Distance to router

Posted by: kiba on 15 June 2011

For my unitiserver I have 15 m to router point. Is it better to have the switch close to Userver and NAS, and then long cable from switch to Router, or should I go for long cable from Userver to Switch + short cable between switch and router. i use a 1GB switch and cat6 Lan cables
Posted on: 15 June 2011 by garyi

Either is good, cat5e will happily transfer over 100m of cable, 15 is nothing.

 

That being said if you can keep the switch and router away from your hifi thats no bad thing as they are usualy resplendent with a crap wall wart

 

Search out CAT5e keystones on ebay and build yourself a nice little wall box for an ethernet point. the switch can be in the loft then or what ever.

Posted on: 15 June 2011 by Simon-in-Suffolk
As Gary says electrical Ethernet is specified to 100m. So no probe what so ever. Agree about RFI rubbish on Ethernet cables, get a ferrite choke/clamp or three on your Ethernet cable to provide a nice 'resistive' load to it.
Simon
Posted on: 15 June 2011 by kiba
Thx
Posted on: 16 June 2011 by kiba

My LAN is Cat6, is a Ferrit clamp still a recomandation ??

Posted on: 16 June 2011 by Simon-in-Suffolk
Yes, it makes no difference to the cable type.
Simon
Posted on: 16 June 2011 by garyi
Or indeed to the music, but fill yer boots.
Posted on: 16 June 2011 by Aleg
Originally Posted by Simon-in-Suffolk:
Yes, it makes no difference to the cable type.
Simon


Would that also go for shielded cable types? (i.e. properly connected on both sides)

Posted on: 16 June 2011 by Simon-in-Suffolk
Aleg, indeed no difference, albeit please read on.... The shield(s) if used around each UTP (un shielded twisted pair) stop common mode pickup on the UTP itself and help reduce cross talk across the UTPs due to inaccuracies in the twisting. The shield around all all the UTPs stop common mode pick up or emission of the combined UTPs.
The RFI typically of concern here is the interference generated from LAN electronics. This carries along all the conductors including shield, hence the term common mode.
One drawback of shielded cable is that although a clamp on inductor will help inhibit the EMI carried on the shield, it won't inhibit the common mode EMI of the twisted pairs in the same way and so shielded cables might be harder to inhibit RFI overall, especially if the source is generating the RF current with respect to earth.

Also one common misconception is that grounding equipment with cable  stops RFI.  This is just  not typically  the case, and frequency/wavelength  and distance to physical  earth (not mains earth - which is there for safety not RF grounding..)  start to have a impact on effectiveness. The best way to stop RF current is to balance it out so it cancels it self or make the RF currentbwork against an inductive (or capacitive)  reactive load (like a resistor to DC).
Simon
Posted on: 17 June 2011 by Aleg
Originally Posted by Simon-in-Suffolk:
Aleg, indeed no difference, albeit please read on.... The shield(s) if used around each UTP (un shielded twisted pair) stop common mode pickup on the UTP itself and help reduce cross talk across the UTPs due to inaccuracies in the twisting. The shield around all all the UTPs stop common mode pick up or emission of the combined UTPs.
The RFI typically of concern here is the interference generated from LAN electronics. This carries along all the conductors including shield, hence the term common mode.
One drawback of shielded cable is that although a clamp on inductor will help inhibit the EMI carried on the shield, it won't inhibit the common mode EMI of the twisted pairs in the same way and so shielded cables might be harder to inhibit RFI overall, especially if the source is generating the RF current with respect to earth.

Also one common misconception is that grounding equipment with cable  stops RFI.  This is just  not typically  the case, and frequency/wavelength  and distance to physical  earth (not mains earth - which is there for safety not RF grounding..)  start to have a impact on effectiveness. The best way to stop RF current is to balance it out so it cancels it self or make the RF currentbwork against an inductive (or capacitive)  reactive load (like a resistor to DC).
Simon

Simon

 

So the RFI generated by the LAN equipment travels 'inside' of the outer-shield of an STP-cable and can therefore not be 'reached' by the RFI-inhibiting effects of the clamp (just to check if I understood correctly ).

 

So, can there be a general advice on cable choice?

Is a properly terminated STP-cable better or worse than an UTP-cable with an RFI-clamp? Does it depend much on RF-surroundings, say WiFi, DECT and microwave?

 

 

Posted on: 17 June 2011 by Aleg

Continued ....... 

 

I read that not all Ferrite clamps are born equal . Especially maximum impedance and 'optimum' frequency differs considerably. Some clamps more focused on power lines and other (high frequency) more focused on LAN and high speed data cables.

 

Would something like this be most appropriate for a Cat5e / Cat 6 cable?

 

Posted on: 17 June 2011 by kiba
If you only have one clamp where to put it? My suggestion A's close A's possible to U.server (input from NAS)
Posted on: 17 June 2011 by Simon-in-Suffolk
Aleg, essentially yes with regard to the shield, and yes ferrite clamps have different properties of the ferrous compound, so they have different inductances, which means different clamps have different impedances for different frequencies.
But in our application here, we are using a blunt tool, and so most clamps will be fine

The best place for the clamp would be prior to the NDX, and additionally you can try right by the LAN switch or 'router' etc.
But remember add more until you hear no difference, and if it makes no difference at all take them off.

For critical industrial applications like in operating theatres special low EMI switchs are used but they are very expensive, and perhaps slightly over the top for our use, so cheap clamps on regular UTP cable (CAT 5) is what I use and recommend.
BTW my teory with wifi and apparent sound degradation is nothing to do with the wireless data network, but more likely due to the RF generated. I
Simon