[c-nsp] T1 Cables

[Ted Mittelstaedt]

> -----Original Message-----
> From: cisco-nsp-bounces at puck.nether.net
> [mailto:cisco-nsp-bounces at puck.nether.net]On Behalf Of Jay Hennigan
> Sent: Friday, May 18, 2007 3:02 PM
> To: Janet Plato
> Cc: cisco-nsp at puck.nether.net
> Subject: Re: [c-nsp] T1 Cables
> 
> 
> Janet Plato wrote:
> > DS1 cabling does not use the normal USOC pairing, so you forgo
> > common mode rejection of noise.  
> 
> Sure it does, USOC RJ48.  This is also the standard EIA pairing, same as 
> just about everything else.  Pairs 1 and 3 of 568A or 1 and 2 of 568B. 
> Regular "ethernet" jumpers are also EIA 568, so the pairing is correct, 
> although one of the active pairs for DS1 is different from 10-base-T or 
> 100-base-T.
> 
> There are several different USOC pairing schemes for 8-pin plug/jack 
> combinations.  USOC RJ48 *IS* the DS1 standard.  USOC RJ45 (which is 
> what the physical jack/plug format is often erroneously called) is 
> something completely different, no it's not ethernet, but the pairing is 
> still compatible.  USOC RJ61 is VERY rare, pairing NOT compatible.  USOC 
> RJ31X also not compatible, used for alarm jacks.
> 
> http://www.hotlinkdata.com/wiring_diagram.htm
> 
> > When a signal pin and that
> > signal's ground share a common path with normal spacing, for
> > example in a single twisted pair, they pick up the same amount of
> > noise from the environment.  The noise is common to both wires,
> > and since the receiver is expecting to see the signal expressed as
> > a difference in voltage between signal pin and ground pin, not
> > an absolute voltage difference, the noise that is common to both
> > wires is rejected.
> 
> But T-1 is a balanced differential signal, so there is no signal pin and 
> ground but instead two signal pins per direction of transmission. 
> Regular "ethernet whips" indeed have the correct wires in twisted pairs 
> for T-1, pins 1 and 2 are a pair and pins 4 and 5 are a pair.
> 
> > Having said that, I've run T1 short distances in a normal data
> > center environment on some really crappy wire without problems.
> > I suspect you either have a problem not related to your cabling,
> > or enough noise that you really need the common-mode noise
> > rejection afforded by correct pairing.
> 
> The pairing of generic 8-pin UTP data cables is correct for T-1.  I 
> agree that the original poster's issue is not with using conventional 
> "ethernet whip" jumpers for DS1.  He may indeed have a layer one 
> problem, but EIA 568A or 568B jumpers (as long as both ends are the 
> same) aren't it.
> 
> > I do not have a source of such cables, I've not used such things
> > in a long time.
> 
> You may be thinking of the old-school individually shielded twisted pair 
> 22-gauge T-1 cable.  This was a special low-capacitance cable with two 
> pairs, insulated with a large diameter foam plastic dielectric, each 
> pair individually shielded.  This cable was designed to carry T-1 
> signals for thousands of feet without repeaters or CSU/DSU equipment.
> 
> Occasionally I've seen a telco use this stuff as a ten-foot jumper from 
> a protector terminating a half-mile or so of 600-pair copper to a 
> smartjack on the backboard.  I guess this gave them a warm fuzzy feeling 
> that at least the last ten feet was cabled better than the first 
> half-mile.
> 
> This cable pre-dates the use of 8-pin modular plugs for T-1 
> terminations.  It was sometimes used with D-sub 15 pin plugs and jacks. 
>   It is very difficult to terminate it correctly into an 8-pin modular 
> plug because the insulation is too thick.

Go to any wire distributor and ask for the "klingon bird of prey"
plug, this is an RJ45 plug that has a back end that is 2 pieces that
use a screw to hold them together.

Ted