NEXT crosstalk binder coupling
model from T1E1.4/95-007R2, Annex B (p138). See calculations
which show this NEXT model closely approximates the Bellcore
NEXT coupling model for f > 20 KHz and 18 Kft of 22 AWG pairs.
Approximation to Bellcore NEXT model for n disturbers used for f
< 20 KHz (see ISDNNEXT.MCD) for comparison.
NEXT power increases as cable length
increases up to about 10 dB loss at the frequency of interest,
so a 10 dB loss cable approximates an infinite cable from a NEXT
perspective. Therefore, this NEXT model is applicable for 24 and
26 AWG cable lengths greater than 6 KFt, for f > 20 KHz. NEXT
varies with cable design and insulation type, but is independent
of wire gauge (Valenti (Bellcore), T1E1.4/97-302, page 5).
FEXT model is from T1.413 issue 2
(actually T1E1.4/98-007R1 section B.4.1, pp 177 and 178).
Both NEXT and FEXT models presented here
are 99% worst-case power sum (long-loop) crosstalk.
;f is in
KHz
;n is
number of disturbers
Note the NEXT coupling in dB is a
logarithmic function of n, using the NEXT equation for f > 20
KHz;
;
logarithmic function of n
Incremental NEXT (dNEXT/dn) is
= 2.6057/n
which is the tangential to the
curve.
Cumulative
NEXT is
which can be interpreted as
follows: Knowing the NEXT for n = 1, the n = 2 NEXT is greater
than the n = 1 NEXT by the y-value increase, in this case, is
1.806 dB. Alternatively, given the n = 10 NEXT, the n = 20 NEXT
is graphically observed to be about 7.8 - 6.0 = 1.8 dB greater.
Also note there is only a 10 dB increase from n = 1 (99
percentile disturber) to n = 49 (99 percentile disturbers). The
first few 99 percentile disturbers contribute the most NEXT.
FEXT crosstalk binder coupling
gain model for n, 99% worst-case disturbers from T1E1.4/98-007R1
Annex B.4 (page 177). Channel amplitude transfer function uses a
1/(f^2) to 1/f(^4) relationship:
;dB loss
per Kft at frequency f1, 24 AWG (large diameter wire maximizes
FEXT).
;f is in
KHz
NOTE the following approximate
expression is used for 24 PIC 70 channel transfer function in
the FEXT equation. Closer, but more complex approximation can be
found in CHANNEL.MCD.
;f is in
KHz
;l is in
Kft
Notice that both NEXT and FEXT
increase with n by the same n^0.6 factor.
Comparing FEXT and NEXT (l = 18
Kft) coupling on the same graph, for f > 20 KHz, 26 AWG
;n = number
of disturbers (1, ... , 49)
;f raised
to the 1.5 implies the crosstalk loss is valid for f > 20 KHz,
see Werner HDSL Environment,
IEEE J of
Communications, 1991.
;dB loss
per Kft at frequency f1, 26 AWG
NOTE the following approximate
expression is used for 26 PIC 70 channel transfer function in
the FEXT equation. Closer, but more complex approximation can be
found in CHANNEL.MCD
;f is in
KHz
;f is in
KHz
;range of
length l in Kft, NOTE: 6 < l < 18
Observations: For subscriber
loops greater than 6 Kft, used for NEXT/FEXT comparison,
worst-case NEXT has higher coupling than any FEXT. NEXT is at
least 8 dB greater than FEXT on short loops, and at least 30 dB
greater on long loops. Thus, on long loops where attenuation and
EDD (leading to ISI), and reduced RX power are the major
impairments, FEXT is minimized.
| 
Home
Technology
