SuperUniti review, ethernet cable and shielding

Posted by: gav111n on 06 June 2012

I have just read the HiFi+ review of the SU. Well there are two reviews actually, one by Chris Thomas and one by Malcolm Steward. Both extremely positive, however Steward says something that confuses me:

 

"…if you want the finest performance use a decent quality CAT6 patch lead for connection to your network, particularly if you wish to exploit the SU's hi-definition capabilities. (Be sure the connection between the SU and your NAS/music storage device is unshielded CAT6 from end to end)."

 

Does this chime with anyone else's experience? I would have though that shielded would be better, to protect against interference from any significant RF interference? Or is this just a case of counting how many angels can dance on the head of a pin? 

 

Gav

Posted on: 07 June 2012 by Peter W
Originally Posted by Phil Harris:
Originally Posted by Peter W:

Ooops. I made my own network cables by clamping RJ45s myself. I had to do this because I ran the cables in plastic conduits. The molded plugs of patch cables could not be fed through those conduits. I did buy specialist clamping tool to do the job and paid as much care as I could.

 

We've all made up our own network cables and it would be naive to say that we never do or never will but if you're running cables through conduit / walls / floors then they really should be terminated (punched down) at wall ports at both ends - then patch leads used to connect the wall ports to the network device.

 

The "rule of thumb" is that you should never need to crimp on an RJ45 connector.

 

Phil

Thanks for your comments, Phil.

 

I always thought it would be better to minimise the number of connection points (such as patch cables / wall port interfaces), so I decided to crimp RJ45s myself.

 

Why is it better to terminate cables at wall ports than RJ45s? IMO the chances of doing the terminations badly by amateurs like me are just the same in both cases?

Posted on: 14 June 2012 by james n

I'd be keen to know that as well Peter ( where is Simon when you need him). I would have thought minimising the number of connections would be beneficial to cable performance (minimising mismatches at the junctions) but willing to be proved wrong. I've got two lengths of Cat5e and Cat 6 across the lounge to my study to try a couple of things out but it looks like I'll be re cabling so i'd like to  know if I should terminate at wall ports or just terminate directly with an RJ45 at each end.

 

James

Posted on: 14 June 2012 by chris2000

The wired v wireless is not so much about whether it's too much hassle and mess, rather it's to what extent can you convince the better half that it's not too much hassle and mess!

Posted on: 14 June 2012 by Simon-in-Suffolk

Peter, James, i would say where you terminate your cables is largely down to your situation. Clearly there are cases where you may have wall ports, or floor ports (common in offices) and there is absolutely nothing wrong with that. 

I think the issue was more about grounded Ethernet cables, and here care should be made to ideally in my opinion just ground the cable at one end of it, rather than both ends to avoid ground loops. However if your wall ports are set for grounded Ethernet, and they are not physically earthed , but are grounded back to the master switch then I can't see there being an issue if both ends of the Ethernet lead are grounded in the port.

With Cat 5e and Cat6a non shielded (UTP) these issues wont occur.

Finally I think Phil was suggesting unless you are proficient at it crimping your own Ethernet connectors can be troublesome and proove unreliable. Remember you can incorrectly or poorly wire an Ethernet patch lead such that it still works but only negotiate to half duplex or <1Gbps which unless you know where to look will behidden from you but mysteriously affect performance.

Simon

Posted on: 15 June 2012 by Simon-in-Suffolk

Just read some summaries of the bonding and shielding specs for comms patch leads. It would appear although it is recognised that multiple earth points do produce earth loops usually at the fundamental and harmonic of the nearby mains frequecy. However if the induced voltage is less than 1Vrms or impedance netween earth connections is less than 1ohm then the induced loop current is not sufficient to disrupt the balanced line driver baluns in the network adapter.

So... If you have a properly designed and measured implementation using qualified structured wiring fitters then shielded Ethernet looks fine, essentially you rely on the network adapter to remove the induced earth loop noise.

 

However in the home I would strongly recommend UTP (UN shielded) as there appear many pitfalls with shielded that could degrade over time.

Simon

Posted on: 15 June 2012 by Phil Harris

So - given that shielded network infrastructures should only ever be installed because they are installed in (almost exclusively commercial and medical) environments that are "hostile" and require shielded infrastructure for correct operation and in most (if not all) domestic cases this is not required then can we regard it as safe to summarise this as the following...

 

"*IF* you have a *SHIELDED* network infrastructure then use shielded Ethernet patch leads, if you don't then use unshielded Ethernet patch leads."

 

Cheers

 

Phil

Posted on: 15 June 2012 by gav111n

So it seems that Malcolm Steward was, in general, offering pretty good advice in his HiFi+ review.

 

Gav

Posted on: 15 June 2012 by Phil Harris
Originally Posted by gav111n:

So it seems that Malcolm Steward was, in general, offering pretty good advice in his HiFi+ review.

 

Gav

 

Pretty much.

 

Phil

Posted on: 15 June 2012 by james n

Thanks Simon, Phil - much appreciated.

 

James

 

Posted on: 21 June 2012 by Warpeon

Is an unshielded version of the ethernet cable simply removing the wrapping of a shielded ethernet cable? Will anyone post a photo for reference?

Posted on: 21 June 2012 by Phil Harris
Originally Posted by Warpeon:

Is an unshielded version of the ethernet cable simply removing the wrapping of a shielded ethernet cable? Will anyone post a photo for reference?

 

If you remove the outer of an Ethernet cable then you will destroy the cable geometry and hence the compliance with the required electrical characteristics...

 

Cheers

 

Phil

Posted on: 23 June 2012 by garyi

Hey Phil taking your comments on board i have purchased a patch paneL and 4u data cabinet.

 

The cabinet has an open back but with a locking front door, which implies the open back should actually be affixed to the wall, effectively making a closed design. But with the switch in there I worry about heat generation.

 

Do you know of a website explaining all this stuff, or how do you have yours set up?