory until it is confirmed by practice.

One should not place too much faith in practice until it is confirmed by theory.

I suppose the counter argument is - Ethernet is a well known / understood carrier. Why is my Streamer sensitive to what's connected to the Ethernet port ?

This type of thread can run and run (as it has done many times) with only subjective opinions and no actual measurements to back it up. I think the mechanisms for why an Ethernet cable can make a difference are well understood but with no actual measurements to back them up. 

Well for most applications these side effects are not an issue. For attached audio devices and a few other specialist scenarios it perhaps is. If you really want to it is straightforward to measure the emissions from an Etherrnet lead with an EM meter.. interesting you never see this quoted - just mystics referenced on the audiophile Ethernet lead lading to pure subjectivity.

I am sure a brand new electrical Ethernet physical standard could be adopted for the audiophile which had reduced emissions, but it may be much more expensive than normal Ethernet leads and almost certainly not compatible with exiting connectors / devices.

Of course there have been alternate electrical Ethernet standards over the years such as token ring coaxial connectors which almost certainly had reduced emissions - but they also had reduced bandwidth.

Now of course if one was serious about this one could simply use send and receive fibres - its used in industry where electrical Cat cable is not appropriate  - but is more expensive and you need specialist adapters on the network equipment. Zero emissions from the bearers however.... interestingly it has not caught on for consumer audio . perhaps because it is genuinely more involved and genuinely more expensive..... and genuinely provides no emissions.

Simon

So just replaced my AQ Cinnamon from NAS (Vortexbox) to switch (GS105) with a Meicord cable. Like Mr H I can report it brought a big boost to the bass, and music sounds overall a lot fuller now. Maybe a slight loss of detail over the all Cinnamon run, but that's splitting hairs (probably more like my ears just getting used to the new sound). I can see how all Meicord would be too much in my system (hard to tame the bass of Ls50's near the wall). Crazy how these things work, but that they do. One possibility is that the Cinnamon is shielded and the Meicord isn't. Definitely like traditional IC's these are a good way to 'tune' a streaming system. 

Did you try the meicord in both directions, as it is directional, even though its not marked. This can make quite a difference if its the wrong way around. Also, like everything else in the system it seems to "settle" after a few weeks and really come on song.

I found the biggest improvement in vocals, especially female, which with 2 cinnamons sounded a bit two dimensional compared to the meicord. Put on something by Eva Cassidy, Norah Jones, Katie Melua or Joss stone and the difference was immediate and obvious.

Another thing to try is using the meicord between the switch and streamer. On my system it sounded better as you are now using it but on a friends system it sounded better between the streamer and switch. We are both using all naim systems, fronted by NDS/555dr. 

Anyway, glad you also found mixing the two cables gave a positive effect in your system. 

Happy listening

Mr H

Nope, you are all delusional. All ethernet cables, mains cables, interconnects, competently designed DACs, etc etc sound the same.

Ah, sorry, for a moment I thought I was on the fishy forum, where forty-page threads on cables are the norm - as Mingh is now discovering 

I have Audioquest and very happy with it, makes a differences!

Fred

Ethernet cables at the data level clearly are not directional as Ethernet link frame and transport packet data flows both ways or they wouldn't work.. The data flows both ways concurrently with duplex links or one at a time with half duplex links.

However if the cable is shielded, the drain might be attached only at one end, which could account for some of the differences the RFI causes.

So yes I say if the Ethernet cable causes a different sound to occur depending on directionality then I suggest you are actually tuning the effects of the RFI on the host system, and it is probably better to reduce the RFI source in the first place to improve consistency rather than trial and error tuning with boutique cables which might only be temporary. 

Simon

The meicord is not sheilded but it makes a very noticable difference which way around you install it. I have no idea technically why but it definitely makes a difference.

Even with UTP there can be a difference, the pitch of the twist isn't rigidly mechanically enforced in the cable, leading to one end having a slightly different pitch than the other.  The pitch can also get disturbed when the cable is bent and straightened.  This variances will again tune the RFI interactions.

Oh bugger... I've just spent £1,750 on a SuperLumia interconnect after reading somewhere that they sound better.  I wish I'd read your post before I did this, Mr Mole.

Indeed the shield drain is one obvious example I used as an illustration, but there are many other variables that an Ethernet lead could tune analogue RFI. Not least the send and receive could have subtly different transmission line properties due to manufacturing tolerances of the twisted pair / connections etc. If you swap the lead around the send becomes the receive and vica versa.

Simon

I did try both directions. One way sounded much more closed in, esp vocals (i.e. midrange) like a strong V curve. The other way more natural. Unfortunately I didn't buy a long enough meicord to reach between the VB and the UQ - my modem, switches and VB are in a closet in the basement below the hifi and powered by a small LPS. I have a 3m Cinnamon that then runs up a hole in the floor (it's an old house). I wasn't prepared to jettison the pricey long run of Cinnamon for the meicord, but if I have time to experiment sometime I'll drag a switch upstairs and try the meicord into the Uq (which I use for source only into a V1). My system isn't nearly as sensitive as a NDS/555 so might not be as noticeable. Strange that it would sound different in two identical systems (were your switches the same?). 

Ah, yes, rj, I feel your pain. Might I suggest a nice glass of Bordeaux would be in order for your next listening session.

I have now installed Chord C-stream for my first and final connections and initial impressions are positive - treble and mid-range appear more forward and clear than previous and there has been a subtle change in the bass (at least to my ears!).  I now intent to keep listening for the next month or so before trying the old cable sin again to see if the improvement then disappears.

Simon - you mention that is is probably more important to change the environment than to have better cables (paraphrasing) - apart from some of the obvious such as attention to cable dressing what, in your opinion would provide the best improvements in SQ.

Mr M - I do tend to listen with glass or two of the grape juice - perhaps this explains why the SQ improves over the session!

Paul

Paul, for me it was decoupling streamer from the DAC wherever possible, using a shielded Ethernet patch lead to streamer, routimg the Ethernet patch lead away from mains and interconnect leads, and putting a large RF choke on the SPDIF lead between streamer and DAC and a large choke around the Ethernet patch lead. I also use mains wiring to the audio equipment that is braided to attenuate RF.

I also do hear a difference from different UPnP media servers on different platforms.. I tend to find the NAS media servers sound least good (I have tried various apps on Netgear NAS's). Currently settled on a RPI 2 running Minimserver and going via a Cisco 2960 switch to the streamer... and that is a nice subtle improvement from my previous RPI 1 running Asset ... which is now retired from hifi duties and doing other 'stuff'.

I hope I have given you a few ideas..

Simon

Many thanks Simon, your sage advice is much appreciated. I shall certainly implement some of your suggestions.

Thanks again

Paul

I was told last weekend by a well respected designer of audiophile networking equipment that there is a difference between Cat 5 and Cat 7. The rise time for the analogue signal to change voltage levels representing 0 and 1 is apparently four times faster in Cat 7 than 5 and thus there is 1/4 the amount of jitter. 

Now, I'm not knowledgeable enough to defend this statement but it sounds logical given the higher data ratIng of 7. Plus I believe it is generally accepted that jitter affects SQ.