There is a difference between designing a cable solution for a
specific installation where you know what you are dealing with, and
designing a solution that will always work. (proper Engineering
design)
One of the most critical parameters in what you can "get away with" in
designing cable for T1 transport is the difference in signal level
between the various T1s. A T1 signal is reduced in strength as it
passes through the cable. At its generation, it is typically 6 Volts
peak to peak. This is called 0dbDSX0 (commonly refered to as 0db or
dsx0). It is the reference level used at DSX cross connect points.
In a CO, at a T1 DSX, there can be literally 1000s of T1s running
through cross connect trays unshielded. This only works because they
are all at the same signal level. There is no significant Near End
Cross Talk (NEXT) generated.
This "same level" phenomenon is why you can bundle multiple transmit
signals or multiple receive signals into a common Unshielded Twisted
Pair (UTP - ie: Cat5) cable or in a shelded bundle of a T-Screened
cable (group shielded bundles in a common cable) without worrying
about NEXT.
However, it is not uncommon for received signals on a normal
(repeatered) T1 line to be significantly below the level they were
transmitted at. (T1 signal degrades at a rate of 4db to 6db per 1000
cable ft depending on wire gauge and splices, etc) At the telco
demark (point of hand off of a T1 line at a customer's location) the
signal will be between dsx0 and -16 db DSX0 by engineering standard.
(-16 db is 16 db below the nominal transmit level and is approximately
1V ptp)
Another place where 0dbDSX0 signals are common is if the telco
delivers the T1 by DSL or M13 multiplexer (fiber optics). These
devices output at or near dsx0 level and if your T1 is delivered by
one of these methods, the signals will be at full strength at the
demark.
Generally, if the signal levels are within 7.5 db of one another
through out a cable length, NEXT is not significant, and you can "get
away with" running UTP for T1. However, if a receive signal at -10 db
and a transmit signal at 0 db are run through a UTP for even
relatively short distances, you can expect to get enough cross talk
to cause bipolar violations (errors). This is especially true if
multiple T1s are run through the same bundle. The lower strength
signals will all be corrupted.
In summary, if you know that the signals are at the same levels and
your distances are not too long (less than 500 ft) you can use UTP
and not worrry about Cross Talk. Just be sure to physically separate
UTP cables carrying signals of different levels.
However, if you know that the signals are not the same level, or you
don't know what the signal levels are or how the T1 lines are
engineered, you can design a solution that will always work. This is
why the telco standards for wiring were developed. They don't know
what they will run into at each installation because they have T1s
coming into an office and cannot always plan for the characteritics
of each one. By using individually shielded twisted pair wire, or
T-Screened cable with all of the transmit signals at the same level
in one bundle and the receive signals at the same level in a separate
bundle NEXT can be avoided.
Otherwise you are left with "sometimes it works...."