Changing Switch-Router from Cat5 to Cat7

You said to your wife, 'listen to how much better this cable sounds' and she humoured you?

I use AudioQuest Cinnamon between my router and the switch, and it sounds considerably better than a standard cable, even when playing from the Unitiserve, which comes after the switch. All my cables are Cinnamon, and the system sounds just great.   

Errrr  ????   no garyi.  She ain't that kinda lady,  & you better watch out as she takes no prisons with people who dare such a remark.

She has a very keen ear & picks up on some stuff I don't. .  We respect each other’s opinions (but not music genre's) frequently getting into deep discussions on different recordings & even the differences in venues when it comes to live classical,

Mike your good old enemy RFI - remember a lot of it travels common mode from devices. Your Cat 7 cable might be rejecting it better - or your old cable is generating from the carrier frequency due to less than perfect twisted pairs. Try disconnecting the router and see what happens - the link to the switch will go down - but the cable will be still connected to the switch - and I bet you hear an improvement.

Simon

I guessed the same Simon,  my thought was the Cat5 with its unused & as you call them "less than perfect twisted pairs" was acting as an aerial & the screened Cat7 does not (or less so)

I tried disconnecting the cable to the router last evening,  but not as your post, I pulled the Cat5 from its switch port & felt it seemed to give the same improvements as adding the Cat7 so my take on that was the Cat5 degraded the SQ rather than the Cat7 improving it.

+1 for RFI differences.

This fits in with all the testing I've done with different cables (with similar and different constructions) and the presence / absence of ferrites.

I've also experimented with vibrational damping, but as yet I don't feel the results are definitive enough to be announced here (even if I gave caveats I'd be risking another round of attacks).

Huge, have you posted all your findings as a summary?

Maybe I missed something,  I've seen lots of posts on various subjects (& the "attacks"), but nothing listing it all as a summary.

I would be interested & I'm sure a few others also.

Hi Mike,

Here's my summary in respect of Ethernet cables

https://forums.naimaudio.com/to...s-overview-a-summary

In terms of use of ferrites in general, it's simple:

Anywhere where a cable enters or leaves a box that

1   Contains electronics that process audio signals.

2   Is/has a SMPS and will be switched on when the audio system is in use

Point 2 is probably overkill, but the only way to test is put them on then remove them

I hope someone will do a review of different ferrite rings to determine which one 'sounds better', it should do great things to further enhance the view others have of audiophiles.

Garyi

you are a proper brainless B****head.

wish you would go from the streaming thread, you've got nothing to add and no knowledge of anything computer audio related.

Please can we keep this civil.

Aleg,

I feel that was somewhat unfair; Garyi has made some positive contributions as well.  Yes he seems to be in the 'bits are bits' camp, but he's entitled to his opinion.

Garyi,

Expressing things that way doesn't help, and doesn't work (unless you were just intending to 'tug peoples chains', in which case Aleg IS right).

Huge, i think this whole thing is Interesting. Aleg said i was on his ignore list so he is fibbing and rather intense on the whole subject.

i am not here to tug anyones chains, i have played this game for a great many years. I moved to streaming as a solution at least five years before it was fashionable. I joined this forum before aleg was even born*

i do not hear differences between ethernet cables. And yes i have tried a few. This area i think is being infested by those purporting to have golden ears but offering entirely nothing in the way of evidence. The only latch we appear to have is rfi. But folks what is the audible effect of rfi? If it exists it can be measured no?

come on then.

until such time as something other than opinion is presented, then sorry my opinion is as valid as the next persons, when this subject is bought up i shall express my opinion, which is one of shear scepticism that someone 'hears' a difference with an ethernet cable between their nas and switch. I do not mean to appear rude or as someone tugging chains. I simply don't believe it. The easy answer is i come from a different perspective, as long as we are polite about the whole thing, then i think it can be fun.

Aleg. Have a shot of whisky and calm down, you are a very rude person, but i like you.

*probably

Garyi,

You are most certainly entitled to your opinion - you've heard no effect, and as yet there's no proof that meets the test of statistical significance, so you have no particular reason to believe that there is an effect.

I am also somewhat sceptical about sound differences arising from Ethernet cables (especially the cable from NAS to Router or Switch) with the exception of the effect of RFI.

For me, changing the cable from the router to the streamer for one with similar RFI rejection had no significant effect, whereas changing to one with better RFI rejection had a considerable effect. Furthermore changing from unshielded to shielded cable gave a broadly similar effect to that obtained using ferrite chokes (supporting the RFI theory).

As to any other (as yet unproven) effects, until I have definitive proof one way or the other I'm going to keep an open mind - I don't know everything!

However I'm not at all sceptical about the effect of RFI.  I have designed audio amps and I am well aware of the perturbations RFI can cause to the operation of audio amp circuitry (there are too many details to give a good description here), but there are many references to this in the various design manuals.

Incidentally all manufacturers put considerable effort into reducing the effects of RFI, but it's not possible for them to make allowance for all circumstances.  I seem to be unlucky where I live and have always had to put in additional protection from RFI (even with an Audiolab 8000s).  It may be that you don't hear an effect because you don't have an RFI problem where you live (you lucky ... ).  And yes, with the right equipment, the effect of RFI on audio circuitry can be measured (frequently the effect is most obvious at the driver or pre-driver transistors).

And yes it was the lack of politeness that disturbed me.

Absolutey RFI filtering and mitigation is a vast and interesting area in electronic design and engineering. Ethernet RFI is also pretty well understood, and you are even challenged on it on various industrial accreditations from the likes of Cisco etc.. And of course I have posted the Texas Instruments white paper on design guidelines for Ethernet RFI mitigation.

Now RFI on audio circuitry can cause many issues, but the one that I am most familiar with and can be measured and therefore mitigated is intermodulation... And this is not new and has been a curse that goes right back to the earliest valve RF circuitry. This is caused by the sums and differences between disparate frequencies... This can produce HF signals that can affect the operation of  feedback circuits and/or induce reflections in leads etc.. This is not benign.. But I am quite prepared to accept that some people when they hear music that has these distortions these might not consciously recognisable.

Interestingly (to me), Moog used this concept and concentrated it to create the Ring Modulator which was used as a creative audio effect in early analogue synthesisers 

Simon

Some more on this now I’m back from a few days living it up worshipping at the altar of Adnams, scoffing north sea bounty & twitching-in-suffolk  

The original router to switch Cat5 routing had about 0.7m running alongside the broadband phone cable, plus about 3m rolled up under the cabinet that also had two DC SMPS DC feeds (both loaded with ferrite close to the SMPS) in the area.

This morning I tried the Cat5 run alongside the phone cable as it was before, then moved it well away & noted a small change in clarity with music supplied UPnP from NAS. 

Pulling the rolled up 3m section out onto the open floor away from the SMPS cables made no further improvement that I could detect.

Installing the Cat7 made a further improvement, but it did not appear to change if run alongside the phone cable or away from it.

With a caveat that I am talking very small changes & I fully accept this is all subjective, I think I have got to a point where I’m happy that the Cat5 was picking up noise from the phone cable & also maybe some other RFI as discussed between Huge & Simon. I am not saying the new cable is responsible for the change.

Anyhow,  I've fitted a length of plastic conduit down the back of the cabinet to house the 2x phone & DC feeds & with the Ethernet hanging loose down the back of the cabinet about 20cm away before diving into the cabinet with the other Ethernet & on to the switch,  so should be no chance of any interference from other cables. 

Mrs Mike is out tonite so might give my favourite albums some wellie,  but no more cable swapping - I done (until the next time).       

I suspect the terminations play a part as well. In addition to CAT7 being shielded and having greater pair balance (noise suppression), the end connectors being metal provide some shielding to the sensitive interface area.

Most switches employ miniature SMPS internally to produce their working voltage rails. When I was using CAT6 I found ports further away from the power end cleaner! I noticed some subtle effect with my c-stream too, along with a short burn- in or settling period. Difficult to describe, but some positions just had an increased ease or 'rightness' to them.

Different models of switch had a far more dramatic influence on 'edge'.

Ooooo ---  do I see a new thread  ---  What Switch Are You Using?  

That'll bring the nay's & yea's into a real frenzy. 

I agree, for me Cisco 2960 fanless  8 port with internal shielded SMPS is the best sounding for me. Netgear's were a car crash for me.. I think mostly due to poor noisy power supplies.

until know down played this on the forum because of nay sayers.

It's sad I have had a few  private emails of fellow enthusiasts who have preferred not to share experiences and advice for being ridiculed on this forum...

Ha! this is the next part of my system on which I want to work.  In my case, all recommendations will be  gratefully received!

It's so irritating when people use ridicule in place of logic and discussion.

Huge.. Well I have found the OS TCP/IP stack of the UPnP DLNA server seems to affect sound quality of TCP transferred media when using Naim. Fairly sure this is windowing protocol activity affecting small amounts of electrical noise on the Naim network player streaming board.

Also RFI from connecting switches has it's part to play too.

Simon

RFI from switches - yes, a given.

Activity from the DLNA server: initial reaction was...

Surprising?,

Second reaction was: OK, thinking about it, that actually makes sense.

Third reaction, it could provide additional explanation as to why some people are reporting different SQ from different NAS drives even with linear PSUs and ferrites in place.

So, which are the best DLNA servers from the perspective of minimal extraneous activity?  I've not tried using different servers on my Synology NAS, but I think installing a good switch is higher priority.

Huge, I think it's the layer 3 transport protocol on TCP/IP. It sets the type of activity and the frequency of data confirmation between the two peers.. That is the TCP/IP stack on the DLNA server and the network player. You can see this using Wireshark if you are interested.

i have found the Raspberry Pi sounds great, and an HP Pavillion PC running Windows 7 the worst. Both were running Asset.

I also found Windows Home Server V1running Asset sounded good.

A Netgear ReadyNAS running various DLNA servers sounded less good than the Pi.

Cheers

Simon

Thanks Simon,

That windows 7 sounds worst of those is no surprise, there are so many background processes that use polling to interrogate the network for all sorts of stuff - it's very chatty.  Additionally a PC SMPS also has very high switching currents so there's a big risk that the filtering won't stop the RFI sufficiently.

My knowledge of the specific behaviours of the separate layer is quite limited, the precise detail here is getting beyond my current understanding.  I think need to do some revision!

I also need to research network switches.

hi Simon,

I note you've said on several occasions you found the Cisco 2960 audibly better than Netgear, and am wondering if you could confirm the model - is it WS-C2960G-8TC-L (7x10/100/1000baseT and 1 multipurpose uplink?), or another variant?

I've used both Netgear Prosafe gs108v3 (not the T-200 model with managed features), and the gs605v4, both on TP 12v linear psus.  I found the gs108(now retired) introduced a distinct 'edge' to the timbre, while the gs605 was more neutral.  However just having the switch connected to the Unitiserve brings a distinct 'hardening' shift to chamber instruments played off local USB which isn't correct. I also think the background naim hiss (in tweeters) becomes slightly crisper with it connected!  I'm attributing this to the internal voltage-reducing SMPS within the switch coupling noise

In all other respects I now have the NAS with TP PSU and c-stream sounding significantly better than the local USB, whereas previously it was a decided second.  Black background, layering, scale/dynamics, openness, clean bass, the speakers disappearing - are all great.  But the NAS source retains this timbral shift.  A good example is the shakers on Rebecca Pidgeon's 'Spanish Harlem', which are so obviously shakers on USB with no network connected, but smear into something more scratchy via NAS.  Period instruments similarly show this affection obviously.  One last thing to fix...

If you could please confirm the Cisco model you've found successful, I'll order one immediately!

Hi this is what I am using closest to my Naim

and use chokes on connected Ethernet patch lead.