JEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, MARCH 1976 235

observed for single straipht wires, Projected efficienciesfor large TheEMCAP (ElectromagneticCornpatabilityAnalysisProgram)
ThirdQuarterTech.Rep.,Part2,TheBDMCorporation,Albu-
wire grids indicate that the banded matrix iterative method may querque, N. Mex. (RA.DC-TR-75-121).
provide solutions for large surfaces at a reasonable cost. [5] -, “Solution of thm wire momentsproblemsbybandedmatrix
iteration,” TheEMCAP (ElectromagneticCompatabilityAnalysis
Program) Fourth Quarter Tech. Rep., Part 2, The BDM Corporation,
Albuquerque, N. Mex. RADC-TR-75-189.
REFERENCES [6] D. J. and
Rose R.
Willoughby,
A. Ed., Sparse Matrices and Their
Applications. New York: Plenum, 1972.
[I] R. F. Harrington, Field Computation byMomentMethods. New [7l G . E.ForsytheandC. B. Moler, Computer Solution of Linear A!gebraic
York: Macmillan, 1968. Systems. Englewood C!iffs, N.J.:Prentice-Hall, 1967.
[2] V. V. Klyuyev and N. 1. Kokovkin-Shcherbak, “On the minimization [8] J. R. Stewart, “Application of optimization theory t o electromagnetic
of the
number of arithmetic
operations for
the
solution of linear
radlatlon
andscattering,”
Ph.D.
dissertation,
Syracuse Univ.,
algebraic systems ofequations,”(transl. by G . J. Tee), Computer Syracuse, N.Y., Jan. 1974.
Science Department,Stanford University, Stanford, Calif., Tech.Rep. [9] F. J. Deadrickand 5. K. Miller, “ProgramWAMP(WireAntenna
ModelingProgram),LawrenceLivermoreLaboratory,Dec. 1973.
CS24, 1965.
[3] F. K.Oshiro et al., “Calculation of radarcrosssection,part 11, (UCID-30084). A Modified Version of aProgramCreated by MB
analyticalreport,”NorthrupCorporationAFAL-TR-70-21,Part 2, Associates.
1970. [lo] M, T. Ma, Theory and Application of Antenna Arrays. New York:
[4] T. R. Ferguson,“Iterativetechniquesinthemethodofmoments,”Wdey, 1974.

. .
Cornmumcations
An Approximate Formula for Calculating the Directivity of and
an Antenna
CHEN-TO TAI, FELLOW, IEEE,AND CLOVIS S. PEREIRA 1 ( ;)I2
E
+
8, - , the H-plane pattern.

From these two patterns we propose that the directivity of the

Absfract-An approximate general formulato calculate the directivity antennacan be obtained approximately, butquite accurately,
of an antenna basedupon the E-planeand H-plane patterns is proposed.
For narrowbeam patterns, the directivity is expressed in terms of the using the formula
half-power beam widths of the main patterns. The better approximation
_1 _ 1- 1 -
of the formulapresentedhereaverthegeometrical
pointed out.
mean formula is
D - 2 (Dl ’$)
where
The directivity of an antenna is defined as

D =
4n
1s urnax
U ( 8 , d ) sin 8 dB dd
(1)

where U(@,q5)denotes the far-zone power density, andit is

related to the far-zone electrical field by
The simplicity of thisapproximateformula is that only
1
u(e,d) = - [iEe(8,4)Iz + IE+(Q,d)l21 one-dimensional integrals are involved in evaluating D land D,.
220 These can be done numerically if the E-plane and the H-plane
where patterns are available. The formula given by (2) will be referred
as the arithmetic-mean formula. The expression for D , defined
z o = (po/&o)”2.
by (3) corresponds tothe directivity of anantenna with a
The functions E, and E+ which represent the two angular rotationally symmetrical pattern IEe(B,0)12 and that for D, with
components in a spherical system are, in general, functions of a rotationally symmetrical pattern !E4(e,(n/2)12. For uniform
B and q5. In experimental work one often measures the power arrays of short dipoles operated as broadsidearrays or as
pattern in two principal planes. These power patterns correspond end-fire arrays it can be shown that the arithmetic-mean formula
to is exact. For arrays made of half-wave dipoles there is sufficient
evidence that the formula is quite accurate.
IEe(6,0)12, the E-plane pattern For antennas with a narrow beam pattern it is desirable to
relate D l and D, in terms of the half-power beamwidth of the
Manuscript received August 18, 1975; revised October 13, 1975. This E-plane and the H-plane pattern, hereby denoted by 8, and 0,.
work was supported in part by the National Science Foundation under An approximate expression for D lin terms of 8, is
Grant GK-36867.
C. T. Tal 1s wlth the Instituto de Pesquisas Ekpaciais, Si0 Paulo, Brasil
on kave from the Department of Electrical and Computer Engineering: Dl E 16 In 2/QlZ
University of Michigan, Ann Arbor, MI 48104. and similarly
C. S . Pereira is with theInstitutodePesquisasEspaciais, SZo Paulo,
Brad D, N 16 In2/t?,2
 236 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, hiARCH 1976

time harmonic solution. For example, we present exact expressions for

thetransientnear-andfar-fieldaroundsmallcircularcurrent loops
involving arbitrarily time-varying excitation. The resulting convolution
These expressions are obtained by considering the asymptotic integrals have been numerically evaluatedfor step function currentswith
expression for the directivity of an antenna with a rotationally k i t e rise time, thus showing the influence of finite loop radius on the
symmetrical power pattern of theform U(8) = C O P 8, for radiation field in comparison with the ideal magnetic dipole.
x12 2 8 2 0, and U ( 8 ) = 0, for n > 8 > n / 2 with very large
value of m. Alternative formulas can be derived by using the I. I~TRODUCITON
asymptotic expressions generated by Legendre polynomials or A recent paper by G . Franceschetti and C. H. Papas [l] pre-
Chebyshev polynomials. This subject will be discussed in detail sents a number of very interesting procedures and results-
on another occasion. besides numerous references-relating tothe computation of
It is recalled that an approximate expression for D proposed transient radiation from elementary sources. Here, we take one
by Kraus [l ] for a narrow beam pattern is of their examples-the small circular loop antenna-to give a
4n new and quite simple approach to thehandling of certain struc-
e,e,
A &
tures radiating arbitrarily in time. The method applies an idea
Consider the same power e defined above. For credited to Cagniard [21 which has been used to solve seismic
m = 2, the exact value of D is six. Equation ( 3 , with = e,,
and electromagnetic pulse problems [3], [4].It allows us to
yields = 4.62 \vhile (@yields D, = 5.09. For = the extend the results of Franceschetti andPapastothe near-field
exact of is 202 while the approximate values obtained radiation zone Ieading to exact expressions forthe components
by (5) and (6) are, respectively, 206 and 233. It would be interest- the Of without the Of the
ing to verify experimentally which formula gives a better overall corresponding time harmonic solution. Thisis especially interest-
approximation, particularly for asymmetrical patterns. ing in those cases where this solution is rather complicated as
for large loop antennas with sinusoidal current distribution [SI.
It should be mentioned that an approximate formula in the
-, The terms “smaI1” or ‘‘large’’ apply the to VT space scale, T
Of = was Once proposed by et [21‘
being the rise t h e o f a step
function
current and
the phase
Their formuladoesnotappear to have the analytical foundation velocity of the isotropic, homogeneous, and nondispersive
as ‘Ompared to OUTS (2) when to uniform Of medium surrounding the antenna.
The consideration of finite
dipoles. For the case of a short horizontal electrical dipole we rise times instead of ideal step functions is of great importance
have in the physical discussion of the results, because the assumption
6 = Eo[cos 8 cos 4 8 - sin 4 61 of constant current over all the small loop is only fulfilled if
hence a << vT, a being the loop radius; thatis to say the results based
01=3 D2=1 on ideal step functions only comply to the limiting case a 0,
thus the arithmetic-mean formula gives D = 8, which is exact,
is., to theidealmagnetic dipole. For finitea > 0 Franceschetti’s
while the ge0metric-m- formula yields D = &=
1.73, an and Papas’ statement, thatthe radiated far-fieldof the small
error of 15 percent. loop is not proportional
the to second time derivative of the
current, is only of mathematical interest because an essential
ACKNOWLEDGMENT
physical assumption is not fulfilled. Our numerical results show
A comment On Our Original manuscript prompted us clearly thatup to a certainapproximation the radiated far-field
to give a better presentation of the approximation relating the of small loop antennas is proportiona~to the second
directivity and the beamwidth for narrow beam patterns. The derivative of the current; a finite loop radusa 0 results only
reviewer’s careful reading of our manuscript isvery much in small modifications of the transient near- and far-field of the
,
appreciated. ideal magnetic dipole. This might be of interest for the investi-
REFERENCES gations of Harmuth
and Fralick [6], [7] concerning the radiation
. ~ . .
[I] John D. Kraus, Antennas. NewYork:McGraw-Hill, 1950, p. 25. of Walsh-wava from small loop antemas.
[2] L. Boithias, R. BCtt, ?fd C. Mathieu, “Le centre d’etudes d’antennes
du CNET a la TurbiC, L’kho des Recherches, July 1967.
Note added in Proof: In ,a privatecommunicationwithProf. John D. 11. MATHEMA~CAL PROBLEM
F O ~ A ~OF O N AND
Kraus, he called our attent~on to general the formula relatmg the dlrect~v~ty
half-power
theand
beamwidth described
book
his in Radio Asfronomy OF SOLUTION
(New York: McGraw-Hill, 1966, p. J58). In the context of that general
formula,thepatternshapefactor wh~chy e have derlved hasthevalue of The electromagnetic field O f a given current density J(r,t),
sr/4 In 2 o r 1.13. It 1s very close to the maxlmumvalue prescr~bedby Kraus being the of position andthe time, can be determined
(1.05 If: 0.05). The authors wish to thank Prof. Kraus for his comment.
from the retarded vector potential A(r,t), where

Transient Fields of Small Loop Antennas

KARL J6RG LANGENBERG The vector r’ denotes the position in the sourceregion V;
v = (co&pop)-1/2is the light phase velocity, eo& the permittivity,
Abstract-A rather simple and general analysis allows us to calculate and pop the permeability of the isotropic, homogeneous, and
directly the transient response of many different radiators of practical
ioterest in the time domainthe of the correspondiog nondispersive medium surrounding v. Laplace transforming
equation (l), s being the variable in the transform space, yields

Manuscript received June 21, 1975; revised September 9, 1975.

Theauthor is withFachrichtung
Elektrotechnlk, Universiat
des
Saarlandes, Saarbriicken, Germany. 471 Jv IP - r‘