GS108 switch ports...

All 8 ports on the GS108 are gigabit.  With the PoE models, only four of the ports have power but I don't see why that would impact data transfer(?)

The cable wiring of Cat 7 and Cat 5e is the same.. There should be no difference. I wonder if some of the switch sockets or cable plugs are somehow damaged or mis shaped slightly.

It might be worth mentioning that Cat 7 is not  recognized and therefore not advised for Ethernet use by the TIA/EIA... But I doubt that really matters either, certainly for home/hobby networks.

Simon

Ken,

My set up is BT Homehub 4 (router) in an upstairs room connected to a Netgear GS116 (16 port) switch.  That connects to a Netgear GS108E switch in my lounge.  Connection from 108E to NAS drive is via AQ Cinnamon 2m cable and connection from switch to NDS is Cinnamon 8m cable.

I can plug the Cinnamon cables into any port on the 108E.

Do you old cables work in any of the ports?  If yes, maybe you have a faulty Cinnamon cable.  If your old cables exhibit the same problem as the AQ ones then maybe you have a faulty switch.

I use the smaller GS105 and can plug my cinnamon cables into any of the five sockets. The 108 should be the same. Sometimes you need to wait for the switch to sort itself out, so plugging the cables in and then rebooting the Naim bits may fix it. 

The GS108 is a straightforward unmanaged switch.

All the Netgear ports are non-specific & the switch will "learn" its connections,  the correct procedure (per Netgear manual) & (all unmanaged switches) is to power up all its connection devices before powering the switch.  I suspect that's why you found the router works only on specific ports, I follow this procedure even when I've had the system powered down (for hols etc.)

The other thing you might be mindful of with Cat7 cables is the screen/shield is carried through the cable & connected via the RJ/E-45 shrouded plugs & the GS series switches have shrouded ports that carry that screen/shield connection across the switch.    This means,  dependant on your NAS port ground/earth connection method,  that this can give you a not required 2nd earth connection on the ethernet network - your NDS port is connected to mains earth & that's the only one you need. Two ethernet earth points is not a good idea & especially so if the NDS & NAS are running on different power circuits.  

I use a UTP male/female adaptor on my NAS to isolate the 2nd earth,  a female/female connector with a short section of Cat6 (UTP) does the same thing as does using a complete section of Cat6 from switch to NAS.   

Wireless hub/routers don't have screen/shield ports & STP ethernet such as Cat7 is not a concern

Mike's suggestion is sensible. I would also use proper TIA/EIA approved Ethernet leads if you can, but I realize its too late for that now if you have bought them already.. And clearly others are having success.

if it is not some chronic earth loop that is interfering with the link setup, it has to lead/connector or switch.

Use a regular Cat5e lead to connect a test device to each port in turn. If the link comes up on each port then it points to an issue with the patch lead / connector.

Simon

What Wot Wat !!! ???  ethernet cables don't have a sound - do they ??   

silly me, i deleted he photo accidentally, not that this really makes any dfference, but here:

enjoy

ken

Or use an unshielded cable such as the Meicord between NAS and switch and then AQ switch to streamer - made a big difference in the sound on mine vs all AQ. 

+1

Not wanting to poo poo opinions but you know its all 1's and 0's don't you? The shielding in cabling is really for extremely high bandwidth application (i.e cat6 10GB Ethernet) or to protect against EFI weakening the signal in a factory environment causing excessive packet loss.

In a home environment there is absolutely no benefit in running cat6. With regard to the way you are transmitting data across your home network (music streaming etc) it will obviously be in TCP/IP format which is a guaranteed delivery method for data. A quick google can explain:-

TCP is a reliable stream delivery service that guarantees that all bytes received will be identical with bytes sent and in the correct order. Since packet transfer over many networks is not reliable, a technique known as positive acknowledgment with retransmission is used to guarantee reliability of packet transfers. This fundamental technique requires the receiver to respond with an acknowledgment message as it receives the data. The sender keeps a record of each packet it sends. The sender also maintains a timer from when the packet was sent, and retransmits a packet if the timer expires before the message has been acknowledged.

So basically, don't worry about the network cable - so long as the data gets there which is will do with any old crappy network cable, you're golden.

Just my two pence (I.T Consultant spending my wholeworking life in I.T/Networking).

I hate to break it to you, but physically the voltages on the Ethernet cable are not 0 and 1s, they are actually analogue three voltage level Manchester encoded signals with a carrier frequency around 31MHz (for 100 Mbps)

The so called "1" and "0" come at the next level up in the TCP/IP stack once the analogue voltages have been converted and signal processed. It's not really surprising that there are RFI artefacts from these RF analogue voltages, although the Manchester encoding is designed to minimise RFI from the analogue voltages.

Simon

Well congratulations for quoting layer 1 on the OSI model. What are you trying to prove... the data is error checked and the packets have to match exactly before it proceeds up the rest of the OSI stack. Come on, really?

Not really, rather I am referring to the fact I see the RFI caused by the physical transmission line characteristics has little to do with the data payload conveyed via the TCP/IP stack above the physical layer, unless there are chronic issues with the transmission line.

Therefore different Ethernet cables have to some extent different transmission line characteristics, and these can affect the interactions that occur when the RF voltage couple to other electronics at either end of the transmission line.

I hope you can see where I am coming from.

Simon

He's IT, not an engineer! 

I respect the fact that everyone is entitled to an opinion of course but I can't agree with you. Again we are both allowed to differ. We are dealing with data here, regardless of what happens at layer 1 ultimately there is no 'character' in data... it will be exactly the same data whether that is sent over WiFi, 100MB Ethernet, 1GB Ethernet, 10GB Ethernet. 

The TCP/IP stack is designed to handle error correction amongst other things. I'll not do a copy and paste but you can read all about it here: https://en.wikipedia.org/wiki/...ion_Control_Protocol. 

Anyway, we are all entitled to buy whatever we wish to spend our money on. IMO, just buy a well made cable and if you have high levels of interference then yes, get a shielded cable but only so you don't get excessive packet loss. I've never experienced issues in an office environment/home environment using regular CAT5E.

That said, I have my whole house cabled with CAT6. Whoops.

Oh wow. What is that meant to mean? I think we are discussing data here, correct me if I'm wrong though - honestly I'm all ears. 

Maybe we are all reading slightly different versions of this forum. I guess if the data is open to interpretation then who knows what everyone else is reading. Genius.

In fairness to simon, I don't think he has ever actually 'come out' audiophile stylee and said it defo makes a difference audibly, he appears to have left that to those willing to spend silly money on an ethernet cable.

I think from the posts I have read is that he believes that the cable ultimately connects the hifi to other stuff in the house and consequently could carry undesirable RFI to the hifi.

Now whether RFI is audible or not is yet another argument, but one thing is for sure, those willing to spend silly money on an ethernet cable, follow the same well beaten path that the more expensive the cable the better it sounds. Though as far as I can establish if you take the earth shield out one end as the cable should be designed this issue should go away.

I could be wrong on all this. I have a mixture of cheap and better quality ethernet cables in the house, some of it off the roll, some pre made, none more than £5 for a 2 metre cable. I have never established a difference in 'sound' with any of these so for sure am not willing to spend silly money testing the theories of audiophiles, 

the argument as to whether cable A sounds different or better than cable B s now new, no matter what the explanation is. you can pick any cable you like - interconnect, mains lead, speaker lead -- there are as many opinions as you care to enumerate.

theoretically, its hard to understand why there would be any SQ difference given the way TCP/IP apparently works, but thanks to garyi for raising an issue i have not considered up to now -- that is, the cable characteristics can affect the hifi that its connected to -- despite the data arriving bit perfect.at its destination. i guess once we are in the hifi world after the data has been received -- then the hifi circuits could be exposed to all sorts that may affect SQ.

i really dont know whether this is actually the case or not as i dont work in this area and i am a total novice as far as networking is concerned. but garyi raises an interest point nevertheless and which doesnt seem to contradict the bit perfect scenario. 'pure data forever'. i shall now go back to my cage before i paint myself into some corner...

i have a 12m AQ Cinnammon cable waiting to replace the bog-std one on the NDS once i have screwed up enough courage to do so. yes, the cost hurts a bit, especially given that my system sounds extremely good with the bog std cable so absolutely no hurry....

enjoy

ken

enjoy

ken

Ken, exactly. I think people get hung up possibly thinking that a cable designed to carry digital data some how magically waives its rights to have analogue characteristics. Clearly that is a nonsense. Whether you can hear this coupling is down to your equipment, your environment and your ears.

Simon

If ever there was a circular argument, this is it.  On one side the 1s and 0s are 1s and 0s and on the other those who hear a difference.  There is room for both views and, if you are open minded enough to actually listen to the options, you can make a purchase decision on that.  As someone who has recently chaned digital transports though I have to say that, whilst I can't comment on cables, for transports there is a huge difference in how those 1s and 0s are received and turned into music

And here is an interesting  engineering guide by Texas Instruments on how to reduce radiated emissions on 100/10 Mbps LAN applications. As you can see not much about 1s and 0s or TCP/IP stacks here... Just good old fashioned electronic engineering and physics.

http://www.ti.com/lit/an/snla107a/snla107a.pdf

Simon

as they say, you can fool some people some of the time but you cant fool all the people all the time. or something like that -- i am probably misquoting. :-)

i dont know whether your view is on cables in general or networking cables in particular -- but in the case of for example, SL IC, to me, in my system, there is absolutely no doubt the difference this single cable has made. The HUGE bonus is that it is a VERY positive difference musically in my active system.

so much so that if i could afford to buy the full loom of  SL cables, i would. This is not to say that high quality cables are unique to Naim -- Chord apparently do some good wires according to extensive reports on this forum -- i dont have any so cant comment.

i am definitely also of the view that there is not much you can do to affect the 0's and 1's in TCP/IP transmission. and the argument that better ethernet cables are possible seemed to me to be snake oil. but now thinking about some of the contributions here -- the 1s and 0s are xmitted over a cable with electrical characteristics -- and these electrical characteristics can definitely affect SQ -- of course i dont know the exact mechanism -- else i would just change jobs and make network cables :-)

very interesting all this -- started off with my noddy questions on behaviour of my switch and now we are dragged again into -- do ethernet cables make a difference? i dont mind. its all great harmless banter.

and all very fascinating indeed...

enjoy

ken

Wat, good points raised and I agree it appears much advantage is taken by some so called audiophile companies  by exploiting many people's lack of knowledge or understanding of physics or electronic engineering.

Ken, as you can hopefully see now from my engineering link from Texas Instruments, there is far more to Ethernet interaction with connected equipment than the transmission of the encoded digital payload. Remember binary data units are simply digitally encoded analogue voltages. It's not magic, it's well understood (by digital electronics and digital transmission engineers) and there are good design principles to minimise the RF interaction. The Ethernet cable can be a source of these RF currents, again read the TI doc for more detail, due to imperfections in it as a transmission line, although radiated emissions of such imperfections can be reduced by screening. But in essence each Ethernet cable design potentially provides variability with RF voltage coupling, emission and interacting. In many applications, this is of no consequence. In applications where it is, such as in sensitive audio equipment, the design considerations discussed by TI (and many others) come into play.

Now where I do take issue is that steps to reduce RFIemmissions and couplings  in cables really is mainly about better compliance of it as a transmission line. This will cost more to manufacture, but need not cost a lot more. There is no magic  or low jitter powder required... Just better compliant twisted pairs... and this is where I agree with Wat, a lack of understanding by many is seemingly taken advantage of by some audiophile companies with ridiculous (in my opinion) sales collateral.

Simon