Network connections - far more crucial than I could ever believe!

Another one down the rabbit hole.

Seth - good to hear your enjoying the fibre bridge, a few gents (note some now dont) here persuaded me to try it a few months ago it was a major improvement, I did remove it a few weeks ago and promptly reinstated it.

Just awaiting linear power supplies to replace the ifi smps ones I use.

There is another cable thread, best to find an obliging dealer to loan AQ, Chord, Miechord, I am also waiting delivery of a Ghent cable which the computer tweaked swear by.

If it works for you good, if it doesn't you tried.

the good thing in the network bridge with fmc is that it costs very little.  I was very happy with mine for quite 2 years, with linear ps on fmc.  But recently, when i replaced my first lan cable connecting the router and the switch by an expensive lan , audioquest diamond, i realized that i prefer now, a little bit, the system without the network bridge.  

So i have now audioquest diamond > cisco 2960 switch > audioquest diamond > nds and unitserve > audioquest vodka > cisco ( unitserve on linear ps).

I feel that high end lan cables are removing noise as do the fmc with fiber optic and linear ps.

With the fmc and 2X audioquest diamond, i have a little bit more softer tones but in the same time a bit less immediacy and prat, and a more complicated network too.  So i choose without finally today.

FR, lan cables can’t remove noise... they can introduce  varying degrees of cross talk rejection to aid longer distance error free transmission at higher bandwidths and they can more effectively reject alien noise (noise from outside the cable) and can act as variable RF loads to the devices they are connected to potentially modulating the ground plane (there is an interesting TI design paper exploring how to mitigate the effects of this for Ethernet loads)

Consumer FMCs will have a propensity however to add PHY layer clock noise... which potentially can produce intermodulation products in connected devices.

From your description I would hazard a guess you are adding more clock noise into the connected streamer with the FMCs but perhaps  you prefer the resultant effect which is fine. My only observation is what you have found may be very streamer dependent.

[@mention:1566878604014936] please post your findings with the Ghent cable

Thanks for the responses. It seems some advise caution that MCs aren’t a magic bullet.  I wasn’t really thinking they are, which is why I explained how my setup was probably far from ideal - the point is more that I expected no effect at all, but changes are so blatant.  Perhaps the rabbit hole has better acoustics?

Ultimately, it’s the first time my system has totally “clicked into place” and thrilled me, without nagging thoughts of slight edginess, lacking bass drive or that things could be better - does that really sound like PHY layer clock noise?  It’s also worth noting that I have a consumer-grade switch, so perhaps I’m still improving on what I had with the MCs.  I’m intrigued that some think MCs add much complexity to their network - I would be very surprised if they increase latency in any meaningful way for audio data transmission, but perhaps someone can clarify that for me?

[@mention:56335183628232089] sounds like your approach is where I might be heading, but it’s encouraging that you can get more music from your system by focussing on the network - after many tweaks on my system (power supply, speaker cables, speaker placement, interconnects, etc) this has been the most fundamental change and I’m still surprised by it.

I’d also be very interested in thoughts about the Ghent cable, when you’ve had a listen [@mention:1566878604014936] :-)

If the sound quality of a system happens to depend crucially on the quality of network connections than I would argue that there is something fundamentally wrong with that system or with its environment. Perhaps you can try another streamer or network player and see whether you observe the same kind of sensitivity. If this is the case, you should probably have a closer look at your LAN. I wouldn't invest any money or time in fancy cables, switches, etc. at this point as you still have to understand where the problem comes from.

Seth posted:

Ultimately, it’s the first time my system has totally “clicked into place” and thrilled me, without nagging thoughts of slight edginess, lacking bass drive or that things could be better - does that really sound like PHY layer clock noise?  It’s also worth noting that I have a consumer-grade switch, so perhaps I’m still improving on what I had with the MCs.  I’m intrigued that some think MCs add much complexity to their network - I would be very surprised if they increase latency in any meaningful way for audio data transmission, but perhaps someone can clarify that for me?

Digital noise can cause a loss of fine detail and smoothen the end result by masking fine detail. Depending on system that can produce an overall pleasing effect with a removal of edginess and fine detail to the benefit of overall coherence and even possibly apparent musicality of the media... it kind of depends on the quality of the DAC and resolution of the speakers, electronics and room reflections.

As far inline media converters, which are usually designed to create long line Ethernet segments of greater than 100metres ... almost certainly there will be no apparent increase in latency, I would have though at most it would be 1mS or less.. less than a thousandth of a second. The biggest issues to me are the potential electrical and digital clock noise they can introduce, you should also check interface stats for corrupt data (not often possible with consumer devices), as Ethernet fibre does not like short distances. If you need a long line of around 100 metres or longer I would try and use fibre SFP modules attached  into switches rather than separate devices. (They are often on the uplink ports for edge switches which typically traverse longer distances). Until Naim support fibre connectivity directly, I say at best all you are doing is noise shaping.. and there are probably more reliable ways of achieving that... possibly by playing with the loading and effects from using different Ethernet patch leads.

Simon-in-Suffolk posted:

FR, lan cables can’t remove noise... they can introduce  varying degrees of cross talk rejection to aid longer distance error free transmission at higher bandwidths and they can more effectively reject alien noise (noise from outside the cable) and can act as variable RF loads to the devices they are connected to potentially modulating the ground plane (there is an interesting TI design paper exploring how to mitigate the effects of this for Ethernet loads)

Consumer FMCs will have a propensity however to add PHY layer clock noise... which potentially can produce intermodulation products in connected devices.

From your description I would hazard a guess you are adding more clock noise into the connected streamer with the FMCs but perhaps  you prefer the resultant effect which is fine. My only observation is what you have found may be very streamer dependent.

 

I tried to say that i don’t use fmc anymore.  With my recent all high quality lans and new cisco 2960, i find that now the sound is  a little better than with fmc.

nbpf posted:

If the sound quality of a system happens to depend crucially on the quality of network connections than I would argue that there is something fundamentally wrong with that system or with its environment. Perhaps you can try another streamer or network player and see whether you observe the same kind of sensitivity. If this is the case, you should probably have a closer look at your LAN. I wouldn't invest any money or time in fancy cables, switches, etc. at this point as you still have to understand where the problem comes from.

i rather tend to think that the problem is the poor quality original home network, and specially the poor commercial router.  

French Rooster posted:

Simon-in-Suffolk posted:

FR, lan cables can’t remove noise... they can introduce  varying degrees of cross talk rejection to aid longer distance error free transmission at higher bandwidths and they can more effectively reject alien noise (noise from outside the cable) and can act as variable RF loads to the devices they are connected to potentially modulating the ground plane (there is an interesting TI design paper exploring how to mitigate the effects of this for Ethernet loads)

Consumer FMCs will have a propensity however to add PHY layer clock noise... which potentially can produce intermodulation products in connected devices.

From your description I would hazard a guess you are adding more clock noise into the connected streamer with the FMCs but perhaps  you prefer the resultant effect which is fine. My only observation is what you have found may be very streamer dependent.

 

I tried to say that i don’t use fmc anymore.  With my recent all high quality lans and new cisco 2960, i find that now the sound is  a little better than with fmc.

 

Ok interesting... so have you the fine detail and that more natural presence now, or is it smoother and softer?

nbpf posted:

If the sound quality of a system happens to depend crucially on the quality of network connections than I would argue that there is something fundamentally wrong with that system or with its environment. Perhaps you can try another streamer or network player and see whether you observe the same kind of sensitivity. If this is the case, you should probably have a closer look at your LAN. I wouldn't invest any money or time in fancy cables, switches, etc. at this point as you still have to understand where the problem comes from.

I completely agree.

I find this all fascinating. I’m not sure how we know when we have it right, more likely it will ‘sound right’ to the individual listener. [@mention:1566878603876589]Suffolks knowledge and expertise is worth its weight in gold! 

Finkfan posted:

[@mention:1566878604014936] please post your findings with the Ghent cable

Will do it is on the Slow boat from China at the moment.

Simon-in-Suffolk posted:

French Rooster posted:

Simon-in-Suffolk posted:

FR, lan cables can’t remove noise... they can introduce  varying degrees of cross talk rejection to aid longer distance error free transmission at higher bandwidths and they can more effectively reject alien noise (noise from outside the cable) and can act as variable RF loads to the devices they are connected to potentially modulating the ground plane (there is an interesting TI design paper exploring how to mitigate the effects of this for Ethernet loads)

Consumer FMCs will have a propensity however to add PHY layer clock noise... which potentially can produce intermodulation products in connected devices.

From your description I would hazard a guess you are adding more clock noise into the connected streamer with the FMCs but perhaps  you prefer the resultant effect which is fine. My only observation is what you have found may be very streamer dependent.

 

I tried to say that i don’t use fmc anymore.  With my recent all high quality lans and new cisco 2960, i find that now the sound is  a little better than with fmc.

 

Ok interesting... so have you the fine detail and that more natural presence now, or is it smoother and softer?

without the fmc and audioquest diamond lans and new cisco 2960, the sound have perhaps a little less details, is a very little bit harder than with fmc and same lans.  But the music seems a bit better balanced and is a little bit quicker, with a bit better prat.

Before, with meicord lan between router and switch,  still same audioquest diamond before nds, i was preferring the sound with fmc.

So it was when i replaced the meicord lan by another audioquest diamond that i choose finally to put off the fmc.     I am not sure to be very clear.

No loss of detail for me - I’m pretty happy so far.   

Seth posted:

No loss of detail for me - I’m pretty happy so far.   

Did you understood that i lost some details with fmc?    i tried to say the contrary : i lost a very little bit some details and softness when i removed the fmc.  But for a very minimal difference.

On the pro side( without fmc but 2 audioquest diamond):   quicker and better balanced sound.

with average quality lans everywhere ( X3):   i would keep the fmc.

I will try again removing the fibre bridge in a few weeks, but as mentioned i did try a few weeks ago and promptly reinstated it, just so much better in every way.

I use a £1000 Chord Indigo Tuned Aray and find the fibre bridge still makes a big difference, upgrading from ifi to lps I am expecting a bigger improvement, but they are still being made ????

The Ghent for me will be the acid test, but that was still £100.

Obsydian posted:

I will try again removing the fibre bridge in a few weeks, but as mentioned i did try a few weeks ago and promptly reinstated it, just so much better in every way.

I use a £1000 Chord Indigo Tuned Aray and find the fibre bridge still makes a big difference, upgrading from ifi to lps I am expecting a bigger improvement, but they are still being made ????

The Ghent for me will be the acid test, but that was still £100.

what is your lan cable between your router and the cisco ? the c stream i think to remember ?

perhaps it is the weak link in your network ?        i don’t know, but perhaps a second indigo in the chain, between the router and cisco, may change the game.   For me it changed the game, if i can express myself like that.  But it can be also an expensive upgrade for a subtle improvement...difficult to predict.

French Rooster posted:

Obsydian posted:

I will try again removing the fibre bridge in a few weeks, but as mentioned i did try a few weeks ago and promptly reinstated it, just so much better in every way.

I use a £1000 Chord Indigo Tuned Aray and find the fibre bridge still makes a big difference, upgrading from ifi to lps I am expecting a bigger improvement, but they are still being made ????

The Ghent for me will be the acid test, but that was still £100.

what is your lan cable between your router and the cisco ? the c stream i think to remember ?

perhaps it is the weak link in your network ?        i don’t know, but perhaps a second indigo in the chain, between the router and cisco, may change the game.   For me it changed the game, if i can express myself like that.  But it can be also an expensive upgrade for a subtle improvement...difficult to predict.

 

I'm saving for 1 Sarum T my friend ????

Chord are of the view the improvement yields from wherever you install the Aray.

Obsydian posted:

French Rooster posted:

Obsydian posted:

I will try again removing the fibre bridge in a few weeks, but as mentioned i did try a few weeks ago and promptly reinstated it, just so much better in every way.

I use a £1000 Chord Indigo Tuned Aray and find the fibre bridge still makes a big difference, upgrading from ifi to lps I am expecting a bigger improvement, but they are still being made ????

The Ghent for me will be the acid test, but that was still £100.

what is your lan cable between your router and the cisco ? the c stream i think to remember ?

perhaps it is the weak link in your network ?        i don’t know, but perhaps a second indigo in the chain, between the router and cisco, may change the game.   For me it changed the game, if i can express myself like that.  But it can be also an expensive upgrade for a subtle improvement...difficult to predict.

 

I'm saving for 1 Sarum T my friend ????

Chord are of the view the improvement yields from wherever you install the Aray.

very good idea, i am glad for you in advance.  I hope to do the same in some months, maybe one year.    

French Rooster posted:

Did you understood that i lost some details with fmc?   ... with average quality lans everywhere ( X3):   i would keep the fmc.

No, Monsieur Rooster - I think your comments are spot on and you make perfect sense.  I do also agree that the quality of my existing network is probably behind the difference between what I hear now and what I heard before.  But, if MCs have gone some way to reducing the issue - to the point that I’m happier than ever listening to my NDX/XPS DR, NAC282/Hicap DR, NAP250, Focal Sopra, Hi-line and Powerlines - I’m not that fussed by comments that I should look elsewhere or that I’m probably just noise shaping. 

Seth posted:

French Rooster posted:

Did you understood that i lost some details with fmc?   ... with average quality lans everywhere ( X3):   i would keep the fmc.

No, Monsieur Rooster - I think your comments are spot on and you make perfect sense.  I do also agree that the quality of my existing network is probably behind the difference between what I hear now and what I heard before.  But, if MCs have gone some way to reducing the issue - to the point that I’m happier than ever listening to my NDX/XPS DR, NAC282/Hicap DR, NAP250, Focal Sopra, Hi-line and Powerlines - I’m not that fussed by comments that I should look elsewhere or that I’m probably just noise shaping. 

Your findings make also perfect sense and i invite you to enjoy your network with fmc as i enjoyed it.  The network bridge is a wonderful upgrade for a very little investment.  

I have a little bit more enjoyment now without it,  but for a cost of 2 audioquest diamond lans, so 15 times more expensive cost of network components.

Minh Nguyen posted:

nbpf posted:

If the sound quality of a system happens to depend crucially on the quality of network connections than I would argue that there is something fundamentally wrong with that system or with its environment. Perhaps you can try another streamer or network player and see whether you observe the same kind of sensitivity. If this is the case, you should probably have a closer look at your LAN. I wouldn't invest any money or time in fancy cables, switches, etc. at this point as you still have to understand where the problem comes from.

I completely agree.

I also agree, but to be fair using the right diagnostic equipment locating the sources of issues is probably beyond most on this forum... if you read the majority of posts one gets the impression what some are calling an ‘upgrade’ is probably a case of moving noise profiles around and finding a noise profile that matches their system, preference and room. This is not as strange as it sounds, as Naim do something a little similar with their firmware assembly code execution timing and resultant electrical noise.. this is why firmwares sound different, even though the actual DSP filter has not changed at all. Therefore tweaking with noise profile  shifters such as consumer FMC devices and specific consumer boutique Ethernet patch leads probably yields good results on some Naim boxes... but possibly  less so with the newer high end streamers with their increased decoupling. So my only real caution is that one may find a noise profile agreeable  with one streamer that might sound somewhat different with another..

But I do agree, if you can and have the engineering skills, it’s probably better to locate and isolate or resolve the issues... rather than relying on metaphorical sticking plasters.