IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 41, NO.

2, MAY 1999 85

Near-Zone Gain of 500 MHz to 2.6 GHz

Rectangular Standard Pyramidal Horns
Motohisa Kanda, Fellow, IEEE, and Stephen F. Kawalko, Member, IEEE

Abstract— Full-wave solutions to the problem of radiation by where

rectangular standard pyramidal horn antennas are presented. electric field strength V/m;
The radiation problem is formulated in terms of an electric field
integral equation (EFIE), which is solved using the method of net power delivered to the transmitting antenna ;
moments (MoM) for four rectangular standard pyramidal horns free-space impedance 120 ;
covering the frequency range from 500 MHz to 2.6 GHz. Results gain of the transmitting antenna at the given frequency
for the near-zone gains as a function distance from the aperture and distance;
of the horn antenna are presented. A comparison is made with distance from the center of the aperture of the trans-
the analytical formula for the fields and near-zone gains.
mitting antenna (horn or OEG) to the on-axis field
Index Terms— Anechoic chamber, near-zone gain, standard point m.
pyramidal horn.
Notice that is the near-zone gain of a transmitting
standard pyramidal horn because EM field measurements in
I. INTRODUCTION an anechoic chamber are usually performed in the near-field
region of a transmitting antenna; where is
M ICROWAVE anechoic chambers are frequently used for
generating a standard (known) electromagnetic (EM)
field. They can be used for calibrating antennas, EM field
the largest dimension of the transmitting antenna and
the free-space wavelength. The approach used to establish the
is

probes and EM radiation hazard monitors. The anechoic cham- standard field strength is to calculate the radiated field strength
ber at the National Institute of Standards and Technology in the near-field region of a transmitting standard antenna using
(NIST) is shown in Fig. 1. This anechoic chamber provides (1). At NIST, a series of rectangular standard pyramidal horn
calculable EM fields over the frequency range of 200 MHz antennas are used above 500 MHz. The purpose of this paper
to 50 GHz. Rectangular standard pyramidal horn antennas are is to present more accurate values for near-zone gains of four
used as transmitting sources of chamber illumination for fre- standard pyramidal horn antennas used at NIST to establish a
quencies above 500 MHz, and open-ended waveguide (OEG) standard EM field strength in the frequency range from 500
antennas are used below 500 MHz. The prime requirement MHz to 2.6 GHz. The radiation problem for these antennas is
for an anechoic chamber is that a locally plane wave field formulated using an electrical field integral equation (EFIE).
be established to simulate a free-space environment over a The EFIE is then solved for the four rectangular standard
test volume (inside the chamber) large enough to perform the pyramidal horn antennas.
measurements. This volume is frequently called a quiet zone. This paper also provides a comparison with the analytical
The reflectivity in the test volume serves as a measure of solutions for the fields and gain found in the literature [2].
the chamber quality. This reflectivity level can be checked by The analytical formula is derived using aperture integration
measuring the relative insertion loss versus separation distance by making an assumption about the distribution of the electric
between a source antenna and a probe [1]. field incident on the aperture and ignoring the generation of
The electric field strength at a specific distance from a higher order modes at the horn aperture [2].
radiating antenna is calculated from measurements of the
power delivered to the transmitting antenna and a knowledge II. FORMULATION OF PROBLEM
of the gain of the antenna as a function of frequency and The geometry of a typical rectangular pyramidal horn is
distance to the field point. The equation used to calculate the shown in Fig. 2. The walls of the horn and the waveguide
electric field along the bore sight axis of the transmitting section are assumed to be perfectly conducting and have
antenna is infinitesimal thickness. The pyramidal horn is excited by a line
source located near the closed end of the waveguide section.
(1) This line source is oriented to excite the dominant TE mode
in the waveguide. Since a thin wall model for a pyramidal
horn antenna is being used, it is necessary to formulate the
Manuscript received May 11, 1998; revised February 5, 1999. problem using an electric field integral equation (EFIE). An
M. Kanda is with the National Institute of Standards and Technology, EFIE suitable for this problem is [3]
Boulder, CO 80303 USA.
S. F. Kawalko is with Electro Magnetic Applications, Inc., Lakewood, CO
80226 USA.
Publisher Item Identifier S 0018-9375(99)04086-7.

0018–9375/99$10.00  1999 IEEE

86 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 41, NO. 2, MAY 1999

Fig. 1. A side view of the NIST anechoic chamber.

Once a solution for the surface current has been obtained,

the directivity can be computed using the following expres-
sion [5]:

(4)

where is the radiation intensity and is given by

Re (5)

is the average value of the radiation intensity,

is the total radiated power, and and
are the total electric field and total magnetic
Fig. 2. Geometry of a rectangular pyramidal horn. field evaluated at the point . and are the incident fields
due to the current distribution on the source, while and
are the scattered fields due to the surface current induced on the
(2) surface of the horns. Expressions for the incident and scattered
fields written in terms of the usual electric vector potential and
electric scalar potential can be found in most electromagnetic
where and are observation and source points, respectively, theory and antenna books [5], [6].
which are located on the surface of the pyramidal horn antenna The physical dimension and the frequency range for the
denoted by is the component of the incident electric individual antennas that cover the frequency range from 500
field which is tangent to is the impedance of free-space, MHz to 2.6 GHz are given in Table I. For each antenna,
is the free-space wave number, is the (equivalent) surface meshes with a mesh size of approximately
surface current density excited on the walls of the pyramidal (where is the free-space wavelength at the highest
horn antenna, and is the free-space Green’s function operating frequency of the antenna) are created. The surface
mesh size of the horn section is generally finer than that of
the waveguide section. The number of surface patches, the
(3) resulting problem size, storage requirements, and computation
time for a single frequency for each pyramidal horn antenna
The notation is used to indicate the tangential (to are given in Table II. In order to make the computation more
the surface ) component of the enclosed vector quantity. manageable, two planes of symmetry were used to reduce the
A general purpose surface-patch method of moments (MoM) number of unknowns.
code, which uses parametric triangular surface patches, vector For each pyramidal horn antenna, a solution of (2) was
basis functions [3], [4], and Galerkin testing, is used to solve obtained at either 25 or 50 MHz steps in the frequency
the EFIE given in (2). range of the horn. A dipole source was used to excite the
KANDA AND KAWALKO: NEAR-ZONE GAIN OF 500 MHz TO 2.6 GHz RECTANGULAR STANDARD PYRAMIDAL HORNS 87

TABLE I
PHYSICAL DIMENSIONS OF FOUR RECTANGULAR PYRAMIDAL HORN ANTENNAS1 WHERE A, B , AND C ARE THE WIDTH, HEIGHT, AND LENGTH OF THE WAVEGUIDE
SECTION OF THE ANTENNA; a AND b ARE THE WIDTH AND HEIGHT OF THE HORN APERTURE; AND `H AND `E ARE THE HORN SLANT HEIGHTS

TABLE II
PROBLEM SIZES AND EXECUTION TIMES USING TWO PLANES OF SYMMETRY TO REDUCE THE NUMBER OF UNKNOWNS. Nfreq: points IS THE
NUMBER OF FREQUENCIES AT WHICH A METHOD OF MOMENTS SOLUTION IS OBTAINED. Npatches IS THE NUMBER OF SURFACE PATCHES IN THE
SURFACE-PATCH MODEL FOR EACH HORN. Nunknowns IS THE TOTAL NUMBER OF UNKNOWNS WHEN TWO PLANES OF SYMMETRY ARE
USED. Tmatrix gen: IS THE MATRIX FILL TIME FOR THE IMPEDANCE MATRIX AND TLU decomp: IS THE TIME REQUIRED TO PERFORM AN
LU DECOMPOSITION ON THE IMPEDANCE MATRIX. RUN TIMES ARE FOR AN SGI POWER CHALLENGE L (t ESTIMATED TIME)

waveguide section of the antenna. This dipole was located at is given by [7]
the approximate location of the actual feed in each antenna.
The actual source in each of these waveguide sections is a
(6)
bowtie monopole. Since the actual current distribution and
exact location of each feed are not precisely known, a solution
was obtained for three different current distributions (sinu- where
soidal, constant, and Hertzian) at three different locations for a
total of nine different sources. The length of the waveguide Re (7)
section is taken to be , where is the width of
the waveguide. The actual length of waveguide section of the is the surface resistance, is the conductivity of the metal,
horns is generally much longer than this, particularly for the is the skin depth of the metal, is the
horns used at higher frequencies and it is impractical to model angular frequency, and is the permeability of the metal. An
the horns using the actual length. The value was estimate of the radiation efficiency
selected to ensure that a TE mode is incident on the throat of the horn antenna can then be computed.
of the horn and also to ensure that the horns could be modeled
using the method of moments.
The EFIE given by (2) assumes that the scatterer is perfectly III. COMPUTATION RESULTS AND VERIFICATION
conducting. The validity of this assumption and its effect on The computed values for the directivity of the four antennas
the gain of the horn antenna being studied can be estimated by as function of frequency computed using a thin wall model are
computing an estimate of the ohmic power losses in the walls summarized in Tables III–VI. In addition, Table VII contains
of the horn and the waveguide section. Given a solution of the computed values for the NARDA-645 horn using a thick
(2) for the surface current density , a first-order estimation wall model. These values are provided since for this frequency
of the ohmic losses , in the walls of the horn antenna range the use of a finite wall thickness in the MoM model
begins to have a noticeable effect on the directivity. The
directivity in Tables III–VII has been computed along the axis
1 Certain commercial equipment, instruments, or materials are identified in of the horn antennas at various distances between 1–5 m. The
this paper in order to adequately specify the experimental procedure. Such values of the directivity given in Table III is for a single source
identification does not imply recommendation or endorsement by the National
Institute of Standards and Technology nor does it imply that the equipment, excitation, while the values in Tables IV–VII are average
instruments, or materials are necessarily the best available for the purpose. values for three sources at three different locations (total of
88 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 41, NO. 2, MAY 1999

TABLE III
COMPUTED DIRECTIVITY AND ESTIMATED EFFICIENCY FOR A SA12-0.5 STANDARD
GAIN-HORN ANTENNA. THE DIRECTIVITY WAS COMPUTED ALONG THE AXIS OF THE HORN

TABLE IV
COMPUTED DIRECTIVITY AND UNCERTAINTY ESTIMATES FOR A SA12-0.5 STANDARD GAIN-HORN ANTENNA. THE DIRECTIVITY WAS COMPUTED ALONG THE AXIS OF
THE HORN. THE TOTAL ESTIMATED UNCERTAINTY IS GIVEN BY utotal =
std: direct: , WHERE std: direct: IS THE STANDARD DEVIATION OF THE DIRECTIVITY

nine different source excitations) and the estimated uncertainty effect of finite range. The gain reduction factors are
is simply the standard deviation of these values.
Also, for comparison, the near-zone gains are calculated
using an analytical formula. This formula was derived using
aperture integration by assuming that no reflection occurs at
the throat or aperture of the horn and that the field incident on (9)
the aperture is essentially a TE mode but with a quadratic
phase distribution across the aperture. The analytical formula
given by Jull [2] for the near-zone gain is where

(8)
(10)
where is the free-space wavelength, and and are
gain reduction factors that include the gain reduction due to (11)
the -plane and -plane flares of the horn as well as the
KANDA AND KAWALKO: NEAR-ZONE GAIN OF 500 MHz TO 2.6 GHz RECTANGULAR STANDARD PYRAMIDAL HORNS 89

TABLE V
COMPUTED DIRECTIVITY AND UNCERTAINTY ESTIMATES FOR A NARDA-646 STANDARD GAIN-HORN ANTENNA. THE DIRECTIVITY WAS COMPUTED ALONG THE AXIS
OF THE HORN. THE TOTAL ESTIMATED UNCERTAINTY IS GIVEN BY utotal =
std: direct: , WHERE std: direct: IS THE STANDARD DEVIATION OF THE DIRECTIVITY

TABLE VI
COMPUTED DIRECTIVITY AND UNCERTAINTY ESTIMATES FOR A NARDA-645 STANDARD GAIN-HORN ANTENNA USING THIN
WALL MODEL. THE DIRECTIVITY WAS COMPUTED ALONG THE AXIS OF THE HORN. THE TOTAL ESTIMATED UNCERTAINTY
IS GIVEN BY utotal =std: direct: , WHERE std: direct: IS THE STANDARD DEVIATION OF THE DIRECTIVITY
90 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 41, NO. 2, MAY 1999

TABLE VII
COMPUTED DIRECTIVITY AND UNCERTAINTY ESTIMATES FOR A NARDA-645 STANDARD GAIN-HORN ANTENNA USING A THICK WALL MODEL
WITH A WALL THICKNESS OF t= 0:2032 cm. THE DIRECTIVITY WAS COMPUTED ALONG THE AXIS OF THE HORN. THE TOTAL ESTIMATED
UNCERTAINTY IS GIVEN BY utotal = std: direct: , WHERE std: direct: IS THE STANDARD DEVIATION OF THE DIRECTIVITY

is the on-axis distance from the aperture plane, is the

-plane slant height, and is the -plane slant height. The
Fresnel integrals and are defined as

(12)

Comparisons of the near-zone gains of the four pyra-

midal horn antennas obtained from the MoM calculation
with those obtained from Jull’s analytical formula are shown
in Figs. 3–6. Although there is generally good agreement,
Jull’s analytical formula [2] does not account for all of the
diffraction effects that occur at the aperture of the horn
antennas. The small oscillations in the gain (directivity) seen in
the results obtained from the direct MoM calculations are due
to multiple diffractions from the edges of the horn aperture
[8], [9].
Also given in Tables III–VII are estimated values for the Fig. 3. Comparison of the on-axis directivity for a SA12-0.5 standard
gain-horn antenna.
radiation efficiency of the horn antennas. The estimated radi-
ation efficiency of the four horn antennas ranges from 0.9904
to 0.9963 assuming m .
The actual conductivities of the walls of the horn and the due to nine different source excitations are also given in
waveguide section are not known, but these estimates for each table.
the radiation efficiency indicate that in the worst case, no
more than 1% of the power delivered to the antenna is
dissipated by resistive losses in the walls of the horn and the IV. CONCLUDING REMARKS
waveguide section. Like the directivity, the estimated values In the paper, we have presented the results of the near-
for the antenna efficiency given in Tables IV–VII are average zone gain obtained from a direct full-wave solution to the
values over total of nine different source excitations. The radiation problem for several rectangular standard pyramidal
standard deviations of the antenna directivity and efficiency horn antennas for the frequency range between 500 MHz and
KANDA AND KAWALKO: NEAR-ZONE GAIN OF 500 MHz TO 2.6 GHz RECTANGULAR STANDARD PYRAMIDAL HORNS 91

Fig. 4. Comparison of the on-axis directivity for a SA12-0.75 standard Fig. 6. Comparison of the on-axis directivity for a NARDA-645 standard
gain-horn antenna. gain-horn antenna.

for the waveguide section. For these reasons, we expect the

discretization error for the horn antennas to be of the same
order of magnitude as or smaller than that reported for the
open-ended waveguides [10].
Other potential sources of uncertainties need to be ex-
amined. For example, there are uncertainties due to small
differences between the mathematical model for the geometry
of the horn antenna and its actual geometry, the uncertainty
introduced by using a mathematical model in which the
walls of the horn section and the waveguide section are
infinitesimally thin versus the real horn antenna which has
walls with finite thickness, and the uncertainty in the radiation
efficiency of the horn antenna due to errors in the conductivity
of the wall metals. Although these sources of uncertainties
have not been examined carefully in this paper, we have found
from the previous study that the uncertainties due to these
sources are less significant than the discretization errors [10].
Further research should be conducted to evaluate uncertainties
Fig. 5. Comparison of the on-axis directivity for a NARDA-646 standard
due to these sources. A comparison with some previously
gain-horn antenna. published results [11] indicates that a reasonable value for the
predicted uncertainty for the near-zone gain of the rectangular
pyramidal horns presented in this paper is 5%.
2.6 GHz. This paper also provides a comparison with the
REFERENCES
analytical solution for the fields and gain.
The detailed uncertainty analysis for estimating the near- [1] M. Kanda, “Methodology for electromagnetic interference measure-
zone gains of the horn antenna has not been accomplished in ments,” in Handbook of Electromagnetic Compatibility. New York:
Academic, 1995, ch. 5.
the paper and will be a very important topic for a future study. [2] E. V. Jull, “Finite-range gain of sectorial and pyramidal horns,” Electron.
However, as is the case of the open-ended waveguide [10], Lett., vol. 6, pp. 680–681, 1960.
we estimate that the largest component of the uncertainty in [3] S. M. Rao, D. R. Wilton, and A. W. Glisson, “Electromagnetic scattering
by surfaces of arbitrary shape,” IEEE Trans. Antennas Propagat., vol.
the computed values for the directivity comes is discretization AP-30, pp. 409–418, May 1982.
error. The singular behavior in the currents at the aperture [4] S. Wandzura, “Electric current basis functions for curved surface,”
Electromagn., vol. 12, pp. 77–91, 1992.
of the horn antenna is not as significant as what occurs [5] W. L. Stutzman and G. A. Thiele, Antenna Theory and Design. New
at the OEG aperture. At the same time, the average edge York: Wiley, 1981.
length for the mesh of the horn section is shorter than that [6] R. F. Harrington, Time-Harmonic Electromagnetic Fields. New York:
McGraw-Hill, 1961.
for the waveguide section. This implies that estimates of the [7] R. E. Collin, Foundation for Microwave Engineering. New York:
discretization error are dominated by the discretization error McGraw-Hill, 1992.
92 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 41, NO. 2, MAY 1999

[8] E. V. Jull, “Errors in the gain of pyramidal horns,” IEEE Trans. Antennas Stephen F. Kawalko (S’86–M’95) received the B.S., M.S., and Ph.D. degrees
Propagat., vol. AP-21, pp. 25–31, Jan. 1973. in electrical engineering from the University of Illinois at Chicago in 1987,
[9] J. F. Nye and W. Liang, “Theory and measurement of the field 1990, and 1994, respectively.
of a pyramidal horn,” IEEE Trans. Antennas Propagat., vol. 44, pp. From 1995 to 1998, he was with the Fields and Interference Metrology
1488–1498, Nov. 1996. Group of the National Institute of Standards and Technology, Boulder,
[10] S. F. Kawalko and M. Kanda, “Near-zone gain of open-ended rect- CO. He is currently a Staff Scientist with ElectroMagnetic Applications,
angular waveguides,” IEEE Trans. Electromagn. Compat., vol. 39, pp. Inc., in Lakewood, CO. His research interests include the application of
408–413, Nov. 1997. numerical methods in electromagnetics, in particular, antenna analysis and
[11] M. Kanda and R. D. Orr, “Near-field gain of a horn and an open-ended electromagnetic scattering.
waveguide: Comparison between theory and experiment,” IEEE Trans.
Antenna Propagat., vol. AP-35, pp. 33–40, Jan. 1987.

Motohisa Kanda (S’67–M’68–SM’83–F’88) re-

ceived the B.S.E.E. degree from Keio University,
Tokyo, Japan, in 1966, and the M.S.E.E. and Ph.D.
degrees from the University of Colorado, Boulder,
in 1968 and 1971, respectively.
From 1965 to 1966, he was a Research Technician
at Keio University where he did research on the
avalanche breakdown in the germanium p-n
junction at a cryogenic temperature. From 1966 to
1971 he was a Research Assistant at the University
of Colorado, where he conducted research on
impact ionization of impurities in n-type germanium, and nonreciprocal
behavior in a solid-state plasma at millimeter and submillimeter wavelengths.
In 1971 he joined the staff of the Electromagnetic Fields Division, the
National Institute of Standards and Technology (NIST, formerly Bureau
of Standards), Boulder, CO, where he is currently Leader of the Fields
and Interference Metrology Group. Concurrently, he serves as a Professor
Adjunct in the Electrical and Computer Engineering Department, University
of Colorado.
Dr. Kanda is the Editor-in-Chief of the IEEE TRANSACTIONS ON
ELECTROMAGNETIC COMPATIBILITY. He was Chairman of the U.S.
Commission A of the International Union of Radio Science (URSI)
from 1990 to 1993. He was Vice Chairman from 1994 to 1996 and was
elected to serve as Chairman of the International Commission A of URSI
from 1997 to 2000. He has received the Bronze Medals (1981 and 1992)
and Silver Medal (1989) from the U.S. Department of Commerce. He has
received three IEEE EMC Transaction Best Paper Awards in 1982, 1989,
and 1992, Honorary Life Member in 1996, and other numerous awards from
the IEEE EMC Society and the Department of Commerce.