432 Chapter 14
a 1000Base-T (1GBase-T) and beyond where data is
split between the four pairs and combined at the desti-
nation device.
Where signals are split up and recombined, the
different cables supplying the components will each
have a measurable delay. The trick is for all the compo-
nent cables to have the same delay to deliver their
portions at the same time. The de facto maximum
timing variation in delay for RGB analog is delivery of
all components within 40 ns. Measuring and adjusting
cable delivery is often called timing. By coincidence,
the maximum delay difference in the data world is
45 ns, amazingly close. In the data world, this is called
skew or delay skew, where delivery does not line up.
In the RGB world, where separate coax cables are
used, they have to be cut to the same electrical length.
This is not necessarily the same physical length. Most
often, the individual cables are compared by a Vector-
scope, which can show the relationship between compo-
nents, or a TDR (time domain reflectometer) that can
establish the electrical length (delay) of any cable.
Any difference in physical versus electrical length
can be accounted for by the velocity of propagation of
the individual coaxes, and therefore, the consistency of
manufacture. If the manufacturing consistency is excel-
lent, then the velocity of all coaxes would be the same,
and the physical length would be the same as the elec-
trical length. Where cables are purchased with different
color jackets, to easily identify the components, they are
obviously made at different times in the factory. It is
then a real test of quality and consistency to see how
close the electrical length matches the physical length.
Where cables are bundled together, the installer then
has a much more difficult time in reducing any timing
errors. Certainly in UTP data cables, there is no way to
adjust the length of any particular pair. In all these
bundled cables, the installer must cut and connectorize.
This becomes a consideration when four-pair UTP
data cables (category cables) are used to deliver RGB,
VGA, and other nondata component delivery systems.
The distance possible on these cables is therefore based
on the attenuation of the cables at the frequency of oper-
ation, and on the delay skew of the pairs. Therefore, the
manufacturers measurement and guarantee (if any) of
delay skew should be sought if nondata component
delivery is the intended application.
14.23 Attenuation
All cable has attenuation and the attenuation varies with
frequency. Attenuation can be found with the equation
(14-10)where,
A is the attenuation in dB/100 ft,
Rt is the total dc line resistance in :/100 ft,
H is the dielectric constant of the transmission line insu-
lation,
p is the power factor of the dielectric medium,
f is the frequency,
Zo is the impedance of the cable.Table 14-29 gives the attenuation for various 50:
: ,and 75: cables. The difference in attenuation is
due to either the dielectric of the cable or
center-conductor diameter.14.24 Characteristic ImpedanceThe characteristic impedance of a cable is the measured
impedance of a cable of infinite length. This impedance
is an ac measurement, and cannot be measured with an
ohmmeter. It is frequency-dependent, as can be seen in
Fig. 14-19. This shows the impedance of a coaxial cable
from 10 Hz to 100 MHz.
At low frequencies, where resistance is a major factor,
the impedance is changing from a high value (approxi-
mately 4000: at 10 Hz) down to a lower impedance.
This is due to skin effect (see Section 14.2.8), where the
signal is moving from the whole conductor at low
frequencies to just the skin at high frequencies. There-
fore, when only the skin is carrying the signal, the resis-
tance of the conductor is of no importance. This can be
clearly seen in the equations for impedance, Eq. 14-13,
for low frequencies, shows R, the resistance, as a major
component. For high frequencies, Eq. 14-14, there is no
R, no resistance, even in the equation.
Once we enter that high-frequency area where resis-
tance has no effect, around 100 kHz as shown in Fig.
14-19, we enter the area where the impedance will not
change. This area is called the characteristic impedance
of the cable.
The characteristic impedance of an infinitely long
cable does not change if the far end is open or shorted.
Of course, it would be impossible to test this as it is
impossible to short something at infinity. It is important
to terminate coaxial cable with its rated impedance or a
portion of the signal can reflect back to the input,
reducing the efficiency of the transmission. Reflections
can be caused by an improper load, using a wrong
connector—i.e., using a 50ȍ video BNC connector atA 4.35Rt
Zo-----+= 2.78pf H