Chapter 5

Point Sources and

Their Arrays

Topics in this chapter include:

Point source radiators Linear arrays of n point sources

Power patterns Null directions and beam widths
Isotropic sources Non uniform amplitude distributions
Radiation intensity Dolph-Tschebycheff (D-T) or optimum
Examples of power patterns field patterns distribution
Field patterns D-T distribution for an array of 8-point sources
Phase patterns A comparison of amplitude distributions
Arrays of two point sources Continuous arrays
Pattern multiplication Huygen’s principle
Pattern synthesis Diffraction by flat sheet
Non-isotropic dissimilar sources Rectangular broadside arrays

5–1 Introduction
In chap. 2 an antenna was treated as an aperture. In this chapter an antenna is first considered as a point source
and later the concept is extend to the formation of arrays of point source. This approach is of great value since
the pattern of any antenna can be regarded as produced by an array of point sources. Further the initial
discussion relating to arrays is confined to isotropic point sources, which may represent different kinds of
antennas. Later the discussion is extended to encompass more general case of non-uniform distribution.
With the information of this chapter and the computer programs on the book’s website, an arrays producing
almost any desired patterns may be designed.

5–2 Point Source Defined

At a sufficient distance in the far field of an antenna, the radiated fields of the antenna are transverse and the
power flow or Poynting vector (Wm−2 ) is radial as at the point O at a distance R on the observation circle
in Fig. 5−1. It is convenient in many analyses to assume that the fields of the antenna are everywhere of this
type. In fact, we may assume, by extrapolating inward along the radii of the circle, that the waves originate
at a fictitious volumeless emitter, or point source, at the center O of the observation circle. The actual field
variation near the antenna, or “near field,” is ignored, and we describe the source of the waves only in terms

86
5–3 Power Patterns 87

of the “far field” it produces. Provided that our observations are made at a sufficient distance, any antenna,
regardless of its size or complexity, can be represented in this way by a single point source.
Instead of making field measurements around the observation circle with the antenna fixed, the equivalent
effect may be obtained by making the measurements at a fixed point Q on the circle and rotating the antenna
around the center O. This is usually the more convenient procedure if the antenna is small.
In Fig. 5−1a, the center O of the antenna coincides with the center of the observation circle. If the center
of the antenna is displaced from O, even to the extent that O lies outside the antenna as in Fig. 5−1b, the
distance d between the two centers has a negligible effect on the field patterns at the observation circle,
provided R  d, R  b, and R  λ. However, the phase patterns will generally differ, depending on d. If
d = 0, the phase shift around the observation circle is usually a minimum. As d is increased, the observed
phase shift becomes larger.

Antenna Antenna

O O d
b b

R Observation R
circle

Q Q
(a) (b)

Figure 5–1 Antenna and observation circle.

As discussed in Sec. 2−3, a complete description of the far field of a source requires three patterns: two
patterns of orthogonal field components as a function of angle [Eθ (θ, φ) and Eφ (θ, φ)] and one pattern of the
phase difference of these fields as a function of angle [δ (θ, φ)]. For many purposes, however, such a complete
knowledge is not necessary. It may suffice to specify only the variation with angle of the power density or
Poynting vector magnitude (power per unit area) from the antenna [Sr (θ, φ)]. In this case the vector nature of
the field is disregarded, and the radiation is treated as a scalar quantity. This is done in Sec. 5−3. The vector
nature of the field is recognized later in the discussion on the magnitude of the field components. Although the
cases considered as examples in this chapter are hypothetical, they could be approximated by actual antennas.

5–3 Power Patterns

Let a transmitting antenna in free space be represented by a point-source radiator located at the origin of the
coordinates in Fig. 5−2 (see also Fig. 2−5). The radiated energy streams from the source in radial lines. The
time rate of energy flow per unit area is the Poynting vector, or power density (watts per square meter). For a
point source (or in the far field of any antenna), the Poynting vector S has only a radial component Sr with
no components in either the θ or φ directions (Sθ = Sφ = 0). Thus, the magnitude of the Poynting vector, or
power density, is equal to the radial component (|S| = Sr ).
 88 Chapter 5 Point Sources and Their Arrays

z z
Polar Polar
axis axis r sin ␪ d␾
S r ( radial component r d␪
Element of area
of Poynting vector ds at radius r
or power density, r sin ␪  r 2 sin ␪ d␪ d␾
W m2, at radius r) Area ds subtends
(r, ␪, ␾) a solid angle
␪ r ␪
Point source Point source
r  sin ␪ d␪ d␾
at origin
y y
␾ ␾

x Equatorial plane x

(a) (b)

Figure 5–2 Spherical coordinates for a point source of radiation in free space.

A source that radiates energy uniformly in all directions is an isotropic source. For such a source the radial
component Sr of the Poynting vector is independent of θ and φ. A graph of Sr at a constant radius as a
function of angle is a Poynting vector, or power-density, pattern, but is usually called a power pattern. The
three-dimensional power pattern for an isotropic source is a sphere. In two dimensions the pattern is a circle
(a cross section through the sphere), as suggested in Fig. 5−3.
Although the isotropic source is convenient in theory, it is ␪0
not a physically realizable type. Even the simplest antennas
have directional properties, i.e., they radiate more energy in ␪
some directions than in others. In contrast to the isotropic Sr
source, they might be called anisotropic sources. As an
example, the power pattern of such a source is shown in
Fig. 5−4a where Srm is the maximum value of Sr .
If Sr is expressed in watts per square meter, the graph is an
absolute power pattern. On the other hand, if Sr is expressed Figure 5–3 Polar power pattern of
in terms of its value in some reference direction, the graph is isotropic source.
a relative power pattern. It is customary to take the reference
direction such that Sr is a maximum. Thus, the pattern radius for relative power is Sr /Srm where Srm is the
maximum value of Sr . The maximum value of the relative power pattern is unity, as shown in Fig. 5−4b. A
pattern with a maximum of unity is also called a normalized pattern.
␪0 ␪0
POWER RELATIVE
1 POWER

Srm Sr
Srm
Sr

(a) (b)

Figure 5–4 (a) Power pattern and (b) relative power pattern for same source. Both patterns
have the same shape. The relative power pattern is normalized to a maximum of unity (1).
5–5 Radiation Intensity 89

5–4 A Power Theorem1 and its Application to an Isotropic Source

If the Poynting vector is known at all points on a sphere of radius r from a point source in a lossless medium,
the total power radiated by the source is the integral over the surface of the sphere of the radial component
Sr of the average Poynting vector. Thus,
 
P =  S · ds =  Sr ds (1)

where
P = power radiated, W
Sr = radial component of average Poynting vector, W m−2
ds = infinitesimal element of area of sphere (see Fig. 3−2b)
= r 2 sin θ dθ dφ, m2
For an isotropic source, Sr is independent of θ and φ so

P = Sr  ds = Sr × 4πr 2 (W) (2)

and
P
Sr = (W m−2 ) (3)
4πr 2
Equation (3) indicates that the magnitude of the Poynting vector varies inversely as the square of the distance
from a point-source radiator. This is a statement of the well-known law for the variation of power per unit
area as a function of the distance.

5–5 Radiation Intensity

As discussed in Sec. 2−5, the radiation intensity U is expressed in watts per unit solid angle (W sr−1 ). The
radiation intensity is independent of radius. It is power per steradian. From (5−4−3) we have
r 2 Sr = P /4π = U (W/sr) (1)
Thus, the power theorem may be restated as follows:
The total power radiated is given by the integral of the radiation intensity over a solid angle of 4π steradians.
As already mentioned in Sec. 2−5, power patterns can be expressed in terms of either the Poynting vector
(power density) or the radiation intensity. A power pattern in terms of U is the same as in Fig. 5−4a with the
maximum Poynting vector (Sm ) replaced by the maximum radiation intensity (Um ) and the Poynting vector

1 This theorem is a special case of a more general relation for the complex power flow through any closed surface as given by

1
P =  (E × H∗ ) · ds (1)
2
where P is the total complex power flow and E and H∗ are complex vectors representing the electric and magnetic fields, H∗ being the
complex conjugate of H. The average Poynting vector is
1
S= Re(E × H∗ ) (2)
2
Now the power flow in the far field is entirely real; hence, taking the real part of (1) and substituting (2), we obtain the special case of (3).
 90 Chapter 5 Point Sources and Their Arrays

as a function of r(Sr ) replaced by the radiation intensity as a function of r(Ur ). The maximum value of Um
is in the θ = 0 direction. Relative Poynting vector (or power density) and relative radiation intensity patterns
are identical.
Applying (1) to an isotropic source gives

P = 4πU0 (W)

where U0 = radiation intensity of isotropic source, W sr −1 .

5–6 Examples of Power Patterns

EXAMPLE 5–6.1 Source with Unidirectional Cosine Power Pattern

A source has a cosine radiation-intensity pattern, that is,

U = Um cos θ (1)

where Um = maximum radiation intensity

The radiation intensity U has a value only in the upper hemisphere (0 ≤ θ ≤ π/2 and 0 ≤ φ ≤ 2π )
and is zero in the lower hemisphere. The radiation intensity is a maximum at θ = 0. The pattern is shown
in Fig. 5−5. The space pattern is a figure of revolution of this circle around the polar axis. Find the
directivity.
To find the total power radiated by the cosine source, we apply (5−4−1) and integrate only over the
upper hemisphere. Thus
 2π  π/2
P = Um cos θ sin θ dθ dφ = πUm (2)
0 0

If the power radiated by the unidirectional cosine source is the same as for an isotropic source, then (2)
and (1) in Sec. 5−5 may be set equal, yielding

πUm = 4πU0

or
Um
Directivity = =4=D Ans (3)
U0

␪0
Thus, the maximum radiation Polar
intensity Um of the unidirectional axis
cosine source (in the direction θ =0) 1
is 4 times the radiation intensity U0
from an isotropic source radiating the 0.5
same total power. The power patterns
Half-power
for the two sources are compared in points
Fig. 5−6 for the same total power
radiated by each.
Figure 5–5 Unidirectional cosine power pattern.
5–6 Examples of Power Patterns 91

EXAMPLE 5–6.2 Source with Bidirectional Cosine Power Pattern

A source has a cosine power pattern that is bidirectional. Find the directivity. With radiation in two
hemispheres instead of one; the maximum radiation intensity is half its value in Example 5−6.1. Thus,
from (3),
D = 4/2 = 2 Ans.

EXAMPLE 5–6.3 Source with Sine (Doughnut) Power Pattern

A source has a radiation intensity pattern given by
U = Um sin θ (4)
The pattern is shown in Fig. 5−6. The space pattern is a
figure-of-revolution of this pattern around the polar axis and ␪0
has the form of a doughnut. Find D. Polar
axis
Solution 4
Applying (5−4−3) the total power radiated is 3 Cosine power
pattern
 2π  π 2
P = Um sin2 θ dθ dφ = π 2 Um (5)
0 0 1
If the power radiated by this source is the same as for an Isotropic
isotropic source taken as reference, we have power pattern

π 2 Um = 4πU0 (6) Figure 5–6

Power patterns of
and
cosine and
Um 4 isotropic sources.
Directivity = = = 1.27 = D Ans. (7)
U0 π

EXAMPLE 5–6.4 Source with Sine-Squared (Doughnut) Power Pattern

A source has a sine-squared radiation-intensity power pattern. The radiation-intensity pattern is given by

U = Um sin2 θ (8)
The power pattern is shown in Fig. 5−7a. This type of pattern is of considerable interest because it is
the pattern produced by a short dipole coincident with the polar (θ = 0) axis in Fig. 5−7a. Applying
(5−4−3), the total power radiated is
 2π  π
8
P = Um sin3 θ dθ dφ = πUm (9)
0 0 3
If P is the same as for the isotropic source,
8
πUm = 4πU0
3
 92 Chapter 5 Point Sources and Their Arrays

␪0
␪0

(a) (b)

Figure 5–7 (a) Sine-squared power pattern and (b) unidirectional cosine-squared power
pattern.
and
Um 3
Directivity = = = 1.5 = D Ans. (10)
U0 2

EXAMPLE 5–6.5 Source with Unidirectional Cosine-Squared Power Pattern

A source with a unidirectional cosine-squared radiation-intensity pattern is given by
U = Um cos2 θ (11)
The radiation intensity has a value only in the upper hemisphere as in Fig. 5−7b. The three-dimensional
or space pattern is a figure-of-revolution of this pattern around the polar (θ = 0) axis. Find the directivity.
Solution
The total power radiated is
 2π  π/2
2
P = Um cos2 θ sin θ dθ dφ = πUm (12)
0 0 3
If P is the same as radiated by an isotropic source,
2
πUm = 4πU0
3
and
Um
Directivity = =6=D Ans. (13)
U0
Thus, the maximum power per unit solid angle (at θ = 0) from the source with the cosine-squared power
pattern is six times the power per unit solid angle from an isotropic source radiating the same power.

Directivities are summarized in Table 5−1.

Example 5−6.6 provides some valuable insights into the effect minor lobes have on directivity or gain. Without
minor lobes the gain of this antenna would be 91.4 or 19.6 dBi as compared to a gain of 18.0 or 12.6 dBi with
minor lobes. The minor lobes have large beam or solid angles because they extend 360◦ in the azimuth or φ
direction at large sin θ values (θ near 90◦ ). The main lobe, on the other hand, is at small θ angles so the Pn (θ )
sin θ product is small, in fact, zero at θ = 0◦ .
5–6 Examples of Power Patterns 93

Table 5–1 Directivities of the Point Source Patterns in Examples 5−6.1 to 5−6.5.
Pattern Directivity
Unidirectional cosine 4
Bidirectional cosine 2
Sine doughnut 1.27
Sine-squared doughnut 1.5
Unidirectional cosine squared 6

EXAMPLE 5–6.6 Pencil Beam with Minor Lobes

As shown in Fig. 5−8, the pattern has pencil beam (symmetrical around the θ = 0 axis) with a main-lobe
HPBW of approximately 22◦ and four minor lobes. Find the directivity.
Solution
The directivity is given by
4π
D =  2π  π (14)
0 0 Pn (θ ) sin θ dθ dφ
where the denominator equals the total beam area A .
Since the pattern is symmetrical (no variation with φ), the integral with respect to φ yields 2π and
(14) reduces to
4π
D= π (15)
2π 0 Pn (θ ) sin θ dθ
We have only the pattern graph available (no analytical expression), so let us divide the pattern (Fig. 5−8)
into 36 steps of 5◦ each. The approximate value of the integral in the first (m = 1) 5◦ section (= π/36 rad)
is given by
π π 1.0 + 0.93
Pn (θ1 )av sin θ1 = sin 2.5◦ (16)
36 36 2
and the approximate directivity is then given by the summation of all 36 sections or by
4π
D (17)

m=36
2π(π/36) Pn (θm )av sin θm
m=1
Completing the summation, we obtain
4π 4π 72
D=  = = 18.0 (18)
A 2π(π/36)(0.25 + 0.37 + 0.46 + 0.12 + 0.07) 1.27π
Main First Second Third Fourth
lobe minor minor minor minor
lobe lobe lobe lobe
(back
lobe)

or D  12.6 dBi
It is noteworthy that the second minor lobe contributes most to the total beam area, the first minor lobe
almost as much, and the main lobe less than either. Thus, the directivity is greatly affected by the minor
 94 Chapter 5 Point Sources and Their Arrays

Figure 5–8 Power patterns of beam antenna in polar plot at (a) and in rectangular plot
at (b). The large shaded area A of (b) is for an isotropic source while the antenna area a
appears as a series of small areas. The directivity D = A / a
lobes, which is a common situation with actual antennas. For this antenna pattern the beam efficiency is
given by
0.25
εM = = 0.20 (19)
1.27
If the second minor lobe were eliminated, the directivity would increase to 14.5 dBi (up 1.9 dB) and if
both first and second minor lobes were eliminated, the directivity would increase to 17.1 dBi (up 4.5 dB).

The directivity obtained in Example 5−6.6 is approximate. By sufficiently reducing the step size (5◦ in
the example), the summation can be made as precise as the available data will allow. Computation of this
numerical integration can be facilitated by using a computer.
The half-power beamwidth of the pattern in the example is about 22◦ . Taking kp = 1 and εM as in (19),
the approximate directivity is then
41,000εM 41,000 × 0.2
D = = 16.9 or 12.3 dBi (20)
kp × HPBW2 (22◦ )2
which is 0.3 dB less than obtained by the 36-step summation.
The beam area of an isotropic source equals 4π steradians. In Fig. 5−8b this corresponds to the area A
under the sin θ curve. The beam area of the source in Example 5−6.6 corresponds to the area a under the
Pn (θ ) sin θ curve. Thus, the directivity is simply A/a or the ratio of the area of the isotropic source to the
area of the source being measured. Hence,
4π A
D= = (21)
A a
 100 Chapter 5 Point Sources and Their Arrays

Solution
The fields north and south are horizontally polarized (in the plane of the page). The fields east and west
are vertically polarized. At 45◦ or NE the field is right circularly polarized (RCP). At 135◦ or SE the
field is left circularly polarized (LCP). At 225◦ or SW the field is again right circularly polarized. Finally,
at 315◦ or NW the field is again left circularly polarized. At intermediate angles the field is elliptically
polarized.

5–9 Arrays of Two Isotropic Point Sources

Let us introduce the subject of arrays of point sources by considering the simplest situation, namely, that of
two isotropic point sources. As illustrations, five cases involving two isotropic point sources are discussed.

Case 1. Two Isotropic Point Sources of Same Amplitude and Phase

The first case we shall analyze is that of two isotropic point sources having equal amplitudes and oscillating in
the same phase. Let the two point sources, 1 and 2, be separated by a distance d and located symmetrically with
respect to the origin of the coordinates as shown in Fig. 5−15a. The angle φ is measured counterclockwise

Figure 5–15 (a) Relation to coordinate system of two isotropic point sources separated by a
distance d. (b) Vector addition of the fields from two isotropic point sources of equal amplitude
and same phase located as in (a). (c) Field pattern of two isotropic point sources of equal
amplitude and same phase located as in (a) for the case where the separation d = λ/2.

from the positive x axis. The origin of the coordinates is taken as the reference for phase. Then at a distant
point in the direction φ the field from source 1 is retarded by 12 dr cos φ, while the field from source 2 is
advanced by 12 dr cos φ, where dr is the distance between the sources expressed in radians; that is,
2π d
dr = = βd
λ
5–9 Arrays of Two Isotropic Point Sources 101

The total field at a large distance r in the direction φ is then

E = E0 e−j ψ/2 + E0 e+j ψ/2 (1)
where ψ = dr cos φ and the amplitude of the field components at the distance r is given by E0 .
The first term in (1) is the component of the field due to source 1 and the second term the component due
to source 2. Equation (1) may be rewritten
e+j ψ/2 + e−j ψ/2
E = 2E0 (2)
2
which by a trigonometric identity is
 
ψ dr
E = 2E0 cos = 2E0 cos cos φ (3)
2 2
This result may also be obtained with the aid of the vector diagram1 shown in Fig. 5−15b, from which (3)
follows directly. We note in Fig. 5−15b that the phase of the total field E does not change as a function of ψ.
To normalize (3), that is, make its maximum value unity, set 2E0 = 1. Suppose further that d is λ/2. Then
dr = π . Introducing these conditions into (3) gives
 
π
E = cos cos φ (4)
2
The field pattern of E versus φ as expressed by (4) is presented in Fig. 5−15c. The pattern is a bidirectional
figure-of-eight with maxima along the y axis. The space pattern is doughnut-shaped, being a figure-of-
revolution of this pattern around the x axis.
The same pattern can also be obtained by locating source 1 at the origin of the coordinates and source 2
at a distance d along the positive x axis as indicated in Fig. 5−16a. Taking now the field from source 1 as
reference, the field from source 2 in the direction φ is advanced by dr cos φ. Thus, the total field E at a large
distance r is the vector sum of the fields from the two sources as given by
E = E0 + E0 e+j ψ (5)
where ψ = dr cos φ.
The relation of these fields is indicated by the vector diagram of Fig. 5−16b. From the vector diagram the
magnitude of the total field is
ψ dr cos φ
E = 2E0 cos = 2E0 cos (6)
2 2
as obtained before in (3). The phase of the total field E is, however, not constant in this case but is ψ/2, as
also shown by rewriting (5) as
 j ψ/2 
e + e−j ψ/2 ψ
E = E0 (1 + ej ψ ) = 2E0 ej ψ/2 = 2E0 ej ψ/2 cos (7)
2 2
Normalizing by setting 2E0 = 1, (7) becomes
ψ ψ
E = ej ψ/2 cos = cos \ψψ/2 (8)
2 2

1 It
is to be noted that the quantities represented here by vectors are not true space vectors but merely vector representations of the time
phase (i.e., phasors).
 102 Chapter 5 Point Sources and Their Arrays

E0 ej ␺ (from source 2)

o s␾
dc ␺/2
␾ ␺/2
1 d 2 x ␾0
E0 (from source 1)

(a) (b)

90
Rotation around source 1
Rotation around center point of array
␺/2 0

90
0 90 180 270 360
␾
(c)

Figure 5–16 (a) Two isotropic point sources with the origin of the coordinate system
coincident with one of the sources. (b) Vector addition of the fields from two isotropic point
sources of equal amplitude and same phase located as in (a). (c) Phase of total field as a
function of φ for two isotropic point sources of same amplitude and phase spaced λ/2 apart.
The phase change is zero when referred to the center point of the array but is ψ/2 as shown by
the dashed curve when referred to source 1.

In (8) the cosine factor gives the amplitude variation of E, and the exponential or angle factor gives the phase
variation with respect to source 1 as the reference. The phase variation for the case of λ/2 spacing (dr = π ) is
shown by the dashed line in Fig. 5−16c. Here the phase angle with respect to the phase of source 1 is given by
ψ/2 = (π/2) cos φ. The magnitude variation for this case has already been presented in Fig. 5−15c. When
the phase is referred to the point midway between the sources (Fig. 5−15a), there is no phase change around
the array as shown by the solid line in Fig. 5−16c. Thus, an observer at a fixed distance observes no phase
change when the array is rotated (with respect to φ) around its midpoint, but a phase change (dashed curve of
Fig. 5−16c) is observed if the array is rotated with source 1 as the center of rotation.

Case 2. Two Isotropic Point Sources of Same Amplitude but Opposite Phase

This case is identical with the one we have just considered except that the two sources are in opposite phase
instead of in the same phase. Let the sources be located as in Fig. 5−15a. Then the total field in the direction
φ at a large distance r is given by
E = E0 e+j ψ/2 − E0 e−j ψ/2 (9)
from which
 
ψ dr
E = 2j E0 sin = 2j E0 sin cos φ (10)
2 2
5–9 Arrays of Two Isotropic Point Sources 103

whereas in Case 1 (3) involves the cosine of ψ/2, (10) for Case 2 involves the sine. Equation (10) also includes
an operator j , indicating that the phase reversal of one of the sources in Case 2 results in a 90◦ phase shift of
the total field as compared with the total field for Case 1. This is unimportant here. Thus, putting 2j E0 = 1
and considering the special case of d = λ/2, (10) becomes
π
E = sin cos φ (11)
2
The directions φm of maximum field are obtained by setting the argument of (11) equal to ±(2k + 1)π/2.
Thus,
π π
cos φm = ±(2k + 1) (11a)
2 2
where k = 0, 1, 2, 3 . . . . For k = 0, cos φm = ±1 and φm = 0◦ and 180◦ .
The null directions φ0 are given by
π
cos φ0 = ±kπ (11b)
2
For k = 0, φ0 = ±90◦ .
The half-power directions are given by
π π
cos φ = ±(2k + 1) (11c)
2 4
For k = 0, φ = ±60◦ , ±120◦ .
The field pattern given by (11) is shown in Fig. 5−17. The pattern is a relatively broad figure-of-eight
with the maximum field in the same direction as the line joining the sources (x axis). The space pattern is a
figure-of-revolution of this pattern around the x axis. The two sources, in this case, may be described as a
simple type of “end-fire” array. In contrast to this pattern, the in-phase point sources produce a pattern with
the maximum field normal to the line joining the sources, as shown in Fig. 5−15c. The two sources for this
case may be described as a simple “broadside” type of array.

90
60

30

ⴚ ⴙ 0
1 2

␭
2

Figure 5–17 Relative field pattern for two isotropic point sources of the same amplitude but
opposite phase, spaced λ/2 apart.
 104 Chapter 5 Point Sources and Their Arrays

Case 3. Two Isotropic Point Sources of the Same Amplitude and In-Phase Quadrature

Let the two point sources be located as in Fig. 5−15a. Taking the origin of the coordinates as the reference
for phase, let source 1 be retarded by 45◦ and source 2 advanced by 45◦ .
Then the total field in the direction φ at a large distance r is given by
   
dr cos φ π dr cos φ π
E = E0 exp +j + + E0 exp −j + (12)
2 4 2 4
From (12) we obtain
 
π dr
E = 2E0 cos + cos φ (13)
4 2
Letting 2E0 = 1 and d = λ/2, (13) becomes
 
π π
E = cos + cos φ (14)
4 2
The field pattern given by (14) is presented in Fig. 5−18. The space pattern is a figure-of-revolution of this
pattern around the x axis. Most of the radiation is in the second and third quadrants. It is interesting to note
that the field in the direction φ = 0◦ is the same as in the direction φ = 180◦ . The directions φm of maximum
field are obtained by setting the argument of (14) equal to kπ, where k = 0, 1, 2, 3 . . . . In this way we obtain
π π
+ cos φm = kπ (15)
4 2
For k = 0,
π π
cos φm = − (16)
2 4

90
120
60

150 30

180 ⴚ ⴙ 0
1 2
␭
2

Figure 5–18 Relative field pattern of two isotropic point sources of the same amplitude and
in phase quadrature for a spacing of λ/2. The source to the right leads that to the left by 90◦ .
5–9 Arrays of Two Isotropic Point Sources 105

and
φm = 120◦ and 240◦ (17)
If the spacing between the sources is reduced to λ/4, (13) becomes
 
π π
E = cos + cos φ (18)
4 4
The field pattern for this case is illustrated by Fig. 5−19a. It is a cardioid-shaped, unidirectional pattern with
maximum field in the negative x direction. The space pattern is a figure-of-revolution of this pattern around
the x axis.
A simple method of determining the direction of maximum field is illustrated by Fig. 5−19b. As indicated
by the vectors, the phase of source 2 is 0◦ (vector to right) and the phase of source 1 is 270◦ (vector down).
Thus, source 2 leads source 1 by 90◦ .
To find the field radiated to the left, imagine that we start at source 2 (phase 0◦ ) and travel to the left, riding
with the wave (phase 0◦ ) like a surfer rides a breaker. The phase of the wave we are riding is 0◦ and does not
change but by the time we have traveled λ/4 and arrived at source 1, a 14 -period has elapsed so the current
in source 1 will have advanced 90◦ (vector rotated ccw) from 270◦ to 0◦ , making its phase the same as that
of the wave we are riding, as in the middle diagram of Fig. 5−19b. Thus, the field of the wave from source
2 reinforces that of the field of source 1, and the two fields travel to the left together in phase producing a
maximum field to the left which is twice the field of either source alone.
Now imagine that we start at source 1 with phase 270◦ (vector down) and travel to the right. By the time
we arrive at source 2 the phase of its field has advanced from 0 to 90◦ so it is in phase opposition and cancels
the field of the wave we are riding, as in the bottom diagram in Fig. 5−19b, resulting in zero radiation to the
right.

Case 4. General Case of Two Isotropic Point Sources of Equal Amplitude and Any Phase Difference

Proceeding now to a more general situation, let us consider the case of two isotropic point sources of equal
amplitude but of any phase difference δ. The total phase difference ψ between the fields from source 2 and
source 1 at a distant point in the direction φ (see Fig. 5−16a) is then

ψ = dr cos φ + δ (19)

Taking source 1 as the reference for phase, the positive sign in (19) indicates that source 2 is advanced in
phase by the angle δ. A minus sign would be used to indicate a phase retardation. If, instead of referring the
phase to source 1, it is referred to the centerpoint of the array, the phase of the field from source 1 at a distant
point is given by −ψ/2 and that from source 2 by +ψ/2. The total field is then
ψ
E = E0 (ej ψ/2 + e−j ψ/2 ) = 2E0 cos (20)
2
Normalizing (20), we have the general expression for the field pattern of two isotropic sources of equal
amplitude and arbitrary phase,
ψ
E = cos (21)
2
where ψ is given by (19). The three cases we have discussed are obviously special cases of (21). Thus, Cases
1, 2, and 3 are obtained from (21) when δ = 0◦ , 180◦ , and 90◦ respectively.
 106 Chapter 5 Point Sources and Their Arrays

90
120 60

150

30

180 1 2 x axis
0

4

(a)

x
1  2
x
4
Going from 2 to 1
field doubles
(b) Going from 1 to 2
field is zero (a null)

Figure 5–19 (a) Relative field pattern of two isotropic sources of same amplitude and
in-phase quadrature for a spacing of λ/4. Source 2 leads source 1 by 90◦ . (b) Vector diagrams
illustrating field reinforcement in the −x direction and field cancellation in the +x direction.

Case 5. Most General Case of Two Isotropic Point Sources of Unequal Amplitude and Any Phase
Difference

A still more general situation, involving two isotropic point sources, exists when the amplitudes are unequal
and the phase difference is arbitrary. Let the sources be situated as in Fig. 5−20a with source 1 at the origin.
Assume that the source 1 has the larger amplitude and that its field at a large distance r has an amplitude of E0 .
Let the field from source 2 be of amplitude aE0 (0 ≤ a ≤ 1) at the distance r. Then, referring to Fig. 5−20b,
the magnitude and phase angle of the total field E is given by

E = E0 (1 + a cos ψ)2 + a 2 sin2 ψ [a sin ψ/(1 + a cos ψ)] (22)

where ψ = dr cos φ + δ and the phase angle (\ ) is referred to source 1. This is the phase angle ξ shown
in Fig. 5−20b.
5–10 Nonisotropic but Similar Point Sources and the Principle of Pattern Multiplication 107

E
aE0

  
1 d 2 x E0

(a) (b)

Figure 5–20 (a) Two isotropic point sources of unequal amplitude and arbitrary phase with
respect to the coordinate system. (b) Vector addition of fields from unequal sources arranged as
in (a). The amplitude of source 2 is assumed to be smaller than that of source 1 by the factor a.

5–10 Nonisotropic but Similar Point Sources and the Principle of Pattern
Multiplication
The cases considered in the preceding section all involve isotropic point sources. These can readily be extended
to a more general situation in which the sources are nonisotropic but similar.
The word similar is here used to indicate that the variation with absolute angle φ of both the amplitude
and phase of the field is the same.1 The maximum amplitudes of the individual sources may be unequal. If,
however, they are also equal, the sources are not only similar but are identical.
As an example, let us reconsider Case 4 of Sec. 5−9
in which the sources are identical, with the modification y
that both sources 1 and 2 have field patterns given by
Short
dipoles
E0 = E0 sin φ (1)

1 2
Patterns of this type might be produced by short x
dipoles oriented parallel to the x axis as suggested by d
Fig. 5−21. Substituting (1) in (5−9−20) and normalizing
by setting 2E0 = 1 gives the field pattern of the array as Figure 5–21 Two nonisotropic sources
with respect to the coordinate system.
ψ
E = sin φ cos (2)
2
where ψ = dr cos φ + δ
This result is the same as obtained by multiplying the pattern of the individual source (sin φ) by the pattern
of two isotropic point sources (cos ψ/2).
If the similar but unequal point sources of Case 5 (Sec. 5−9) have patterns as given by (1), the total
normalized pattern is

E = sin φ (1 + a cos ψ)2 + a 2 sin2 ψ (3)

Here again, the result is the same as that obtained by multiplying the pattern of the individual source by the
pattern of an array of isotropic point sources.

1 The patterns not only must be of the same shape but also must be oriented in the same direction to be called “similar.”
 108 Chapter 5 Point Sources and Their Arrays

These are examples illustrating the principle of pattern multiplication, which may be expressed as follows:

The field pattern of an array of nonisotropic but similar point sources is the product of the pattern of
the individual source and the pattern of an array of isotropic point sources having the same locations,
relative amplitudes, and phase as the nonisotropic point sources.

This principle may be applied to arrays of any number of sources provided only that they are similar. The
individual nonisotropic source or antenna may be of finite size but can be considered as a point source situated
at the point in the antenna to which phase is referred. This point is said to be the “phase center.”
The above discussion of pattern multiplication has been concerned only with the field pattern or magnitude
of the field. If the field of the nonisotropic source and the array of isotropic sources vary in phase with space
angle, i.e., have a phase pattern which is not a constant, the statement of the principle of pattern multiplication
may be extended to include this more general case as follows:

The total field pattern of an array of nonisotropic but similar sources is the product of the individual source pattern
and the pattern of an array of isotropic point sources each located at the phase center of the individual source and
having the same relative amplitude and phase, while the total phase pattern is the sum of the phase patterns of the
individual source and the array of isotropic point sources.

The total phase pattern is referred to the phase center of the array. In symbols, the total field E is then
E = f (θ, φ)F (θ, φ) fp (θ, φ) + Fp (θ, φ) (4)
Field pattern Phase pattern

where
f (θ, φ) = field pattern of individual source
fp (θ, φ) = phase pattern of individual source
F (θ, φ) = field pattern of array of isotropic sources
Fp (θ, φ) = phase pattern of array of isotropic sources
The patterns are expressed in (4) as a function of both polar angles to indicate that the principle of pattern
multiplication applies to space patterns as well as to the two-dimensional cases we have been considering.
To illustrate the principle, let us apply to it two special modifications of Case 1 (Sec. 5−9).

EXAMPLE 5–10.1 Assume two identical point sources separated by a distance d, each source having
the field pattern given by (1) as might be obtained by two short dipoles arranged as in Fig. 5−21. Let
d = λ/2 and the phase angle δ = 0. Then the total field pattern is
 
π
E = sin φ cos cos φ (5)
2

This pattern is illustrated by Fig. 5−22c as the product of the individual source pattern (sin φ) shown
at (a) and the array pattern {cos[(π/2) cos φ]} as shown at (b). The pattern is sharper than it was in Case
1 (Sec. 5−9) for the isotropic sources. In this instance, the maximum field of the individual source is
in the direction φ = 90◦ , which coincides with the direction of the maximum field for the array of two
isotropic sources.
5–10 Nonisotropic but Similar Point Sources and the Principle of Pattern Multiplication 109

Figure 5–22 Example of pattern multiplication. Two nonisotropic but identical point
sources of the same amplitude and phase, spaced λ/2 apart and arranged as in Fig. 5−21,
produce the pattern shown at (c). The individual source has the pattern shown at (a), which,
when multiplied by the pattern of an array of two isotropic point sources (of the same
amplitude and phase) as shown at (b), yields the total array pattern of (c).

EXAMPLE 5–10.2 Let us consider next the situation y

in which d = λ/2 and δ = 0 as in Example 5−10.1
but with individual source patterns given by Short
E0 = E0 cos φ (6) dipoles
This type of pattern might be produced by short 
1 2
dipoles oriented parallel to the y axis as in Fig. 5−23. x
Here the maximum field of the individual source is 
in the direction (φ = 0) of a null from the array, 2
while the individual source has a null in the direc-
tion (φ = 90◦ ) of the pattern maximum of the array. Figure 5–23 Array of two nonisotropic
By the principle of pattern multiplication the total sources with respect to the coordinate system.
normalized field is

(a) (b) (c)

Figure 5–24 Example of pattern multiplication. Total array pattern (c) as the product of
pattern (a) of individual nonisotropic source and pattern (b) of array of two isotropic sources.
The pattern (b) for the array of two isotropic sources is identical with that of Fig. 5−22b, but
the individual source pattern (a) is rotated through 90◦ with respect to the one in Fig. 5−22a.
5–13 Linear Arrays of n Isotropic Point Sources of Equal Amplitude and Spacing 115

f (φ) = relative field pattern of source 1

fp (φ) = phase pattern of source 1
F (φ) = relative field pattern of source 2
Fp (φ) = phase pattern of source 2
In (1) the phase angle (/ ) is referred to the phase of the field from source 1 in some reference direction
(φ = φ0 ).
In the special case where the field patterns are identical but
the phase patterns are not, a = 1, and
y
f (φ) = F (φ) (4)
from which
ψ 
E = 2E0 f (φ) cos fp (φ) + ψ/2 (5)
2 x
where phase is again referred to source 1 in some reference 1 2
direction φ0 . d
As an illustration of nonisotropic, dissimilar point sources,
let us consider an example in which the field from source 1 is Figure 5–29 Relation of two
given by nonisotropic dissimilar sources to
E1 = cos φ/0 (6) coordinate system.
and from source 2 by
E2 = sin φ ψ (7) 90

where ψ = dr cos φ + δ
The relation of the two sources to the coordinate system
and the individual field patterns is shown in Fig. 5−29.
Source 1 is located at the origin. The total field E is then
the vector sum of E1 and E2 , or 
180 1 2 0
E = cos φ + sin φ ψ (8)
Let us consider the case for λ/4 spacing (d = λ/4) and
phase quadrature of the sources (δ = π/2). Then
π
ψ = (cos φ + 1) (9)
2
The calculation for this case is easily carried out by graphi- 270

cal vector addition. The resulting field pattern for the total
field E of the array is presented in Fig. 5−30, and the Figure 5–30 Field pattern of array of
resulting phase pattern for the angle ξ is given in Fig. 5−31. two nonisotropic dissimilar sources of
The angle ξ is the phase angle between the total field and Fig. 5−29 for d = λ/4 and δ = 90◦ .
the field of source 1 in the direction φ = 0.

5–13 Linear Arrays of n Isotropic Point Sources of Equal Amplitude and

Spacing
Introduction
Let us now proceed to the case of n isotropic point sources of equal amplitude and spacing arranged as a linear
array, as indicated in Fig. 5−32, where n is any positive integer. The total field E at a large distance in the
direction φ is given by
E = 1 + ej ψ + ej 2ψ + ej 3ψ + · · · + ej (n−1)ψ (1)
 116 Chapter 5 Point Sources and Their Arrays

360

Total phase angle, 

270

180

90

0
0 90 180 270 360


Figure 5–31 Phase pattern of array having field pattern of Fig. 5−30. The phase angle ξ is
with respect to source 1 as phase center.
where ψ is the total phase difference of the fields from adjacent sources as given by
2πd
ψ= cos φ + δ = dr cos φ + δ (2)
λ
where δ is the phase difference of adjacent sources, i.e.,   90
source 2 with respect to 1, 3 with respect to 2, etc.   0
(Schelkunoff-1, Stratton-1).
The amplitudes of the fields from the sources are all
To distant point
equal and taken as unity. Source 1 (Fig. 5−32) is the 
phase reference. Thus, at a distant point in the direction

φ the field from source 2 is advanced in phase with respect os 
dc
to source 1 by ψ, the field from source 3 is advanced in 0
phase with respect to source 1 by 2ψ, etc. d d d
Equation (1) is a geometric series. Each term repre- 1 2 3 4 5 n
sents a phasor, and the amplitude of the total field E and
Figure 5–32 Arrangement of linear
its phase angle ξ can be obtained by phasor (vector) addi-
array of n isotropic point sources.
tion as in Fig. 5−33. Analytically, E can be expressed in
a simple trigonometric form which we now develop as
follows:


E E 5 5
 4
4 
 3 E 0 E
5
 
4 3
5 3   2 1


4


 2 2 1 2 3 
  1 
 
1
(a) (b)
Figure 5–33 (a) Vector addition of fields at a large distance from the linear array of five
isotropic point sources of equal amplitude with source 1 as the phase center (reference for
phase). (b) Same, but with midpoint of array (source 3) as phase center.
5–13 Linear Arrays of n Isotropic Point Sources of Equal Amplitude and Spacing 117

Multiply (1) by ej ψ , giving

Eej ψ = ej ψ + ej 2ψ + ej 3ψ + · · · + ej nψ (3)
Now subtract (3) from (1) and divide by 1 − ej ψ , yielding
1 − ej nψ
E= (4)
1 − ej ψ
Equation (4) may be rewritten as
 
ej nψ/2 ej nψ/2 − e−j nψ/2
E = j ψ/2 (5)
e ej ψ/2 − e−j ψ/2
from which
sin(nψ/2) sin(nψ/2)
E = ej ξ = ξ (6)
sin(ψ/2) sin(ψ/2)
where ξ is referred to the field from source 1. The value of ξ is given by
n−1
ξ= ψ (7)
2
If the phase is referred to the centerpoint of the array, (6) becomes
sin(nψ/2)
E= (8)
sin(ψ/2)
In this case the phase pattern is a step function as given by the sign of (8). The phase of the field is constant
wherever E has a value but changes sign when E goes through zero.
When ψ = 0, (6) or (8) is indeterminate so that for this case E must be obtained as the limit of (8) as ψ
approaches zero. Thus, for ψ = 0 we have the relation that
E=n (8a)
This is the maximum value that E can attain. Hence, the normalized value of the total field for Emax = n is
1 sin(nψ/2)
E= (9)
n sin(ψ/2)
The field as given by (9) will be referred to as the “array factor.” Values of the array factor as obtained from (9)
for various numbers of sources are presented in Fig. 5−34. If ψ is known as a function of φ, then the field
pattern can be obtained directly from Fig. 5−34.
We may conclude from the above discussion that the field from the array will be a maximum in any direction
φ for which ψ = 0. Stated in another way, the fields from the sources all arrive at a distant point in the same
phase when ψ = 0. In special cases, ψ may not be zero for any value of φ, and in this case the field is usually
a maximum at the minimum value of ψ.
To illustrate some of the properties of linear arrays (9) will now be applied to several special cases. See
programs on the book’s web site involving these different cases. See also discussion in Appendix C.

Case 1. Broadside Array (Sources in Phase)

The first case is a linear array of n isotropic sources of the same amplitude and phase. Therefore, δ = 0 and
ψ = dr cos φ (10)
 118 Chapter 5 Point Sources and Their Arrays

1.0
n1
.9
n2
.8

.7 n3
| Array factor |

.6 n4
.5
n5
.4
n  10
3
.3 n  20 ⴚ
4
5
20 10 ⴚ ⴚ 5
.2 ⴚ ⴚ
20 10 ⴙ
10
.1 ⴙ 20 ⴙ ⴚ 10
ⴚ 20 20 20 ⴙ20
20 ⴙ20 ⴚ ⴙ ⴚ ⴚ ⴚ
0
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180
360 350 340 330 320 310 300 290 280 270 260 250 240 230 220 210 200 190 180



Figure 5–34 Universal field-pattern chart for arrays of various numbers n of isotropic point
sources of equal amplitude and spacing.
To make ψ = 0 requires that φ = (2k +1)(π/2), where k = 0, 1, 2, 3, . . . . The field is, therefore, a maximum
when
π 3π
φ= and (10a)
2 2
That is, the maximum field is in a direction normal to the array. Hence, this condition, which is characterized
by in-phase sources (δ = 0), results in a “broadside” type of array.
As an example, the pattern of a broadside array of four in-phase isotropic point sources of equal amplitude
is shown in Fig. 5−35a. The spacing between sources is λ/2.1 The field pattern in rectangular coordinates
and the phase patterns for this array are presented in Fig. 5−35a.

Case 2. Ordinary End-Fire Array

Let us now find the phase angle between adjacent sources that is required to make the field a maximum in the
direction of the array (φ = 0). An array of this type may be called an “end-fire” array. For this we substitute
the conditions ψ = 0 and φ = 0 into (2), from which
δ = −dr (11)
Hence, for an end-fire array, the phase between sources is retarded progressively by the same amount as the
spacing between sources in radians. Thus, if the spacing is λ/4, source 2 in Fig. 5−32 should lag source 1 by
90◦ , source 3 should lag source 2 by 90◦ , etc.
As an example, the field pattern of an end-fire array of four isotropic point sources is presented in Fig. 5−36a.
The spacing between sources is λ/2 and δ = −π. The field pattern in rectangular coordinates and the phase
patterns are shown in Fig. 5−36b. The same shape of field pattern is obtained in this case if δ = +π since,
with d = λ/2, the pattern is bidirectional. However, if the spacing is less than λ/2, the maximum radiation
is in the direction φ = 0 when δ = −dr and in the direction φ = 180◦ when δ = +dr .

1 If
the spacing between elements exceeds λ, sidelobes appear which are equal in amplitude to the main (center) lobe. These are called
grating lobes (see Sec. 19−6)
5–13 Linear Arrays of n Isotropic Point Sources of Equal Amplitude and Spacing 119

  90
1

|E|
0
2 01  90
d
2

Total phase angle, 

0
d d d Phase center
180   0 90 of midpoint
1 2 3 4
of array
Array 180

270 Phase center

at source 1
Field
360
pattern
450
0 90 180 270 360

(a) (b)
Figure 5–35 (a) Field pattern of broadside array of four isotropic point sources of the same
amplitude and phase. The spacing between sources is λ/2. (b) Field pattern in rectangular
coordinates and phase patterns of same array with phase center at midpoint and at source 1.
The reference direction for phase is at φ = 90◦ .

Figure 5–36 (a) Field pattern of ordinary end-fire array of four isotropic point sources of
same amplitude. Spacing is λ/2 and the phase angle δ = −π . (b) Field pattern in rectangular
coordinates and phase patterns of same array with phase center at midpoint and at source 1.
The reference direction for phase is at φ = 0.

Case 3. End-Fire Array with Increased Directivity

The situation discussed in Case 2, namely, for δ = −dr , produces a maximum field in the direction φ = 0
but does not give the maximum directivity. It has been shown by Hansen (1) and Woodyard that a larger
directivity is obtained by increasing the phase change between sources so that
 
π
δ = − dr + (12)
n
 120 Chapter 5 Point Sources and Their Arrays

This condition will be referred to as the condition for “increased directivity.” Thus for the phase difference
of the fields at a large distance we have
π
ψ = dr (cos φ − 1) − (13)
n
As an example, the field pattern of an end-fire array of four isotropic point sources for this case is illustrated
in Fig. 5−37.
The spacing between sources is λ/2, and
therefore δ = −(5π/4). Hence, the conditions 90
are the same as for the array with the pattern
of Fig. 5−36, except that the phase difference
between sources is increased by π/4. Compar-
ing the field patterns of Figs. 5−36a and 5−37,
it is apparent that the additional phase difference
yields a considerably sharper main lobe in the 180 0
direction φ = 0. However, the back lobes in this
case are excessively large because the large value
of spacing results in too great a range in ψ. 
d
To realize the directivity increase afforded by 2
the additional phase difference requires that |ψ|
be restricted in its range to a value of π/n at d d d
φ = 0 and a value in the vicinity of π at φ = 180◦ .
1 2 3 4
This can be fulfilled if the spacing is reduced. For
Array
example, the field pattern of an end-fire array of
10 isotropic point sources of equal amplitude and
Figure 5–37 Field pattern of end-fire array of
spaced λ/4 apart is presented in Fig. 5−38a for
four isotropic point sources of equal amplitude
the phase condition giving increased directivity
spaced λ/2 apart. The phasing is adjusted for
(δ = −0.6π). In contrast to this pattern, one is
increased directivity (δ = − 54 π).
presented in Fig. 5−38b for the identical antenna
with the phasing of an ordinary end-fire array (δ = −0.5π ). Both patterns are plotted to the same maximum.
The increased directivity is apparent from the greater sharpness of the pattern. Integrating the pattern, including
the minor lobes, the directivity is found to be about 19 and of the ordinary endfire about 11. The beamwidths
and directivities for the two cases are compared in Table 5−2.

Table 5–2 Comparison of end-fire arrays

Ordinary end-fire End-fire array with
array increased directivity
Beamwidth between half-power points 69◦ 38◦
Beamwidth between first nulls 106◦ 74◦
Directivity 11 19

The maximum of the field pattern of Fig. 5−38a occurs at φ = 0 and ψ = −π/n. In general, any increased
directivity end-fire array, with maximum at ψ = −π/n, has a normalized field pattern given by
 
π sin(nψ/2)
E = sin (14)
2n sin(ψ/2)
5–13 Linear Arrays of n Isotropic Point Sources of Equal Amplitude and Spacing 121

Phase shift 180ⴗ (0.6␲) versus 90ⴗ (0.5␲)

Increased-directivity
end fire Ordinary end fire
D ⴝ 19
or 12.8 dBi D ⴝ 11
or 10.4 dBi

Array

(a) (b)

Figure 5–38 Field patterns of end-fire arrays of 10 isotropic point sources of equal amplitude
spaced λ/4 apart. The pattern at (a) has the phase adjusted for increased directivity
(δ = −0.6π), while the pattern at (b) has the phasing of an ordinary end-fire array (δ = −0.5π ).

Case 4. Array with Maximum Field in an Arbitrary Direction. Scanning Array

Let us consider the case of an array with a field pattern having a maximum in some arbitrary direction φ1 not
equal to kπ/2 where k = 0, 1, 2, or 3. Then (2) becomes
0 = dr cos φ1 + δ (15)
By specifying the spacing dr , the required phase difference δ is then determined by (15). Conversely, by
changing δ the beam direction φ1 can be shifted or scanned.
As an example, suppose that n = 4, d = λ/2, and that we wish to have a maximum field in the direction of
φ = 60◦ . Then δ = −π/2, yielding the field pattern shown in Fig. 5−39.
 Chapter 6
Electric Dipoles, Thin
Linear Antennas and
Arrays of Dipoles and
Apertures

Topics in this chapter include:

Fields of Short dipole Radiation efficiency

Radiation resistance of Short dipole Arrays of n driven elements
Thin linear antenna: λ/2, λ, and 3 λ/2
Horizontal and vertical antennas above ground
Radiation resistance of λ/2 dipole
Shaped dipole arrays
Arrays of two λ/2 dipoles
Phased arrays
Broadside, end-fire and close-spaced arrays
Grid and chain arrays
Radiation resistance at a point which is not a
Digital beam-forming or adaptive or smart
current maximum
arrays
Traveling wave antennas
Long-wire antennas: V, Rhombic and
Arrays of two driven elements; broadside and Beverage
end-fire
Curtain arrays
Patterns, driving point impedance and gain
Arrays feed points
Arrays of two driven elements; general case
Folded dipoles
Closely spaced elements

6–1 Introduction
This chapter first develops the concepts of electric dipoles and thin linear antennas. Later it is extended to
the arrays of dipoles and apertures. The essential background for this later part is covered in chap. 5 on point
sources and their arrays. The dipoles referred to herein are mostly thin linear dipoles, whereas the apertures
in general may be helices, horns, big reflectors or arrays of dipoles (arrays of arrays).
The far or radiation field pattern, the driving point impedance and the array gains are first derived in that
order for several different arrays of dipoles. The method of analysis is general and applicable to other dipole
arrays; the specific types discussed are merely examples. Array gain is calculated by treating the dipoles as
circuit elements having self and mutual impedances. Although direct pattern integration could be used to
determine the gain, the circuit approach is simpler provided impedance values are available (patterns having
been utilized in the impedence calculations).

156
6–3 The Fields of a Short Dipole 157

6–2 The Short Electric Dipole

Since any linear antenna may be considered as consisting of a large number of very short conductors connected
in series, it is of interest to examine first the radiation properties of short conductors. From a knowledge of
properties of short conductors, we can then proceed to a study of long linear conductors such as are commonly
employed in practice.
A short linear conductor is often called a short dipole. In the following discussion, a short dipole is always
of finite length even though it may be very short. If the dipole is vanishingly short, it is an infinitesimal dipole.
Let us consider a short dipole such as shown in Fig. 6–1a.
The length L is very short compared to the wavelength (L  λ). End plate provides
Plates at the ends of the dipole provide capacitive loading. loading with little
The short length and the presence of these plates result in a effect on pattern
uniform current I along the entire length L of the dipole. The q
dipole may be energized by a balanced transmission line, as d
shown. It is assumed that the transmission line does not radiate
and, therefore, its presence will be disregarded. Radiation from I L L I
the end plates is also considered to be negligible. The diameter Transmission
d of the dipole is small compared to its length (d  L). Thus, line
q
for purposes of analysis we may consider that the short dipole
appears as in Fig. 6–1b. Here it consists simply of a thin con- (a) (b)
ductor of length L with a uniform current I and point charges
q at the ends. The current and charge are related by Figure 6–1 A short dipole
antenna (a) and its equivalent (b).
dq
=I (1)
dt
6–3 The Fields of a Short Dipole
Let us now proceed to find the fields everywhere around a short
dipole. Let the dipole of length L be placed coincident with
z
the z axis and with its center at the origin as in Fig. 6–2. The
relation of the electric field components, Er , Eθ and Eφ , is then Er
as shown. It is assumed that the medium surrounding the dipole
E
is air or vacuum. P
In dealing with antennas or radiating systems, the propaga- Dipole 
tion time is a matter of great importance. Thus, if a current is r
flowing in the short dipole of Fig. 6–3, the effect of the current is
not felt instantaneously at the point P , but only after an interval L
y
E
equal to the time required for the disturbance to propagate over
the distance r. We have already recognized this in Chap. 5 in 
connection with the pattern of arrays of point sources, but here
we are more explicit and describe it as a retardation effect.
x
Accordingly, instead of writing the current I as1
Figure 6–2 Relation of dipole to
I = I0 ej ωt (1) coordinates.

1 It is assumed that we take either the real (cos ωt) or imaginary (sin ωt) part of ej ωt .
 158 Chapter 6 Electric Dipoles, Thin Linear Antennas and Arrays of Dipoles and Apertures

which implies instantaneous propagation of the effect of the z

current, we introduce the propagation (or retardation) time
P
as done by Lorentz and write
[I ] = I0 ej ω[t−(r/c)] (2)
s
where [I ] is called the retarded current. Specifically, 1
Dipole
the retardation time r/c results in a phase retardation s
ωr/c = 2πf r/c radians = 360◦ f r/c = 360◦ t/T , where  r
T = 1/f = time of one period or cycle (seconds) and f = dz s2
frequency (hertz, Hz = cycles per second). The brackets may z
be added as in (2) to indicate explicitly that the effect of the y
L 0
current is retarded.
Equation (2) is a statement of the fact that the disturbance
at a time t and at a distance r from a current element is caused
by a current [I ] that occurred at an earlier time t − r/c. The d
time difference r/c is the interval required for the distur- Figure 6–3a Geometry for short dipole.
bance to travel the distance r, where c is the velocity of light
(= 300 Mm s−1 ).
Electric and magnetic fields can be expressed in terms of vector and scalar potentials. Since we will be
interested not only in the fields near the dipole but also at distances which are large compared to the wavelength,
we must use retarded potentials, i.e., expressions involving t − r/c. For a dipole located as in Fig. 6–2 or
Fig. 6–3a, the retarded vector potential of the electric current has only one component, namely, Az . Its value
is 
μ0 L/2 [I ]
Az = dz (2)
4π −L/2 s
where [I ] is the retarded current given by
[I ] = I0 ej ω[t−(s/c)] (3a)
In (3) and (3a),
z = distance to a point on the conductor
I0 = peak value in time of current (uniform along dipole)
μ0 = permeability of free space = 4π × 10−7 H m−1
If the distance from the dipole is large compared to its length (r  L) and if the wavelength is large compared
to the length (λ  L), we can put s = r and neglect the phase differences of the field contributions from
different parts of the wire. The integrand in (3) can then be regarded as a constant, so that (3) becomes
μ0 LI0 ej ω[t−(r/c)]
Az = (4)
4πr
The retarded scalar potential V of a charge distribution is

1 [ρ]
V = dτ (5)
4πε0 V s
where [ρ] is the retarded charge density given by
[ρ] = ρ0 ej ω[t−(s/c)] (6)
6–3 The Fields of a Short Dipole 159

and dτ = infinitesimal volume element

ε0 = permittivity or dielectric constant of free space = 8.85 × 10−12 F m−1
Since the region of charge in the case of the dipole being considered is confined to the points at the ends
as in Fig. 6–1b, (5) reduces to
 
1 [q] [q]
V = − (7)
4πε0 s1 s2

From (6–1–1) and (3a),

  To
[I ] point
[q] = [I ] dt = I0 ej ω[t−(s/c)] dt = (8) s1
P
jω
s
L co
Substituting (8) into (7), 2
 j ω[t−(s1 /c)]  r
I0 e ej ω[t−(s2 /c)] 
V = − (9)
4πε0 j ω s1 s2 L
s2
Referring to Fig. 6–3b, when r  L, the lines connecting
the ends of the dipole and the point P may be considered
as parallel so that s
Dipole L co
L 2
s1 = r − cos θ (10)
2
Figure 6–3b Relations for short
and dipole when r  L.
L
s2 = r + cos θ (11)
2
Substituting (10) and (11) into (9), it may be shown that the fields of a short electric dipole are:

 
I0 L cos θej ω[t−(r/c)] 1 1
Er = + (12)
Electric fields 2πε0 cr 2 j ωr 3 General
of short dipole   case
I0 L sin θej ω[t−(r/c)] j ω 1 1 (13)
Eθ = + +
4πε0 c2 r cr 2 j ωr 3

In obtaining (12) and (13) the relation was used that μ0 ε0 = 1/c2 , where c = velocity of light.
Turning our attention now to the magnetic field, this may be calculated from curl of A as follows:

   
r̂ ∂(sin θ )Aφ ∂(Aθ ) θˆ ∂Ar ∂(r sin θ)Aφ
∇ ×A = − + −
r sin θ ∂θ ∂φ r sin θ ∂φ ∂r
 
φ̂ ∂(rAθ ) ∂Ar
+ − (14)
r ∂r ∂θ
Since Aφ = 0, the first and fourth terms of (14) are zero, since Ar and Aθ are independent of φ, so that the
second and third terms of (14) are also zero. Thus, only the last two terms contribute, so that ∇ × A, and
hence also H, have only a φ component. Thus,
 160 Chapter 6 Electric Dipoles, Thin Linear Antennas and Arrays of Dipoles and Apertures

 
I0 L sin θej ω[t−(r/c)] j ω 1 (15)
Magnetic fields |H| = H = +
r 2 General
φ
4π cr
of short dipole case
Hr = Hθ = 0 (16)

Thus, the fields from the dipole have only three components Er , Eθ and Hφ . The components Eφ , Hr and Hθ
are everywhere zero.
When r is very large, the terms in 1/r 2 and 1/r 3 in (12), (13), and (15) can be neglected in favor of the
terms in 1/r. Thus, in the far field Er is negligible, and we have effectively only two field components, Eθ
and Hφ , given by

j ω[t−(r/c)]
Electric and E = j ωI0 L sin θ e =j
I0 βL
sin θej ω[t−(r/c)] (17)
θ 2
magnetic 4πε0 c r 4π ε0 cr Far-field
fields of j ωI0 L sin θ e j ω[t−(r/c)] I0 βL case
short dipole Hφ = =j sin θej ω[t−(r/c)] (18)
4πcr 4π r

Taking the ratio of Eθ to Hφ as given by (17) and (18), we obtain

Eθ 1 μ0
= = = 376.7 Impedance of space (19)
Hφ ε0 c ε0

This is the intrinsic impedance of free space (a pure resistance). It is a very important constant.
Comparing (17) and (18) we note that Eθ and Hφ are in time phase in the far field. We note also that the field
patterns of both are proportional to sin θ. The pattern is independent of φ, so that the space pattern is doughnut-
shaped, being a figure-of-revolution of the pattern in Fig. 6–4a about the axis of the dipole. Referring to the
near-field expressions given by (12), (13) and (15), we note that for
a small r the electric field has two components Er and Eθ , which
are both in time-phase quadrature with the magnetic field, as in
a resonator. At intermediate distances, Eθ and Er can approach
time-phase quadrature so that the total electric field vector rotates
in a plane parallel to the direction of propagation, thus exhibiting
Dipole Dipole
the phenomenon of cross-field. For the Eθ and Hφ components,
the near-field patterns are the same as the far-field patterns, being (a) (b)
proportional to sin θ (Fig. 6–4a). However, the near-field pattern
for Er is proportional to cos θ as indicated by Fig. 6–4b. The space Figure 6–4 Near- and far-field
pattern for Er is a figure-of-revolution of this pattern around the patterns of Eθ and Hφ
dipole axis. components for short dipole (a)
Let us now consider the situation at very low frequencies. This and near-field pattern of Er
will be referred to as the quasi-stationary, or dc case. Since from component (b).

[I ] = I0 ej ω[t−(r/c)] = j ω[q] (20)

6–3 The Fields of a Short Dipole 161

(12) and (13) can be rewritten as

 
[q]L cos θ j ω 1
Er = + (21)
2πε0 cr 2 r3
and
 
[q]L sin θ ω2 jω 1
Eθ = − 2 + 2+ 3 (22)
4πε0 c r cr r
The magnetic field is given by (15) as
 
[I ]L sin θ j ω 1
Hφ = + 2 (23)
4π cr r
At low frequencies, ω approaches zero so that the terms with ω in the numerator can be neglected. As ω → 0,
we also have
[q] = q0 ej ω[t−(r/c)] = q0 (24)
and
[I ] = I0 (25)
Thus, for the quasi-stationary, or dc, case, the field components become from (21), (22) and (23)
q0 L cos θ
Er = (26)
2πε0 r 3
Electric and magnetic q0 L sin θ
E = Low-frequency case (27)
fields of short dipole θ 4πε0 r 3
I0 L sin θ (28)
Hφ =
4πr 2

The restriction that r  L still applies.

The expressions for the electric field, (26) and (27), are identical to those obtained in electrostatics for the
field of two point charges, +q0 and −q0 , separated by a distance L. The relation for the magnetic field, (28),
may be recognized as the Biot-Savart relation for the magnetic field of a short element carrying a steady
or slowly varying current. Since in the expressions for the quasi-stationary case the fields decrease as 1/r 2
or 1/r 3 , the fields are confined to the vicinity of the dipole and there is negligible radiation. In the general
expressions for the fields, (21), (22) and (23), it is the 1/r terms which are important in the far field and hence
take into account the radiation.
The expressions for the fields from a short dipole developed above are summarized in Table 6–1.
Setting
1
|A| =
2rλ
1
|B| =
4πrλ2
1
|C| =
8π 2 rλ3
 162 Chapter 6 Electric Dipoles, Thin Linear Antennas and Arrays of Dipoles and Apertures

Table 6–1 Fields of a short electric dipole†

GeneraI Quasi-
Component expression Far field stationary
 
[I ]L cos θ 1 1 q 0 L cos θ
Er + 0
2πε0 c r2 j ωr 3 2πε0 r 3
 
[I ]L sin θ j ω 1 1 [I ]Lj ω sin θ j 60π [I ] sin θ L q 0 L sin θ
Eθ 2
+ 2 + 3
=
4πε0 c r cr j ωr 4πε0 c 2 r r λ 4πε0 r 3
 
[I ]L sin θ j ω 1 [I ]Lj ω sin θ j [I ] sin θ L I0 L sin θ
Hφ + 2 =
4π cr r 4πc r 2r λ 4πr 2
†The restriction applies that r  L and λ  L. The quantities in the table are in SI units, that is, E in volts per meter, H in amperes per meter, I in
amperes, r in meters, etc. [I ] is as given by (20). Three of the field components of an electric dipole are everywhere zero, that is,
Eφ = Hr = H θ = 0

for the three components of Eθ , their variation with distance is as shown in Fig. 6–5. For rλ greater than the
radian distance [1/(2π)], component A of the electric field is dominant, for rλ less than the radian distance
component C of the electric field is dominant, while at the radian distance only B contributes (= π ) because
although |A| = |B| = |C| = π, A and C are in phase opposition and cancel.

100
Energy Energy mostly
mostly radiated
stored r E

10
Components A, B, C of E

1/r 2␭ term
 (near field)

1
1/r␭ term
|A| (radiated
|B| or far field)
|C|

0.1 1/r3␭ term

(nearest field)
Radian
distance

0.01
0.01 0.1 1 1 10 100
2
r/

Figure 6–5 Variation of the magnitudes of the components of Eθ of a short electric dipole as
a function of distance (r/λ). The magnitudes of all components equal π at the radian distance
1/(2π ). At larger distances energy is mostly radiated, at smaller distances mostly stored.
6–3 The Fields of a Short Dipole 163

For the special case where θ = 90◦ (perpendicular to the dipole in the xy plane of Fig. 6–2) and at
rλ  1/(2π),
I0 Lλ
|Hφ | = (A m−1 ) (29)
2r
while at rλ  1/(2π),
I0 L
|Hφ | = (30)
4πr 2
which is identical to the relation for the magnetic field perpendicular to a short linear conductor carrying
direct current as given by (28).
The magnetic field at any distance r from an infinite linear conductor with direct current is given by
I0
Hφ = (31)
2πr
which is Ampere’s law.
Remarkably, the magnitude of the magnetic field in the equatorial plane (θ = 90◦ ) in the far field of an
oscillating λ/2 dipole is identical to (31) (Ampere’s law). It is assumed that the current distribution on the
λ/2 dipole is sinusoidal. This is discussed in more detail in Sec. 6–5.
Rearranging the three field components of Table 6–1 for a short electric dipole, we have

[I ]Lλ Z cos θ 1 1
Er = 2
−j 2 3 (32)
λ 2πrλ 4π rλ

[I ]Lλ Z sin θ 1 1 1
Eθ = j + 2
−j 2 3 (33)
λ 2rλ 4πrλ 8π rλ

[I ]Lλ sin θ 1 1
Hφ = j + (34)
λ 2rλ 4πrλ2

We note that the constant factor in each of the terms in brackets differs from the factors of adjacent terms by
a factor of 2π .
At the radian distance (rλ = 1/2π) the fields of (32), (33) and (34) reduce to
√
2 2π[I ]Lλ Z cos θ
Er = − 45◦ (35)
λ
π[I ]Lλ Z sin θ
Eθ = (36)
λ
√
2π [I ]Lλ sin θ
Hφ = 45◦ (37)
λ
The magnitude of the average power flux or Poynting vector in the θ direction is given by
Sθ = 1
2 Re Er Hφ∗ = 12 Er Hφ Re 1 − 90◦ = 12 Er Hφ cos(−90◦ ) = 0 (38)
indicating that no power is transmitted. However, the product Er Hφ represents imaginary or reactive energy
that oscillates back and forth from electric to magnetic energy twice per cycle.
 164 Chapter 6 Electric Dipoles, Thin Linear Antennas and Arrays of Dipoles and Apertures

In like manner the magnitude of the power flux or Poynting vector in the r direction is given by
1
Sr = 21 Eθ Hφ cos(−45◦ ) = √ Eθ Hφ (39)
2 2
indicating energy flow in the r direction.
Much closer to the dipole [rλ  1/(2π)], (32), (33) and (34) reduce approximately to
[I ]Lλ Z cos θ
Er = −j (40)
4π 2 λrλ3
[I ]Lλ Z sin θ
Eθ = −j (41)
8π 2 λrλ3
[I ]Lλ sin θ
Hφ = (42)
4πλrλ2
From these equations it is apparent that Sr = Sθ = 0. However, the products Er Hφ and Eθ Hφ represent
imaginary or reactive energy oscillating back and forth but not going anywhere. Thus, close to the dipole
there is a region of almost complete energy storage.
Remote from the dipole [rλ  1/(2π)], (32), (33) and (34) reduce approximately to
Er = 0 (43)
[I ]Lλ Z sin θ
Eθ = j (44)
2λrλ
[I ]Lλ sin θ
Hφ = j (45)
2λrλ
Since Er = 0, there is no energy flow in the θ direction (Sθ = 0). However, since Eθ Hφ are in time phase,
their product represents real power flow in the outward radial direction. This power is radiated.
Many antennas behave like the dipole with large energy storage close to the antenna.
The region near the dipole is one of stored energy (reactive power) while regions remote from the dipole
are ones of radiation. The radian sphere at rλ = 1/(2π) marks a zone of transition from one region to the
other with a nearly equal division of the imaginary and real (radiated) power.
The region close to the dipole may be likened to a spherical resonator within which pulsating energy is
trapped, but with some leakage which is radiated. There is no exact boundary to this resonator region, but if
we arbitrarily put it at the radian distance a qualitative picture may be sketched as in Fig. 6–6.

6–4 Radiation Resistance of Short Electric Dipole

Let us now calculate the radiation resistance of the short dipole of Fig. 6–1b. This may be done as follows.
The Poynting vector of the far field is integrated over a large sphere to obtain the total power radiated. This
power is then equated to I 2 R where I is the rms current on the dipole and R is a resistance, called the radiation
resistance of the dipole.
The average Poynting vector is given by
S= 1
2 Re(E × H ∗ ) (1)
The far-field components are Eθ and Hφ so that the radial component of the Poynting vector is
Sr = 12 Re Eθ Hφ∗ (2)
where Eθ and Hφ∗ are complex.
6–4 Radiation Resistance of Short Electric Dipole 165

Radian sphere
resonator
Energy storage

Radiation Radiation

Dipole

Figure 6–6 Sketch suggesting that within the radian sphere at r = λ/2π = 0.16λ the
situation is like that inside a resonator with high-density pulsating energy accompanied by
leakage which is radiated.

The far-field components are related by the intrinsic impedance of the medium. Hence,
μ
Eθ = Hφ Z = Hφ (3)
ε
Thus, (2) becomes
μ
Sr = 1
2 Re ZHφ Hφ∗ = 12 |Hφ |2 Re Z = 12 |Hφ |2 (4)
ε
The total power P radiated is then
  
1 μ 2π π
P = Sr ds = |Hφ |2 r 2 sin θ dθ dφ (5)
2 ε 0 0
where the angles are as shown in Fig. 6–2 and |Hφ | is the absolute value of the magnetic field, which from
(6–3–18) is
ωI0 L sin θ
|Hφ | = (6)
4πcr
Substituting this into (5), we have
 
1 μ β 2 I02 L2 2π π 3
P = sin θ dθ dφ (7)
32 ε π 2 0 0
The double integral equals 8π/3 and (7) becomes
μ β 2 I02 L2
P = (8)
ε 12π
This is the average power or rate at which energy is streaming out of a sphere surrounding the dipole. Hence,
it is equal to the power radiated. Assuming no losses, it is also equal to the power delivered to the dipole.
 166 Chapter 6 Electric Dipoles, Thin Linear Antennas and Arrays of Dipoles and Apertures

Therefore, P must be equal to the square of the rms current I flowing on the dipole times a resistance Rr
called the radiation resistance of the dipole. Thus,
 
μ β 2 I02 L2 I0 2
= √ Rr (9)
ε 12π 2

Solving for Rr ,

μ β 2 L2
Rr = (10)
ε 6π
√ √
For air or vacuum μ/ε = μ0 /ε0 = 377 = 120π so that (10) becomes1

 2
Dipole with L Radiation
uniform current Rr = 80π 2 = 80π 2 L2λ = 790L2λ ( ) resistance (11)
λ

As an example suppose that Lλ = 10 1

. Then Rr = 7.9 . If Lλ = 0.01, then Rr = 0.08 . Thus, the
radiation resistance of a short dipole is small.
In developing the field expressions for the short dipole, which were used in obtaining (11), the restriction
was made that λ  L. This made it possible to neglect the phase difference of field contributions from different
parts of the dipole. If Lλ = 12 we violate this assumption, but, as a matter of interest, let us find what the
radiation resistance of a λ/2 dipole is, when calculated in this way. Then for Lλ = 12 , we obtain Rr = 197 .
The correct value is 168 (see Prob. 6–6–1), which indicates the magnitude of the error introduced by
violating the restriction that λ  L to the extent of taking L = λ/2.
It has been assumed that with end loading (see Fig. 6–1a) the dipole current is uniform. However, with no
end loading the current must be zero at the ends and, if the dipole is short, the current tapers almost linearly
from a maximum at the center to zero at the ends, as in Fig. 2–12, with an average value of 12 of the maximum.
Modifying (8) for the general case where the current is not uniform on the dipole, the radiated power is

μ β 2 Iav
2 L2
P = (W) (12)
ε 12π
where Iav = amplitude of average current on dipole (peak value in time)
The power delivered to the dipole is, as before,

P = 12 I02 Rr (W) (13)

where I0 = amplitude of terminal current of center-fed dipole (peak value in time). Equating the power
radiated (12) to the power delivered (13) yields, for free space (μ = μ0 and ε = ε0 ), a radiation resistance
 2
Iav
Rr = 790 L2λ ( ) (14)2
I0

1 √μ
0 /ε0 = 376.73 . 377 and 120π are convenient approximations.
2 As already given by (2–10–9).
6–5 The Thin Linear Antenna 167

For a short dipole without end loading, we have Iav = 12 I0 , as noted above, and (14) becomes
Rr = 197L2λ ( ) (15)
6–5 The Thin Linear Antenna
In this section expressions for the far-field patterns of thin linear antennas will be developed. It is assumed that
the antennas are symmetrically fed at the center by a balanced two-wire transmission line. The antennas may
be of any length, but it is assumed that the current distribution is sinusoidal. Current-distribution measurements
indicate that this is a good assumption provided that the antenna is thin, i.e., when the conductor diameter is
less than, say, λ/100. Thus, the sinusoidal current distribution approximates the natural distribution on thin
antennas. Examples of the approximate natural-current distributions on a number of thin, linear center-fed
antennas of different length are illustrated in Fig. 6–7. The currents are in phase over each λ/2 section and in
opposite phase over the next.
Referring to Fig. 6–8, let us now proceed to develop the far-field equations for a symmetrical, thin, linear,
center-fed antenna of length L. The retarded value of the current at any point z on the antenna referred to a
point at a distance s is
  
2π L
[I ] = I0 sin ± z ej ω[t−(r/c)] (1)
λ 2

In (1) the function

  
2π L
sin ±z
λ 2
is the form factor for the current on the antenna. The expression (L/2) + z is used when z < 0 and (L/2) − z
is used when z > 0. By regarding the antenna as made up of a series of infinitesimal dipoles of length dz, the
field of the entire antenna may then be obtained by integrating the fields from all of the dipoles making up
the antenna with the result1

 3  5 3 2
2 4 4 2

Figure 6–7 Approximate natural-current distribution for thin, linear, center-fed antennas of
various lengths.

1 For complete development, see the second edition of this book, pp. 220–221.
 168 Chapter 6 Electric Dipoles, Thin Linear Antennas and Arrays of Dipoles and Apertures

To
z distant
point
s

dz r

z


L y

I0

Figure 6–8 Relations for symmetrical, thin, linear, center-fed antenna of length L.

 
j [I0 ] cos[(βL cos θ)/2] − cos(βL/2) (2)
Hφ =
2πr sin θ
Far fields of
center-fed dipole  
j 60[I0 ] cos[(βL cos θ)/2] − cos(βL/2)
Eθ =
r sin θ (3)

where [I0 ] = I0 ej ω[t−(r/c)] and

Eθ = 120πHφ (3a)

Equations (2), (3) and (3a) give the far fields Hφ and Eθ of a symmetrical, center-fed, thin linear antenna of
length L. The shape of the far-field pattern is given by the factor in the brackets. The factors preceding the
brackets in (2) and (3) give the instantaneous magnitude of the fields as functions of the antenna current and
the distance r. To obtain the rms value of the field, we let [I0 ] equal the rms current at the location of the
current maximum. There is no factor involving phase in (2) or (3), since the center of the antenna is taken as
the phase center. Hence any phase change of the fields as a function of θ will be a jump of 180◦ when the
pattern factor changes sign.
As examples of the far-field patterns of linear center-fed antennas, three antennas of different lengths will
be considered. Since the amplitude factor is independent of the length, only the relative field patterns as given
by the pattern factor will be compared.
6–5 The Thin Linear Antenna 169

EXAMPLE 6–5.1 λ/2 Antenna

When L = λ/2, the pattern factor becomes
cos[(π/2) cos θ ]
E= (4)
sin θ
This pattern is shown in Fig. 6–9a. It is only slightly more directional than the pattern of an infinitesimal
or short dipole which is given by sin θ. The beamwidth between half-power points of the λ/2 antenna is
78◦ as compared to 90◦ for the short dipole.

␭/2 antenna
␪

78 (a)

␭
2
Rr  73 

␭ antenna

47 (b)

␭ Rr 2000 

3 ␭/2 antenna

 

  (c)

 
3
␭
2
Rr  106 

Figure 6–9 Three-dimensional and polar plots of the patterns of λ/2, λ, and 3 λ/2 antennas.
The antennas are center-fed with current distributions assumed sinusoidal as indicated.
170 Chapter 6 Electric Dipoles, Thin Linear Antennas and Arrays of Dipoles and Apertures

EXAMPLE 6–5.2 Full-Wave (λ) Antenna

When L = λ, the pattern factor becomes
cos(π cos θ ) + 1
E= (5)
sin θ
This pattern is shown in Fig. 6–9b. The half-power beamwidth is 47◦ .

EXAMPLE 6–5.3 Three Half-Wave (3λ/2) Antenna

When L = 3 λ/2, the pattern factor is

cos 32 π cos θ
E= (6)
sin θ
The pattern for this case is presented in Fig. 6–9c. With the midpoint of the antenna as phase center, the
phase shifts 180◦ at each null, the relative phase of the lobes being indicated by the + and − signs. In all
three cases, (a), (b) and (c), the space pattern is a figure-of-revolution of pattern shown around the axis
of the antenna.

EXAMPLE 6–5.4 Field at Any Distance from

Center-Fed Dipole
The geometry for the field at the point P from a symmet- z
rical center-fed dipole of length L with sinusoidal current
distribution is presented in Fig. 6–10. The maximum cur-
rent is I0 . It may be shown that the z component of the s1 Ez
electric field at the point P is given by I0 P

L 
  r
−j I0 Z e−jβs1 e−jβs2 βL e−jβr
Ez = + − 2 cos (7) y
4π s1 s2 2 r
s2
The φ component of the magnetic field at the point P
(Fig. 6–10) is given by
 
j I0 −jβs1 −jβs2 βL −jβr
Hφ = e +e − 2 cos e (8)
4πr sin θ 2
Whereas the other field equations for oscillating dipoles
given in this chapter apply only with the restrictions of Figure 6–10 Symmetrical center-fed
λ  L and r  L, (7) and (8) apply without distance dipole with sinusoidal current distribution.
restrictions. The field component Ez at any distance
◦
If P lies on the y axis (θ = 90 ) and the dipole is can be expressed as the sum of three
λ/2 long, (7) becomes components radiating from the ends
 and the center of the dipole.
I0 Z ◦ 1 ◦
Ez =  −360 + rλ2 − 90 (V m−1 ) (9)
2πλ 16 + rλ
1 2 16
6–6 Radiation Resistance of λ/2 Antenna 171

and (8) becomes


I0 1
Hφ = −360◦ + rλ2 + 90◦ (A m−1 ) (10)
2πr 16

where
rλ = r/λ
Z = 377
I0 = maximum current = terminal current
At a large distance the ratio of Ez as given by (9) to Hφ as given by (10) is

Ez
= Z = 377 = intrinsic impedance (resistance) of space (11)
Hφ

The magnitude of Hφ is
I0
|Hφ | = (A m−1 ) (12)
2πr

6–6 Radiation Resistance of λ/2 Antenna

To find the radiation resistance, the Poynting vector
√ is integrated over a large sphere yielding the power
radiated, and this power is then equated to (I0 / 2)2 R0 , where R0 is the radiation resistance at a current
maximum point and I0 is the peak value in time of the current at this point. The total power P radiated
was given in (6–4–5)1 in terms of Hφ for a short dipole. In (6–4–5), |Hφ | is the absolute value. Hence, the
corresponding value of Hφ for a linear antenna is obtained from (6–5–2) by putting |j [I0 ]| = I0 . Substituting
this into 6–4–5, we obtain
 
15I02 2π π {cos[(βL/2) cos θ ] − cos(βL/2)}2
P = dθ dφ (1)
π 0 0 sin θ
 π
{cos[(βL/2) cos θ ] − cos(βL/2)}2
= 30I02 dθ (2)
0 sin θ
Equating the radiated power as given by (2) to I02 R0 /2 we have
I02 R0
P = (3)
2
and
 π {cos[(βL/2) cos θ ] − cos(βL/2)}2
R0 = 60 dθ (4)
0 sin θ
where the radiation resistance R0 is referred to the current maximum. In the case of a λ/2 antenna this is at
the center of the antenna or at the terminals of the transmission line (see Fig. 6–7).

 √ 
1P = S · ds = 12 μ/ε |Hφ |2 ds