Page 86 - Antennas for Base Stations in Wireless Communications

Base Station Antennas for Mobile Radio Systems 59

At the end of this procedure the currents in each element are given
by the vector sum of the currents associated with each excitation that
we have applied. To allow for inaccuracy in the practical achievement
of the desired current, a margin should be left between the calculated
values of nulls and sidelobes and those specified—a margin of 3 dB
should generally be adequate. After a suitable array excitation has been
produced, it may be modified to suit various practical constraints. It is
often convenient to feed adjacent elements with currents of the same
amplitude, so for every element pair the mean of the calculated currents
can be applied to both. After checking the computed pattern following
any adjustments, the required electrical tilt can be arranged by applying
a linear phase shift across the array.
To provide a fixed electrical beamtilt as a function of frequency, we
must delay the currents in lower elements by a fixed time rather than
a fixed phase, as our objective is to tilt the phase-front of the radiated
signal away from the normal to the array; the phase angle associated
with this delay is directly proportional to frequency. This characteristic
is provided by different lengths of transmission line rather than by
fixed phase shifters. If we compute the cable length differences required
at one frequency (typically we do this at midband), the electrical beam-
tilt will, in principle, remain constant over the whole band. Errors in
the element currents can be regarded as random in phase and with a
circular error probable (CEP)—the median length of the error vector
of, say, 0.5 dB; the CEP can be increased and a Monte Carlo analysis
rerun until a significant number of specification infringements occur.
This exercise will indicate the precision that must be achieved in the
practical control of the element currents. As an example of what can be
achieved, Figure 2.11 shows the superimposed measured elevation pat-
terns of a batch of ten 12-element antennas. The degree of consistency
of the result is very good above −25 dB.
A wide variety of computer-based tools is used for network coverage
planning, and unfortunately many of these have adopted inconsistent
conventions regarding the format of the input radiation pattern data.
Most antenna manufacturers provide radiation pattern data files on
their websites and make them freely available in a selection of the more
common commercial program formats.

2.3.3.3 Input Impedance The usual specification for the VSWR at the
input of a base station antenna is 1.4:1 or 1.5:1. As well as regarding the
antenna as a single unit with an input VSWR specification, we need to
consider it as a complex assembly of radiating elements, power dividers,
and interconnecting transmission lines, and must consider the control
of impedance matching at each stage between the radiating elements
and the array input. When designing the intermediate components,