IEEE Standard Test Procedures for Antennas · IEEE Std 149-1979 (Revision of IEEE Std 149-1965)...

IEEEStd 149-1979

(Revision ofIEEE Std 149-1965)

IEEE StandardTest Procedures for Antennas

Sponsor

Antenna Standards Committee

December 19. 1979

Library of Congress Catalog Number 79-92425

@Copyright 1979 by

The Institute of Electrical and Electronics Engineers, Inc

No part of this publication may be reproduced in any form,in an electronic retrieval system or otherwise,

without the prior written permission of the publisher.

SH07682

Approved December 15, 1977

IEEE Standards Board

William R. Kruesi, Chairman Irvin N. Howell, Jr, Vice Chairman

Ivan G. Easton, SecretaryWilliam E. Andrus R. 0. DuncanJean Jacques Archambault Charles W. FlintMark Barber Jay ForsterEdward J. Cohen Ralph I. HauserWarren H. Cook Joseph L. KoepfingerLouis Costrell Irving KolodnyR. L. Curtis Benjamin J. LeonDavid B. Dobson Thomas J. Martin

Donald T. MichaelVoss A. MooreWilliam S. MorganWilliam J. NeiswenderRalph M. ShowersRobert A. SoudermanLeonard W. Thomas, SrB. W. Whittington

Foreword

(This Foreword is not a part of IEEE Std 149-1979, IEEE Standard Test Procedures for Antennas.)

This document is a major revision of IEEE Std 149-1965 which it supersedes. It represents thesecond revision of the standard since the original issuance in 1948 of 48IRE2S2, Standards onAntennas - Methods of Testing. Practically every topic contained in the previous standard has beenexpanded to reflect the great changes that have taken place, since 1965, in metrology and instru-mentation technology as applied to antenna measurements.

This document contains sections on the design, evaluation, and operation of antenna ranges,electromagnetic radiation hazards, and environmental factors which did not appear in the precedingstandard. The section on the determination of scattering cross-section, which appeared previously,has been omitted since it will appear as a separate standard at a later date.

Suggestions for the improvement of this standard will be welcome. They should be sent to:

SecretaryIEEE Standards BoardThe Institute of Electrical and Electronics Engineers, Inc345 East 47th StreetNew York, NY 10017

This standard was prepared by the Subcommittee 2.11 on Methods of Testing Antennas of theIEEE Antenna Standards Committee. The Subcommittee preparing this revision had the followingmembership :

W. H. Kummer, Chairman

J. D. DysonE. S. GillespieT. Mukaihara

A. T. Villeneuve

A. C. NewellA. F. SeatonG. P. Tricoles

Members of the IEEE Antenna Standards Committee who contributed to this standard documentwere:

E. S. Gillespie, Chairman

C. C. AllenK. G. BalmainP. L. BurgmyerH. V. Cottony*G. A. DeschampsJ. D. DysonE. S. GillespieP. W. HannanH. JasikW. K. KahnE. M. KennaughW. H. Kummer

M. S. Wheeler

D. J. LeVineT. MukaihataA. C. NewellD. C. PortsL. J. RicardiA. C. SchellA. F. SeatonC. J. Sletten*P. H. SmithW. T. TilstonG. P. TricolesA. T. Villeneuve

*Past chairmen

Contents

SECTION

1 . S c o p e . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2. Standards References ...................................................

3. Antenna-Range Measurements of Radiation Patterns ............................

3.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.2 Pattern Cuts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.3 Basic Antenna-Range Configurations. ..................................

4. Antenna-Range Design ..................................................

4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2 Antenna-Range Design Criteria .......................................

4.2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2.2 Effect of Mutual Coupling Between Source and Test Antennas. .......4.2.3 Effect of a Transverse Amplitude Taper over the Test Aperture .......4.2.4 Effect of a Longitudinal Amplitude Taper at the Test Antenna .......4.2.5 Effect of Phase Variation over the Test Aperture ..................

4.3 Design of Elevated Ranges. ..........................................4.3.1 Elevated Ranges over Flat Surfaces. ............................4.3.2 Elevated Ranges over Irregular Surfaces .........................4.3.3 Estimation of Errors Due to Reflections. ........................4.3.4 Use of Diffraction Fences and Longitudinal Ramps to Redirect Re-

flected Energyy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.4 Design of Ground-Reflection Ranges. ..................................4.5 Other Ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.5.1 Slant Rangee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.5.2 Compact Rangee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.5.3 Image-Plane Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.5.4 Anechoic Chambers. ........................................

5. Antenna-Range Instrumentation. . .........................................

5.15.25.35.45.5

5.65.7

General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Source Antennas for Antenna Ranges. .................................Transmitting Systems ..............................................

Receiving Systems. ................................................Positioning Systems. ...............................................5.5.1 Antenna Positioners ........................................5.5.2 Antenna-Positioner Errors. ...................................Antenna-Pattern Recorder. ..........................................Data-Processing and Control Computers ................................

PAGE

13

14

14141617

171718181819192020202222

23252828293030

32323232343737394145

SECTION PAGE

6. Antenna-Range Evaluation ...............................................6.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6.2 Field-Probe Measurements over Test Aperture ...........................6.3 Incident-Field Measurements Near the Range Axis on an Elevated Range. ......6.4 Incident-Field Measurements Near the Range Axis on a Ground-Reflection

Range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6.5 Wide-Angle Incident-Field Measurements ...............................

6.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6.5.2 Antenna-Pattern-Comparison Method ...........................6.5.3 Longitudinal-Field-Probe Method ..............................

6.6 Evaluation of Anechoic Chambers. ....................................

47474749

515252525454

7. Special Measurement Techniques ..........................................7.1 Modeling Techniques. ..............................................7.2 Antenna-Focusing Technique .......................... . .............7.3 Near-Field Probing with Mathematical Transformation .....................7.4 Swept-Frequency Technique. ........................................7.5 Indirect Measurements of Antenna Characteristics ........................

565658596263

8. Antenna-Range Operation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . _ 65

9. On-Site Measurements of Amplitude Patterns. ................................ 67

10. Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6910.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6910 .2 Phase Patternss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7010.3 Antenna Phase Center. .......................................... ... 7010.4 Phase Measurements ............................................... 71

10.4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7110.4.2 Instrumentation ........................................... 7410.4.3 Sources of Error ........................................... 75

11. Polarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11.2 Polarization Measurements ........................................

11.2.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11.2.2 Measurement of the Polarization Pattern. ......................11.2.3 Rotating-Source Method. ..................................11.2.4 Multiple-Amplitude-Component Method. ......................11.2.5 Phase-Amplitude Method ..................................

7676858586888890

12. Measurement of Power Gain and Directivityy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9412.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9412.2 Gain Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95

12.2.1 Types of Gain Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9512.2.2 Calibration of Gain Standards on a Free-Space Antenna Range. . . . . . . . 9612.2.3 Calibration of Gain Standards on a Ground-Reflection Antenna Range . . 9712.2.4 Calibration of Gain Standards on an Extrapolation Antenna Range . . . . 99

SECTION

13.

1 4 .

15.

16.

17.

18.

19.

20.

21.

12.3 Gain-Transfer Measurements .........................................12.3.1 Measurement of Linearly Polarized Antennas .....................12.3.2 Measurement of Circularly and Elliptically Polarized Antennas. .......12.3.3 Measurement in the High-Frequency Range (3-30 MHz) ............

12.4 Measurement of the Power Gain of Electrically Large Antennas. .............12.4.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12.4.2 Use of Extraterrestrial Radio Sources for Power-Gain Measurements ...12.4.3 Measurement of Absolute Antenna Noise Temperature and Figure of

Merit G/TT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12.4.4 Measurement of the Power Gain of Electrically Large Antennas Using

the Gain-Transfer Method. ...................................12.5 Errors in Power-Gain Measurements ...................................

12.5.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12.5.2 Sources of Error ...........................................

12.5.3 Estimation of Uncertainty in Gain Measurements ...................

12.6 Directivity Measurements ...........................................

Determination of Radiation Efficiency. .....................................

Special Measurements for Angle-Tracking Antennas ............................

14.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14.2 Conical Scanning Angle Tracking .....................................

14.3 Monopulse Angle Tracking ..........................................

14.4 Electrical Boresight Measurements ....................................

Measurement of the Electrical Properties of Radomes. ..........................

15.1 Generali . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15.2 Significant Antenna-Radome Parameters. ...............................15.3 Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15.4 Testing of Wet Radomes ............................................

Measurement of Impedances. .............................................16.1 Input-Impedance Measurements ......................................

16.2 Mutual-Impedance Measurements .....................................

Ground-Wave Measurements ..............................................

Power-Handling Measurements ............................................

Electromagnetic Radiation Hazards. ........................................19.1 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19.2 Safe Radiation Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19.3 Measurement and Instrumentation ....................................

Environmental Factors ..................................................

Bibliography.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

PAGE

100100100101102102103

106

106107107107110110

112

113113114115115

116116116117119

119119120

123

125

127127127128

129

130

FIGURES PAGE

Fig 1Fig 2Fig 3

Fig 4Fig 5

Fig 6Fig 7Fig 8Fig 9Fig 10Fig 11Fig 12Fig 13Fig 14

Fig 15Fig 16

Fig 17

Fig 18

Fig 19Fig 20Fig 21Fig 22Fig 23Fig 24Fig 25

Fig 26

Fig 27

Fig 28

Fig 29Fig 30Fig 31Fig 32

Fig 33Fig 34

Coordinate System of Inter-Range Instrumentation Group. .................Standard Spherical Coordinate System Used in Antenna Measurements ........Calculated Radiation Patterns Illustrating the Effect of Quadratic Phase ErrorsEncountered in Measuring Patterns at the Ranges Indicated .................Elevated-Range Geometry. ..........................................Possible Error in Measured Relative Pattern Level Due to Coherent ExtraneousSignals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Example Configurations of a 686 Meter Elevated Range with Diffraction Fences . .Example Configuration of a Diffraction Fence with Serrations. ..............Ground-Reflection-Range Geometry. ..................................Possible Layout for a Ground-Reflection Range ..........................Slant-Range Geometry .............................................Schematic Representation of a Compact Range Using a Reflector and Feed. ....Specular Reflections from Sidewalls in Anechoic Chamber. ..................Block Diagram of Typical Antenna-Measurement System. ..................Basic Heterodyne Receiving System Using Double Conversion and PhaseLocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Dual-Channel Heterodyne Receiving System for Phase Measurements. .........The Two Orthogonal Axes of Rotation Required by an Antenna Positioner UsingSpherical Coordinates ..............................................Positioner Configuration in which the Source Antenna is Supported by a Gantrythat Provides the 0 Rotation. ........................................Two Standard Positioner Configurations and their Associated Spherical Co-ordinate Systems. .................................................Model Tower and Its Associated Spherical Coordinate System ...............Polar and Rectangular Logarithmic Plots of a Normalized Radiation Pattern ....Power, Field, and Decibel Plots of the Same Antenna Pattern. ...............Radiation Distribution Table Recorded by Scanning in Q and Stepping in 8 .....Typical Field-Probe Mechanism. ......................................Spatial Interference Pattern Due to a Reflected Wave. .....................Amplitude of Spatial Interference Pattern Versus Ratio of Reflected-Signal toDirect-Signal Strengths for an Aperture-Probe Cut in the Plane of E, andE, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Illustration of How the Side-Lobe Level of the Test Antenna is Affected DuringAntenna-Pattern-Comparison Measurement. .............................Azimuthal Pattern Comparisons for Incremental Longitudinal Displacements ofthe Center of Rotation .............................................Amplitude Pattern of a Broad-Band Antenna Taken as a Function of Frequencywith Angular Coordinate Taken in 5º Steps .............................Geometry for Free-Space VSWR Method ...............................Geometry for Geometric Optics Approach to Focusing ....................Block Diagram of Automatic Position and Measurement System .............Amplitude Pattern of a Broad-Band Antenna Taken as a Function of Frequencywith Angular Coordinate Taken in 5” Steps .............................Schematic Illustration of Analytical Photogrammetric Triangulation ..........Geometry Used to Relate the Coordinates of a Point on the Surface of a Reflec-tor to the Measured Curvature. .......................................

1516

2122

232425262829293133

3637

38

38

40414344464849

50

53

54

55575961

6364

65

FIGURES PAGE

Fig 35Fig 36Fig 37Fig 38Fig 39

Fig 40Fig 41Fig 42

Fig 43

Fig 44Fig 45Fig 46Fig 47Fig 48

Fig 49Fig 50Fig 51Fig 52Fig 53Fig 54Fig 55Fig 56Fig 57Fig 58Fig 59Fig 60

Fig 61

Fig 62Fig 63Fig 64

Fig 65Fig 66Fig 67

System for On-Site Measurements of Amplitude Patterns. ..................Phase Shift of Single-Frequency Field Propagating in the x Direction. .........Arrangements for Measuring Phase Patterns .............................Phase Measurement Between Two Ports of a Multiport Antenna. .............Geometry and Phase Change as a Displaced Source is Rotatedabout a Given Origin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Illustration of the Sense of Rotation. ..................................Polarization Ellipse in Relation to Antenna Coordinate System ..............Relation Between Polarization Properties of an Antenna whenTransmitting and Receiving ..........................................Illustration of the Division of Power Between Two OrthogonalElliptical Polarizations A and B .......................................The Polarization Ellipses of Cross-Polarized Field Vectors ...................Poincare Sphere Representation of the Polarization of a Plane Wave W. ........Representation of Polarization on the Poincare Sphere. .....................Definition of Phase Reference for Orthogonal Circular Components ..........Polarizations of Incident Wave W and Receiving Antenna A,,Plotted on the Poincare Sphere .......................................Polarization Box and Its Relation to the Poincare Sphere. ..................Polarization Pattern of a Wave. .......................................Poincare Sphere Representation of the Polarization-Pattern Method. ..........Generation of the Polarization Pattern Using the Polarization Matrix Result ....Continuously Scanned Polarization Pattern as a Function of Angle 19 ..........Multiple-Amplitude-Component Method of Polarization Measurement. ........Instrumentation of the Phase-Amplitude Method of Polarization Measurement . .Phase-Amplitude Measurements with Circularly Polarized Antennas. ..........Phase-Amplitude Measurements with Linearly Polarized Sampling Antennas ....Three-Antenna Absolute Method of Polarization Measurement. ..............Two-Antenna System Illustrating the Friis Transmission Formula ............Typical Instrumentation for Two-Antenna and Three-Antenna Methods ofPower-Gain Measurement ...........................................Typical Instrumentation for Swept-Frequency Two-Antenna andThree-Antenna Methods of Power-Gain Measurement ......................Ground-Reflection-Range Geometry. ..................................Flux-Density Spectra of Several Radio Stars. ............................Typical Instrumentation for the Measurement of Antenna PowerGain and Temperature Using a Radio-Star Method ........................Signals Received by Tracking Antenna Versus Angle of Target ...............Orthogonal Arrays for Beam-Shift Measurement. .........................Decay of Surface-Wave Component of the Ground Wave for a Plane Earth. .....

68707273

747777

78

8080818282

838486878788899091929396

97

9898

102

105114118124

TABLES

Table 1 Information about Several Radio Sources [92]. . . . . . . . . . . . . . . . . . . . . . . . . . . 103Table 2 Errors in the Measured Gain of a Purely Circularly Polarized

Antenna Due to a Finite Axial Ratio of the Transmitting Antenna . . . . . . . . . . . . 109Table 3 Errors in the Measured Gain of a Linearly Polarized Antenna Due

to a Finite Axial Ratio of the Transmitting Antenna . . . . . . . . . . . . . . . . . . . . . . . 109Table 4 Electromagnetic Radiation Safety Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

APPENDIXES PAGE

Appendix A: Field Regions .......................... ...................... 139

Al. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I . . . . . . . . . . . . . . . . . . . 139A2. Bibliography .............................. . . . . . . . . . . . . . . . . . . . . . . 141

Appendix B: Reciprocity. ........................... . . . . . . . . . . . . . . . . . . . . . . 141B1. General.. . . . . . . . . . . . . . . . . . . . . . . 141. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .B2. Antenna Patterns. 141.......................... . . . . . . . . . . . . . . . . . . . . . .B3. Gain and Effective Area 142..................... . . . . . . . . . . . . . . . . . . . . . .B4. Expanded Reciprocity Relations. .............. . . . . . . . . . . . . . . . . . . . . . . 142

B5. Bibliography . . . . . . . . . . . . . . . . . . . . . . 143..............................

APPENDIX FIGURE

Fig Al Field Regions for Two Antenna Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . s 140

IEEE StandardTest Procedures for Antennas

1. Scope

This document comprises test procedures forthe measurement of antenna properties. It is acornprehensive revision and extension of theprevious test procedure ANSI/IEEE Std 149-1965 (Reaff 1971).

Throughout this standard it is assumed thatthe antenna to be measured can be treated as apassive, linear, and reciprocal device. Thereforeits radiation properties can be measured ineither the transmitting or the receiving mode.Many of the test procedures decribed can, how-ever, be adapted for use in the measurement ofantenna systems containing circuit elementsthat may be active, nonlinear, or nonreciprocal.For these cases there is no simple relationshipbetween the antenna system’s transmitting andreceiving properties. Therefore measurementsshall be performed for the mode or modes inwhich the antenna system has been designed tobe used.

A fundamental property of any antenna is itsradiation pattern. The measurement of radiationpatterns on an antenna range is discussed inSection 3, with the emphasis placed on amplitude patterns. The design of antenna ranges, orantenna test facilities, is described in Sec tion 4.

The instrumentation required for the antennarange, directions for the evaluation of an (exist-ing) range, and the operation of ranges are dis-cussed in Sections 5, 6, and 8, respectively. Avariety of special measurement techniques areincluded in Section 7.

The working environment in which anantenna is installed may substantially modifythe intrinsic pattern of an antenna. Conse-quently measurements in situ are frequentlyrequired. These are discussed in Section 9.

For each direction of space, the radiationpattern is characterized by amplitude, phase,and polarization. The latter characteristics aretaken up in Sections 10 and 11, respectively.

The relative amplitude-pattern informationmay be converted into absolute intensitiesthrough information derived from the measure-ment of antenna gain. The determination ofgain and closely related directivity is describedin Section 12. Errors in conventional gain cali-bration measurements are discussed particularlyin 12.5. Losses in the antenna itself can be ofimportance in some types of antennas, Theselosses can be accounted for by the radiationefficiency. Procedures for their determinationare treated in Section 13.

Section 14 deals with boresight measure-ments, which are concerned with the precisedetermination of the direction of the beam ortracking axis of an antenna system. The sensi-tive components of the antenna frequentlyrequire protection from harsh influences ofthe environment. The electrically transmissiveshield or radome which provides this protectionshall frequently be evaluated so that its effecton the radiation pattern is understood. Thistopic is treated in Section 15.

Power transfer from generator to antenna iscontrolled by the input impedance to the

13

IEEEStd 149-1979

antenna. This important parameter frequentlylimits the useful bandwidth of the antenna.Measurement procedures and network descrip-tions appropriate from low to microwave fre-quencies are presented in Section 16.

An important class of antennas relies onground to enhance the received signal. In thiscase the ground_ shall be considered as anintegral part of the antenna. The modificationof antenna concepts and additional data onthe ground-wave propagation are presented inSection 17.

The antenna and its associated circuits ratherthan the capacity of the transmitter generatormay limit the amount of power, either averagepower or peak power, that can be effectivelyradiated. It is desirable therefore to determinethese limitations as well as the environmentalfactors that may modify them independentlyof the system context. Procedures for testingpower-handling capacity are outlined in Sec-tion 18.

Another concern to the antenna engineer isthat of radiation hazards. It is well known thatradio-frequency (rf) fields of sufficient intensitycan cause damage to biological tissue. Thereforeit is usually necessary to determine the level ofthe radiation intensity in the vicinity ofantennas radiating high radio-frequency powerso that appropriate safety precautions can betaken before personnel enter the area. Thisimportant aspect of antenna measurements isdiscussed in Section 19.

Mechanical or structural properties alongwith environmental factors are described inSection 20. Because these properties are sovaried and specialized in nature, no attempt hasbeen made to include descriptions of actualmeasurements in this test procedure. Theenvironmental impact of an antenna is also animportant consideration for the antenna engi-neer. One aspect of environmental impact isthat of aesthetics. Large antenna structuresare necessarily conspicuous, and their appear-ance is of concern to those who live in theirvicinity. This is particularly true in an urbansetting. Since the aesthetic quality of theantenna structure is highly subjective, it isbeyond the scope of this document to suggestany evaluation procedure.

ANTENNA-RANGE MEASUREMENTS

Throughout this test procedure an attempthas been made to discuss measurement tech-niques as thoroughly as is practicable. How-ever, in general step-by-step procedural descrip-tions have been avoided. References are provid-ed which are illustrative of measurement tech-niques and in which details may be found.Because measurement techniques undergocontinuing refinement, the reader should bealert to references on the subject of antennameasurement that will have appeared after thistest procedure was prepared.

Many commonly used terms used in this testprocedure are defined in ANSI/IEEE Std 145-1973, Definitions of Terms for Antennas.Commonly used terms that do not appear intha t s tandard are i ta l ic ized in th is tes tprocedure.

2. Standards References

When the following standard documentsreferenced in the text are superseded by anapproved revision, the revision shall apply.

ANSI C95.1-1974, Safety Level of Electro-magnetic Radiation with Respect to Personnel.

ANSI C95.3-1973, Techniques and Instru-mentation for the Measurement of PotentiallyHazardous Electromagnetic Radiation at Micro-wave Frequencies.

ANSI/IEEE Std 100-1977, Dictionary ofElectrical and Electronics Terms.

ANSI/IEEE Std 145-1973, Definitions ofTerms for Antennas.

ANSI/IEEE Std 148-1959 (Reaff 1971),Measurement of Waveguides and Components.

IEEE Std 211-1977, Standard Definitions ofTerms for Radio Wave Propagation.

IEEE Std 291-1969, Standards Report onMeasuring Field Strength in Radio WavePropagation.

3. Antenna-Range Measurements ofRadiation Patterns

3.1 General. Associated with the antennaunder test is an operational coordinate system

14

OF RADIATION PATTERNS

IEEEStd 149-1979

8 = 9 0 º4 = 90º

*Assuming nopitch or yaw

Fig 1Coordinate System of Inter-Range Instrumentation Group

[ 1, pp 5.4-5.71],1 usually a spherical one. Gen-erally this coordinate system is defined by thesystem in which the antenna is used, althoughat times, for testing a specific antenna, it maybe necessary to define a different coordinatesystem. The Inter-Range Instrumentation Groupof the Range Commanders Council [2, p 1201,for example, has defined a coordinate systemspecifically for use with rockets, missiles, andspace vehicles (Fig 1).

The antenna’s coordinate system is typicallydefined with respect to a mechanical referenceon the antenna. A means of establishing the

‘Numbers in brackets correspond to those of the Bib-liography, Section 21 of this standard.

5

mechanical reference should be provided. Thestandard spherical coordinate system used inantenna measurements is shown in Fig 2.

To completely characterize the radiationfield of an antenna, one shall measure itsrelative amplitude, relative phase, polarization,and the power gain on the surface of a spherethe center of which is located at the antennaunder test. A representation of any of theseradiation properties as a function o f spacecoordinates is defined as a radiation pattern, orantenna pattern, of the test antenna. Since thedistance R from the antenna under test to themeasuring point is fixed, only the two angularcoordinates are variables in a given radiationpattern. Usually the radio frequency of opera-

IEEEStd 149-1979 ANTENNA-RANGE MEASUREMENTS

e = 9o”+ = 270’=

Fig 2Standard Spherical Coordinate System Used in Antenna Measurements

tion is treated as a parameter, with the radiationpattern being measured at specified frequencies.For some antenna applications it is necessaryto make frequency a variable. If frequency isvaried continuously, such a procedure is calleda swept-frequency technique; it is discussed in7.4. It is impractical to measure the radiationpattern of an antenna completely, and thereforeit is necessary to resort to sampling techniques.For example, with the frequency of operationand polarization fixed, the I3 coordinate can bevaried incrementally, and for each incrementthe desired antenna property can be measuredcontinuously over the range of 4. If the incre-ments are small enough, then for all practicalpurposes the complete antenna pattern is ob-tained. The resulting patterns taken for allincrements of 0 are usually referred to as a setof radiation patterns.

There are situations in which the operational

antenna illuminates structures in its immediatevicinity which alter the radiation field of theisolated antenna. In these cases it may benecessary for measurements of the radiationfield to include with the antenna those relevantparts of the nearby structures. The use of scalemodels is quite common for such cases (see 7.1).Throughout this standard the expression “testantenna” or, alternately, “antenna under test”shall mean the antenna itself plus any structureincluded with it. This means that the testantenna can be physically larger than theantenna alone.

3.2 Pattern Cuts. A direct method of measur-ing the radiation pattern of a test antenna isto employ a suitable source antenna, whichcan be positioned in such a manner that itmoves relative to the test antenna along linesof constant 13 and constant 4 (see Fig 2). Theloci of constant 6’ directions describe cones;

16

OF RADIATION PATTERNS

hence measurements made with 4 as the variableand 8 as a parameter are called conical cuts or4 cuts Those made with 8 as the variable and 4as a parameter are called great-circle cuts or 8cuts Note, however, that the conical cut for 8= 90” is also a great-circle cut.

Though rarely done, it is possible to positionthe antenna in such a manner that the loci ofdirections describe a spiral. For this case both0 and @ are variables, and the resultant motionis called a spiral cut. When spiral cuts are beingmade, it is usual for the motion in B to beslowly varying with respect to that in 4; hencefor each 360” rotation in $J the resulting pat-tern is approximately a conical cut.

Principal-plane cuts refer to orthogonalgreat-circle cuts which are through the axis ofthe test antenna’s major lobe. For this defini-tion to hold, the beam axis shall lie either inthe equator of the spherical coordinate system(0 = 90’) or at one of the poles (0 = 0’ or0 = 180”).

If the positioner system is designed to pro-vide 8 and @ cuts, then the alignment of theaxis of a pencil-beam antenna along the poles isusually avoided. This is because the @ cut forthe 8 = 0’ orientation yields only a polariza-tion pattern (see 11.2.2). When near the poles0 cuts involve radical changes in the direction ofpolarization of the incident field relative tothe test antenna’s polarization, except whenthe incident field is identically circularlypolarized. For these reasons if the test antennais of a pencil-beam type, it is usually orientedwith its beam axis in any desired direction inthe equator (most often in the Q = 0” orQ = 180” direction).

3.3 Basic Antenna-RangeConfigurations. Thereare two basic range configurations that accom-plish the position requirement for 0 and @ cuts[1, pp 5.12-5.24]. One is the fixed-line-of-sightconfiguration. Here the test antenna and itsassociated coordinate system are rotated abouta suitable axis (usually one passing through thephase center (see 10.3) of the test antenna). Ifthe test antenna is operating in the receivemode, then the signal that it receives from an

.7

IEEEStd 149-1979

appropriately located fixed source antenna isrecorded. The other one is called the movable-line-of-sight configuration. For this case thesource antenna is moved incrementally or con-tinuously along the circumference of a circlecentered approximately at the phase center ofthe antenna under test. If it is moved incre-mentally, then for each position of the sourceantenna the test antenna is rotated and thereceived signal is recorded. Alternately the testantenna can be rotated incrementally, and foreach of its positions the source antenna is movedcontinuously along its circumferential path.

If the test antenna and the source antennaare both reciprocal devices, the functions ofreceive and transmit may be interchanged.The measured patterns should be identical.This means that the test antenna may be usedin either the receive or the transmit mode;most often it is used in the receive mode. Inthe following, unless otherwise stated, the testantenna shall be considered as the receivingantenna, and it will be illuminated by thefield of the transmitting source antenna.

4. Antenna-Range Design

4.1 General. Antenna ranges have been de-veloped for the purpose of measuring theradiation patterns of antennas independent oftheir operational environment. The antennarange consists of the appropriate instrumenta-tion and the physical space required for themeasurements. In this section the design isdiscussed. Evaluation and use of antenna rangesare described in Sections 6 and 8. Emphasis isplaced on the measurement of the relativeamplitude patterns. The measurement of rela-tive phase, polarization, and power gain is dis-cussed in Sections 10, 11, and 12, respectively.It should be pointed out that the variouscriteria for antenna-range design presentedhere, as well as the methods of evaluation,apply equally well to other types of ranges usu-ally found at antenna testing facilities, such asradome ranges and scattering ranges. Radomemeasurements are discussed in Section 15, and

IEEEStd 149-1979

scattering ranges are beyond the scope of thisstandard.

The ideal incident field for measuring theradiation characteristics of the test antenna isthat of a uniform plane wave. In practice it isonly possible to approximate such a field.Attempts to do this have led to the develop-ment of two basic types of ranges:

(1) Free-space ranges. This type of range isdesigned in such a manner that all the effectsof the surroundings are suppressed to accept-able levels.

(2) Reflection ranges. This type of range isdesigned to judiciously use reflections in orderto produce an approximated plane wave.

Typical ranges that come under the free-space-range classification are: the elevatedrange, the slant range, the compact range, andmost anechoic chambers.

The elevated range [3] includes those rangesin which the test and source antennas arelocated on towers, on adjacent mountain peaks,on the roofs of adjacent buildings or ondiagonally opposing sides of abandoned quar-ries. Generally, however, it is designed over anapproximately flat area, and the effects of thesurroundings are suppressed :

(1) by careful choice of the source antennawith regard to directivity and side-lobe level

(2) by clearance of the line of sight alongthe range surface

(3) by redirection or absorption of energyreaching the range surface or obstacles thatcannot be removed

(4) by special signal-processing techniquessuch as modulation tagging of the desiredsignal

(5) by use of short pulses

The ground-reflection range [4] is designedto make use of the energy that is specularlyreflected from the range surface to create aconstructive interference with the direct-pathenergy in the region of the test antenna. Withproper design the illuminating field will have asmall, essentially symmetric amplitude taper.This is usually achieved by adjusting the heightof the source antenna above the range surfacewith the test antenna at a fixed height.

ANTENNA-RANGE

Since the elevated range over a flat surfaceand the ground-reflection range are so basic toantenna-pattern measurements, their designwill be discussed in detail in 4.3 and 4.4,respectively. Other range types will be dis-cussed in 4.5.

4.2 Antenna-Range Design Criteria4.2.1 General. To establish the design criteria

for either basic range type, one must considerthe following:

(1) the coupling between source and testantennas

(2) the transverse and longitudinal ampli-tude taper of the illuminating wave front

(3) the phase curvature of the illuminatingwave front

(4) spatial variations in the illuminatingwave front caused by reflections

(5) interference from spurious radiatingsources

Items (l)-(3) are discussed in this section.The problems associated with reflections fromthe range surface and other obstacles are dis-cussed in 4.3 and Section 6.

Outdoor ranges are usually subjected tointerference from signal sources -outside therange area. These sources can be mobile com-munications, radar, or telemetry systems.The use of filters can often suppress the effectof these interfering signals.

4.2.2 Effect of Mutual Coupling BetweenSource and Test Antennas. The total field ofany antenna consists of a radiation part and areactive part. The radiation field decays as thereciprocal of the distance from the antenna,whereas the reactive field decays at least asrapidly as the reciprocal of the square of thedistance from the antenna. Usually the spacingbetween source and test antennas is largeenough so that the level of the reactive field ofthe source antenna is negligible. However,practical situations do arise for which thismay not be the case. In these cases the testantenna will couple to the reactive field, andfor some types of measurements this producesan error. Such effects are considered negligiblewhen the spacing is greater than about 10wavelengths. Based upon calculations made fora very short dipole antenna at 10 wavelengths

18

DESIGNIEEE

Std 149-1979

from the antenna, the level of the inductivefield (which is the part of the reactive fieldthat would couple) will be 36 dB below that ofthe radiation field.

In addition to the reactive field coupling, thereradiative coupling or mutual coupling be-tween source and test antennas is also of con-cern [5]. For this case, part of the energyreceived by the test antenna is reradiated towardthe source antenna. In turn, part of the energyreceived by the source antenna is again re-radiated toward the test antenna. While thislevel is quite low, it can cause a measurableerror in the level of the signal observed nearthe peak of the test antenna’s major lobe. Theeffect on the side-lobe levels is usually negligible.A careful analysis [ 1, pp 14.3-14.5] was per-formed; it was based upon identical source andtest antennas, each with a (sin x)/x antennapattern, a 50 percent aperture efficiency, andan equivalent power reflection coefficient of0.25. The level of the reradiated signal reachingthe test antenna was at least 45 dB below thatof the original received signal when the ratioof the plane angle subtended at the sourceantenna by the diameter of the test antenna,to the 3 dB beamwidth of the source antennawas made equal to or less than 0.3. This ratiocorresponds to a subtended angle that isapproximately equal to the source antenna’sbeamwidth at the 0.25 dB level. It is generallyconsidered good practice to adhere to a 0.25 dBcriterion for the amplitude taper of the directilluminating field [ 1, pp 14.13-14.15], [4].

4.2.3 Effect of a Transverse Amplitude Taperover the Test Aperture. An amplitude taper ofthe illuminating field can produce an error inthe measured antenna pattern of the testantenna. This effect is dependent upon theaperture excitation function of the test an-tenna. For typical excitation functions theeffect of the amplitude taper of the incidentfield is a reduction in measured gain with aslight modification to the close-in side lobes.This effect can be evaluated by recognizingthat the variation of the amplitude of theincident field over the aperture of the testantenna on receive is analogous to a modifica-tion of the aperture illumination from its feed

on transmit. Suppose that the test antenna hasan aperture illumination from its feed on trans-mit given by f(x, y). If it is used on receive andis illuminated by an incident field with anamplitude variation given by g(x, y), then themeasured radiation pattern is essentially thesame as that for the transmitting case, with thefeed modified to produce a distribution overthe test antenna aperture given by f (x, y)-g(x, y).For example, consider a test antenna havinga circular aperture with a 10 dB cosine-on-pedestal distribution the far-field radiationpattern of which is to be measured with theuse of a source antenna that produces a circu-larly symmetric amplitude taper over theaperture of the test antenna. The taper isassumed to have a (sin x ) \x form. One cannow compute the directivity of the equivalentdistribution, which is the product of thecosine-on-pedestal and (sin x)/x functions forvarious tapers. If the amplitude taper variesto -0.5 dB at the periphery of the aperture, anapparent directivity reduction of approximately0.15 dB will result as compared to the casewithout taper. If the amplitude taper variesto -0.25 dB at the periphery of the aperture,one obtains a reduction in directivity of about0.10 dB.

Generally it is better to choose a sourceantenna that yields a 0.25 dB taper, rather thanone with a broader beamwidth which wouldproduce a smaller amplitude taper. This is be-cause the use of broader beamwidth sourceantennas usually results in an increased errordue to the reflections (see Section 6). Con-versely, there are situations in which it may benecessary to use a source antenna with a nar-rower beam in order to reduce the effect ofreflections. If a more directive source antennais used, the alignment of the source antennabecomes more critical. Care shall be exercisedin orienting the source antenna so that thepeak of its beam is centered on the test antennato prevent an excessive and asymmetrical il-lumination taper with a resultant increase inmeasurement error.

4.2.4 Effect of a Longitudinal AmplitudeTaper at the Test Antenna. To achieve a givenaccuracy in the measurement of the radiationpattern of a test antenna, the illuminating

19

IEEEStd 149-1979 ANTENNA-RANGE

field shall be sufficiently constant in amplitudealong the range axis, as well as in planes normalto the range axis. Consider an end-fire antennaunder test with an active region having a maxi-mum dimension L along the range axis. If theseparation between the source antenna and thecenter of the active region is R,, then theratio p,, of the power density at the forwardextreme of the active region to that at the rearis given by [7]

lOlog = 2Olog z” 1;;; [dBl0

Severe axial variations of the illuminating fieldcan cause a measurement error, particularly inthe minor lobe structure of radiation pat-terns. For most antenna types that have asignificant depth to their active regions, suchan error is usually considered to be negligiblewhen the power density over the region is con-stant to within 1 dB. This condition corre-sponds to an approximate restraint on R, of

R, 2 1OL

The criterion for such structures as high-gainend-fire antennas often is more restrictive thanthe greater of the previously discussed range-length criteria that were based on the suppres-sion of reactive-field coupling and phasecurvature.

4.2.5 Effect of Phase Variation over the TestAperture. A criterion has to be establishedfor the phase variation of the illuminating fieldover the test antenna. For most practical situa-tions the phase variation is a function solelyof the separation between the source antennaand the antenna under test. If, in the absenceof reflections, the 0.25 dB criterion for ampli-tude variation is adhered to, then the corre-sponding phase variation will be very close tothat of a spherical wave emanating from thephase center of the source antenna. This istrue for spacings considerably less than 2d’/X,where d is the diameter of the source antenna.For example, calculations [l, pp 14.5-l4.11]made for a source antenna with a 30 dB Taylordistribution revealed that its phase pattern dif-

fered from that of a spherical wave by at most2° between the 1 dB points of its amplitudepattern at a spacing of d* /X. Thus the expectedphase variation can be determined by assumingthat the phase front at the test antenna is asphere. It can readily be shown [4] that for atest-antenna diameter D and a separation R, aphase deviation A@ will be given by

A commonly employed criterion for deter-mining the minimum allowable separationbetween source and test antennas is to restrictA+ to n/8 rad. This results in the restrictionthat R 2 W* /h.

The effect of phase variation is that the nullsof the pattern are partially filled, and theamplitudes of the side lobes are changed. Anexample is shown in Fig 3 which depicts calcu-lated patterns of a circular aperture with a30 dB Taylor distribution at separations of2D2 /h, 4D2 /h, and approaching infinity. Itis possible to correct for phase deviations bychanging the focus of the test antenna, a tech-nique that is discussed in 7.2.

4.3 Design of Elevated Ranges4.3.1 Elevated Ranges over Flat Surfaces. In

view of the interdependence between thediameter of the test antenna and the range orseparation between source and test antennas,it is convenient to specify the separation asR = KD* JX, where K is a number to be chosenfor a particular measurement (see Fig 3). Itcan be shown [l, pp 14.15-14.28] that forelevated ranges the 0.25 dB criterion for ampli-tude taper leads to a criterion for the diameterof the source antenna, namely, that

d < 0.37KD

This assumes that the main lobe of the sourceantenna has a (sin x ) /x amplitude characteristic.

If the elevated range is designed over a flatsurface as shown in Fig 4, then steps shall betaken to suppress the reflections from therange surface in the region near the range axis.

20

DESIGNIEEE

Std 149-1979

0

- 1 0

m2

G3

s0CKw - 2 0

sw>

5-I:

- 30

- 40

HDU= (-_)sinO

A

Fig 3Calculated Radiation Patterns Illustrating the Effect of Quadratic Phase Errors

Encountered in Measuring Patterns at the Ranges Indicated.A 30 dB Taylor Distribution Is Assumed

As a basic design goal, the range surface in front antenna, namely,of the test antenna should not intercept anyenergy contained in the main lobe of the d > 1.5KD2/h,source antenna. It is suggested that the firstnull in the radiation patternantenna be directed toward theantenna tower. This leads toof a minimum-size restriction

of the source where h, is the height of the receiving antenna.base of the test This result is based upon the fact that typicalthe imposition source antennas have a main-lobe width ofon the source approximately 3X/d rad. A comparison of the

21


two criteria for the diameter of the sourceantenna shows that the height of the testantenna should satisfy the inequality

h, 2 40

where D is the diameter of the antenna undertest. The source antenna should be elevated tothe same height as the test antenna. Forphysically small antennas this requirement iseasily met. However, as the antenna’s size isincreased, it becomes more difficult, and insome cases impractical, to satisfy.

4.32 Elevated Ranges over Irregular Sur-faces. Elevated ranges are often designed overirregular terrain, for example, between twoadjacent mountain peaks or perhaps betweentwo hilltops. Such ranges are useful for thetesting of physically large antennas. For theseranges the location of the points of specularreflection, which reflect energy toward thetest antenna, may not be immediately obvious.When designing these ranges it is usually neces-sary to construct a scaled drawing of thevertical profile of the range, showing the exactground contour [8] from data obtained fromthe U.S. Geological Survey map of the area.These maps give detailed information on the

topography of the ground by means of contourlines plotted every 20 ft in elevation above sealevel, which sould be adequate for ranges operat-ing as high in frequency as UHF. For microwavefrequencies it may be necessary to locate thepoints experimentally using directive antennas(see Section 6). Once the points of specularreflection are determined, the level of thereflected energy at the test antenna can beestimated [8] , and corrective measures can betaken if the level of reflected energy is exces-sive. For these ranges the restriction on theheight of the test antenna is removed, sincethe mountain itself is part of the test-antennatower.

4.3.3 Estimation of Errors Due to Reflec-tions. To illustrate the effect of reflectionson the measured radiation patterns of a testantenna, consider the following examples. Sup-pose that the nominal side lobe levels of bothsource and test antennas are 25 dB below theirrespective maximum levels and the range sur-face produces a 10 dB attenuation of the re-flected wave. Then the extraneous signal levelwill be approximately 60 dB below the direct-path signal level for the case where the mainlobe of the test antenna is pointing toward thesource antenna. The graph of Fig 5 can be used

Fig 4Elevated-Range Geometry

TEST ANTENNA

22

DESIGNIEEE

Std 149-1979

(L

OUT-OF-PHASE

w

2 0 IO 0 -IO - 2 0 - 3 0

Em-_(dElIELI

2

4 if /

3 I

,0 0 0 2 1 j

-75 - 6 5 - 5 5 - 4 5 -35 - 2 5

$dE,

Fig 5Possible Error in Measured Relative Pattern Level Due to Coherent Extraneous Signals

Linear Scales Are Employed for Signal Ratios of + 20 to -30 dB; the Plus-or-MinusErrors Are Essentially Equal for Ratios of - 25 dB or Less, as Indicated in the

Logarithmic Plot for Ratios down to - 75 dB

to determine the possible error in the relativeamplitude pattern of the test antenna due tothis reflected signal. Fig 5 shows that themeasured value of the peak of the beam will bein error by less than 0.01 dB. However, if thepeak of the main lobe of the test antenna ispointed toward the range surface and a -25 dBside lobe is directed toward the source an-tenna, the relative level of the extraneous signalwill be only 10 dB below the level of thedirect-path signal, and the measurementerror is -3.3 to +2.4 dB.

This example illustrates the fact that it isnever advisable to direct the main beam of thetest antenna toward the ground when measur-ing side lobes. The operational procedure forthe range should be planned in such a way as toensure that this does not happen. A means ofrotating the test antenna through 180° about

the axis of the main lobe should be providedfor this purpose.

4.3.4 Use of Diffraction Fences and Longi-tudinal Ramps to Redirect Reflected Energy.When the antenna under test is essentiallyomnidirectional or when the required measure-ment calls for extremely small errors, thenother means of suppressing the reflected signalwill be required. Two methods which havefound wide usage are diffraction fences andlongitudinal ramps [ 1, pp 14.28-14.37].

Diffraction fences are metallic screens whichare strategically located on the range in such amanner as to redirect, from the antenna undertest, a portion of the energy that would, in theabsence of the fence, be reflected from therange surface toward the test antenna. Diffrac-tion fences can be of simple wooden frameconstruction covered with an ordinary conduct-

23

IEEEStd 149-1979

TOP VIEW

- 268 m

ANTENNA-RANGE

2,m 2 7 m 34 m

f14 m

mi

I- SHADOW REGION -

Fig 6Example Configurations of a 686 Meter Elevated Range with Diffraction Fences

ing screen that has a mesh size of approximately0.1 wavelength or less. When designing diffrac-tion fences care shall be exercised to reduce thediffraction by their top edges. The effect ofthis diffracted field can be estimated by meansof the theory of Fresnel diffraction by ahalf-plane. This leads to a result given in termsof Fresnel integrals, which are usually presentedin graphical form as Cornu’s spiral. The fencesshould be designed and located in such a waythat they will not be illuminated by the mainlobe of the source antenna. Thus it might bedesirable to use several fences, each with alower height than if a single fence were used.Additional suppression of the diffracted fieldcan be achieved by installing either tuned slotsor serratidns on the top edge of the fence.Tuned slots are more effective; however, theyare frequency sensitive. The fences should alsobe tilted, especially if multiple fences are used,to suppress multiple reflections.

The necessary width of the fences has beenempirically determined to span approximately20 Fresnel zones constructed on the rangesurface. A possible diffraction-fence configura-tion for the 686 meter elevated antenna rangeis shown in Fig 6, and a sketch of a typicaldiffraction fence with serrations is shown inFig 7. To save in cost, all fences are usuallymade identical in shape. The actual number offences required and their positions and orienta-tion should be determined experimentally(see 6.2).

A longitudinal ramp is a wedge-shaped re-flecting surface symmetrically placed on theground with its apex parallel to the rangeaxis. Its purpose is to scatter the incidentenergy away from the test antenna. It can pro-vide a moderate suppression of reflected energyat the test antenna; but generally it is lesseffective than a diffraction fence. Ideally thewidth of the base of the ramp should be such

24

DESIGN

IEEEStd 149-1979

IIm

3 m

Fig 7Example Configuration of a Diffraction Fence with Serrations

All Lines Indicate Supporting Structure; Not Shown Is theWire-Mesh Covering

that it spans about 20 Fresnel zones as laidout on the ground surface between source andtest antennas. The apex angle is usually chosento be about 120’. For many situations thesedimensions yield a prohibitively large structure,and a compromise is required, thus reducingthe effectiveness of the ramp. For those rangeswhere the grazing angles of the incident energyare very small, more Fresnel zones are spannedby a ramp of a given size. Hence for theseranges one would expect the ramp to be moreeffective. Another design factor is the construc-tion of the apex itself. Generally it is necessaryto round it off rather than having it as a sharpedge. This results in a degradation at the higherfrequencies. An obvious disadvantage of theramp is that it fixes the range axis. If either

the source antenna or the test antenna ismoved in the transverse direction, the sym-metry of the range is destroyed, and theunwanted reflected energy level at the testantenna increases. If the range has permanenttracks installed for the longitudinal motion ofthe antenna towers, then this is of little concern.4.4 Design of Ground-Reflection Ranges. Forthe ground-reflection range the amplitude taperof the illuminating field at the test antenna inthe horizontal direction will be the same as forthe elevated range, and the criterion for thesource-antenna size applies. In the verticaldirection the reflected signal from the surfacebetween the antennas combines with the directsignal to form an interference pattern (seeFig 8). By an adjustment of the height of the

25

LEE3d 149-1979

source antenna, the first interference lobe canbe peaked at the center of the test antenna [ 41.For a given frequency the height of the sourceantenna h, is given approximately by

ARht =c

r

where h, is the height of the test antenna. Theexpected field at the test antenna will bedependent upon the complex reflection coef-ficient of the reflecting surface. Attempts aremade, in the design of these ranges, to makethe magnitude of the reflection coefficientapproximately equal to 1.

The vertical amplitude taper of the illuminat-ing field is determined by the height of the testantenna. In particular if the 0.25 dB criterion isapplied, the height of the test antenna shouldbe

h, > 3 . 3 0

where D is the diameter of the test antenna [ 41,[9]. Usually it is recommended that thecriterion be

h, 2 40

It is not always practical to position physically

ANTENNA-RANGE

large test antennas at this height. For example,if it is necessary that the height of the testantenna be only 20, then the amplitude taperwill be about 1 dB. This yields less accurateresults but might be satisfactory for someapplications. As previously discussed, theeffect of amplitude taper is usually a reductionin measured gain and a slight modification inthe side lobes. Also if the excitation of thetest antenna is known, the error in gainproduced by a known amplitude taper can beestimated.

Thus far it has been assumed that the magni-tude of the reflection coefficient of the groundis equal to 1. Since this is not exactly true, thevirtual center of radiation will not be midwaybetween the source antenna and its image, butrather it will be some distance h: above thatpoint, as shown in Fig 8. It has been foundthat for any range length greater than 20’ /Xthe apparent phase center is located directlybeneath the source antenna at an approximateheight above the range surface given by

h; = l- PI1+ 1rj ht

Fig 8Ground-Reflection-Range Geometry

RD------/.

SOURCEANTENNA _ - - - - - -

u - - - / . . . - --L

RR/’/

26

DESIGN

where / r 1 is the magnitude of the reflectioncoefficient of the range surface for the polari-zation used at the angle of incidence determinedfrom the range geometry. It is therefore neces-sary to calculate the direction from the testantenna to the virtual center of radiation andto orient the test antenna accordingly by pro-viding the test antenna positioner with a tiltaxis so that the entire positioner can be ap-propriately tilted.

It is crucial to the operation of a ground-reflection range that the surface of the rangebe prepared in such a manner that the energyincident upon it is specularly reflected [lo] .This necessitates the specification of the“smoothness” of the surface. The Rayleighcriterion for smoothness is usually employed[ 1, pp 14.37-14.391, [3]. A useful form of itis given by

A h < - + - -M sin $

where Ah is the root-mean-square deviation ofthe irregularities relative to the median surface,X is the wavelength, $ is the angle an incidentray makes with the horizontal (the grazingangle), and M is the smoothness factor. Smooth-ness factors ranging from 8 to 32 or greater arecommonly used. This range of values corre-sponds to surfaces that are tolerable to verysmooth. Of course Ah shall be computed forthe highest operating frequency to be used.

The grazing angle $ shall be kept small inorder to avoid operating at the Brewster angle,which occurs at approximately 14” (measuredfrom the horizontal) over land [ll] . As theBrewster angle is approached, the magnitude ofthe reflection coefficient of the flat rangesurface decreases to a low value for waves thatare vertically polarized, whereas the reflectioncoefficient for horizontal polarization isunaffected. This is undesirable if the range is tobe used for both polarizations (see 6.4). Itshould be noted that the Brewster angle isdependent upon the electric properties of thematerial used in the construction of the rangesurface (and subsurface). For example, if therange surface were water, then the Brewster

IEEEStd 149-1979

angle would decrease to approximately 4’ [ 121.In order for the polarization characteristics tobe about the same as those for a range overland, the grazing angle shall be correspondinglysmaller. This usually means that the length ofthe range would have to be increased.

Obviously the region directly between testand source antennas is the most critical, and asone moves from the centerline of the range,the requirement for surface smoothness be-comes less demanding. A possible ground-reflection-range layout is depicted in Fig 9[l, pp 14.39-14.411. Notice that the range isdivided into three areas: the primary areawhich is a nominally rectangular region locateddirectly between source and test antennas, asecondary surface extending beyond the pri-mary area, and a cleared area beyond that. Thewidth of the primary surface is determined byconstructing Fresnel zones over the rangesurface centered upon the specular reflectionpoint that had been determined by the use ofgeometrical optics (ray tracing). The width ofthe primary surface should be such that about20 Fresnel zones are contained within itsboundaries. The surface tolerance of thesecondary region can be reduced by a factor of2 or 3. Its overall width should be about twicethat of the primary region, but it should ex-tend past the position of the test antenna byan amount of at least one quarter of the rangelength. In the actual design there would be noabrupt change in surface tolerance, but rathera gradual change from the primary to thesecondary regions. The cleared area should bemaintained by range personnel and kept freeof all major reflecting obstacles. If practicable,it should extend from the edge of the secondaryregion at the source end of the range to pointsdefined by the width of the first nulls in thehorizontal plane of the source antenna’s radia-tion pattern at the test-site end of the range.The cleared region should then continue as arectangular area beyond the test site to adistance approximately equal to one half therange length.

In an alternate form of the ground-reflectionrange, the test antenna is mounted with itspositioner on a nonconducting tower of suf-

27


I------ N ULL WIDTH AT R

SECONDARYSURFACE

PRIMARYSURFACE

/

TEST’ END

--I-

4

I

- I - -

- w 20 -

SOURCE END

Fig 9Possible Layout for a Ground-Reflection Range

ficient height so that the source antenna isplaced within a few wavelengths of the groundand the test antenna is centered upon the firstinterference lobe. This type of range is a varia-tion of the slant range discussed in 4.5 [9] ,[13]. It is common pract ice to cover thereflection region of the ground in front ofthe source antenna with a smooth conductingsurface. Since the source antenna is close tothe ground, the extent of the conducting sur-face is relatively small so that it becomeseconomically feasible.

Ground-reflection ranges find their greatestuse at lower frequencies, especially in the VHFregion of the spectrum, because of the dif-ficulty of suppressing reflections in equivalentfree-space ranges. However, ground-reflection

ranges have been designed to operate at fre-quencies as high as about 35 GHz.

4.5 Other Ranges4.5.1 Slant Range. The slant range [9], as

the name implies, is designed with the sourceantenna near the ground and the test antennaalong with its positioner mounted on a non-conducting tower at a fixed height, as depictedin Fig 10. The source antenna is located andoriented so that its free-space radiation-patternmaximum points toward the center of the testregion and its first null is pointing toward thespecular reflection point on the ground. Inthis way the reflected signal is suppressed.Alternate forms of the slant range are classedas reflection ranges in as much as reflections

28

DESIGNIEEE

Std 149-1979

\

Fig 10S l a n t - R a n g e G e o m e t r y

from the ground are used in order to approxi-mate a plane wave at the test region [ 131, [ 141.This type of slant range is a variation of theground-reflection range and is discussed in 4.4.Slant ranges have been designed with towerswhere the height can be changed so that thespacing between source and test antennas canbe varied. In general a slant range requires lessland for a given spacing than an elevated range.

4.5.2 Compact Range. The compact range[ 151, is one in which the test antenna is illu-minated by the collimated energy in the aper-ture of a larger point or line focus antenna. Forexample, a precision paraboloidal antenna canbe used to collimate the energy as shown sche-matically in Fig 11. The linear dimensions ofthe reflector is usually chosen to be at leastthree times that of the test antenna so that theillumination at the test antenna sufficientlyapproximates a plane wave. An offset feed forthe reflector is recommended to preventaperture blockage and to reduce the diffractedenergy from the feed structure which maycontaminate the field in the test region. Tofurther reduce the effects of the diffractionfrom the .feed structure and also to suppressany direct radiation from the feed antenna inthe direction of the test region, the reflectorcan be designed with a focal length longenough that the feed antenna can be mounteddirectly below the test antenna. If this is done,

Fig 11Schematic Representation of a Compact

Range Using a Reflector and Feed

then high-quality absorbing material can beplaced between test and feed antennas toabsorb the unwanted radiation. The use of arelatively long-focal-length reflector has theadditional advantage that for a given size reflec-tor the depolarization effect associated withcurved reflectors is lessened. Diffraction fromthe edges of the reflector can be retiuced bydesigning the reflector with serrations aboutthe edges.

In order to obtain good results with a com-pact range, the reflector shall be constructedwith sufficient accuracy. Small deviations inthe fabricated reflector surface can result insignificant variations in the amplitude andphase distribution of the incident field at thetest antenna. To assess the effect of surfacedeviations not only their shapes and maximumdeviations shall be specified but, also veryimportantly, their areas. For example, if thereflector has small deviations that do notexceed X/100 and their individual sizes arealso small (less than one square wavelength),then the integrated effect of all the deviationsover the entire reflector will be quite small,and hence a fairly uniform amplitude distri-bution of the incident field over the test an-tenna will be obtained. On the other hand,suppose the reflector had a single surface

29

IEEEStd 149-1979

deviation near the center of the reflector,extending over an area comparable to 1 Fresnelzone. Then a very significant change in theincident field would be expected [16]. Asindicated by this example, the range reflectorshah be fabricated with great care.

The compact range can be evaluated in thesame manner as conventional ranges by the useof field-probing techniques (see 6.2). Since theilluminating field is obtained by the reflectionfrom a curved surface, some depolarization isto be expected. The field probing should,therefore, include measurements of polariza-tion as well as amplitude and phase, especiallyif the measurements to be made depend uponthe polarization characteristics in the illuminat-ing field.

4.5.3 Image-Plane Range. The image-planerange consists of a precision ground plane sur-rounded by absorbing material or with serra-tions along its edges which reduce the edgeeffects due to its finite size. The antenna undertest is cut into two parts along its image plane,and the part above the image plane is mountedonto the ground plane.

NOTE: The image plane of an antenna refers to animaginary plane through the antenna over which thetangential electric field is zero.

The radiation pattern of the resulting configu-ration in the half-space above the ground planeIS equivalent to that of the entire antennaradiating into free space. (This is not true forthe input impedance.) That this is so can beshown by evoking image theory, hence thename image-plane range. The radiation patternabove the ground plane is measured by movinga source antenna in such a manner as to ap-propriately sample the field. This type of rangehas the obvious disadvantage that its use isrestricted to those antennas that have an imageplane, that is, antennas that exhibit linearpolarization over a plane with excitationshaving even symmetry for polarizations normalto the plane (sum mode) or odd symmetry forpolarizations parallel to the plane (differencemode). Perhaps even more restrictive is the factthat if the antenna is not designed to operateas an image-plane antenna, it shall be cut alongits image plane. This means that this range is

ANTENNA-RANGE

generally relegated to developmental andresearch work.

4.5.4 Anechoic Chambers. There are twobasic types of anechoic chambers, the rectangu-lar and the tapered types [ 171.

(1) The rectangular anechoic chamber i susually designed to simulate free-space condi-tions. High-quality absorbing material is usedon surfaces that reflect energy directly towardthe test region in order to reduce the reflectedenergy level. Even though the sidewalls, floor,and ceiling are covered with absorbing material,significant specular reflections can occur fromthese surfaces, especially for the case of largeangles of incidence. One precaution that canbe taken is to limit the angles of incidence tothose for which the reflected energy is belowthe level consistent with the accuracy requiredfor the measurements to be made in the cham-ber. Often, for high-quality absorbers, this limitis taken to be a range of incidence angles of0” to 70” (as measured from the normal tothe wall). For the rectangular chamber thisleads to a restriction of the overall width orheight of the chamber such that

RW>---

2.75

where R is the separation between source andtest antennas and W is the overall width orheight of the chamber. The actual width andheight chosen shall depend upon the magni-tude of the errors that can be tolerated andupon the measured characteristics of theabsorbing material used to line the walls. Ad-ditionally the room width and the size of thesource antenna should be chosen such that nopart of the main lobe of the source antennais incident upon the sidewalls, ceiling, andfloor.

(2) The tapered anechoic chamber is de-signed in the shape of a pyramidal horn thattapers from the small source end to a largerectangular test region, as shown in Fig 12.This type of anechoic chamber has two modesof operation. At the lower end of the fre-quency band for which the chamber is designedit is possible to place the source antenna close

30

DESIGN

J,‘.‘=‘LITUDE OF VvAvEFRONT

(b)

Fig 12Specular Reflections from Sidewalls in Anechoic Chamber

(a) Rectangular Chamber. (b) Tapered Chamber

enough to the apex of the tapered section sothat the reflections from the sidewalls, whichcontribute directly to the field at the test an-tenna, occur fairly close to the source antenna.Using ray-tracing techniques, one can showthat for a properly located source antennathere is little change in the phase differencebetween the direct-path and the reflected-pathrays at any point in the test region of thetapered chamber. The net effect is that theserays add vectorially in such a manner as toproduce a slowly varying spatial interferencepattern and hence a relatively smooth illumina-tion amplitude in the test region of the cham-ber. It should be emphasized that the source

31

IEEEStd 149-1979

antenna shall be positioned close enough to theapex for this condition to exist. The positionis best determined experimentally, although auseful way of estimating its required positioncan be obtained by drawing an analogy be-tween the tapered anechoic chamber and theground-reflection range (see 4.4). By use ofthis analogy the perpendicular distance h,from the source antenna to the chamber wallshould satisfy the inequality

ARh, < - -

4hr

where X is the wavelength, R is the separation

t

IEEEstcl 119-1979

between source and test antennas, and h, is theperpendicular distance from the fixed testantenna to the chamber wall. If the sourceantenna is moved forward in the chamber, thenthe interference pattern becomes more pro-nounced, with deep nulls appearing in theregion of the test antenna.

As the frequency of operation is increased,it becomes increasingly difficult to place thesource antenna near enough to the apex. Whenthis occurs, a higher gain source antenna isused in order to suppress reflections. It ismoved away from the apex, and the chamberis then used in the free-space mode similar tothe rectangular chamber.

The field does not spread in the manner of aspherical wave, hence the tapered chamber can-not be used for any gain measurement baseddirectly upon the Friis transmission formula,such as the two-antenna or three-antenna gain-transfer methods [ 61.

5. Antenna-Range Instrumentation

5.1 General. The instrumentation required foran antenna test range can be classified into fivesubsystems:

(1) source antenna and transmitting system(2) receiving system(3) positioning system(4) recording system(5) data-processing systemThe extent of the instrumentation required

depends upon the functional requirementsir:.posed by the measurements to be made. Theim.rumentation may range from very simplesysit ms designed for making only principal-plane patterns of simple antennas to highlyautomated systems in which computers areprogrammed to control all aspects of themeasurement, including the data processing forthe complete characterization of highly com-plex antennas.

A block diagram of the basic instrumentationelements of an antenna test range is shown inFig 13. The control units, indicators, receiver,recorders, and tuning units for the transmittersare usually located in a master-control console.

ANTENNA-RANGE

The signal source itself is usually located at aremote antenna tower.

5.2 Source Antennas for Antenna Ranges. Withthe exception of a few highly specialized in-stallations, antenna test ranges are designed tooperate over a wide band of frequencies. Thismeans that they shall be equipped with afamily of source antennas and transmitterscovering the entire band. The antennas shall,of course, have the beamwidths and polariza-tion properties consistent with the measure-ments to be performed on the range.

For frequencies below 1.0 GHz, log-periodicarrays are often used, and for frequenciesabove 400 MHz families of parabolas withbroad-band feeds are most often used. In somecases large horn antennas have been utilized. Ameans for controlling the polarization isusually required. As will be discussed in 5.5,this can be accomplished by mounting alinearly polarized source antenna on a polariza-tion positioner. Then the direction of polariza-tion can be continuously rotated. This featureis required for some types of measurements.Some other types of measurements requirecircular polarization so that a well-equippedantenna test range usually has families of bothlinear and circularly polarized source antennas.The circularly polarized antennas can be de-signed so that they can produce right-hand orleft-hand circular polarization, as well asorthogonal linear polarizations. Crossed log-periodic arrays are examples of such antennas.In addition, a set of gain-standard antennas isnecessary if power-gain measurements are to bemade using the gain-transfer method (see 12.3).

5.3 Transmitting Systems. The selection of thetransmitter depends upon several system con-siderations [l, pp 15.11-15.371. There are anumber of types of signal sources availablesuch as triode cavity oscillators, klystrons,magnetrons, backward wave oscillators, andvarious solid-state oscillators.

Whatever type of signal source is chosen, thefollowing performance requirements apply:

(1) Frequency control. A means shall beavailable to tune the signal source to the

32

INSTRUMENTATION

T E S TA N T E N N A

f Lb<,T E S TPOSITIONER

IEEEStd 149-1979

Fig 13Block Diagram of Typical Antenna-Measurement System

desired frequency. Depending upon the typeof signal source used, mechanical, electro-mechanical, or electrical means may be em-ployed. If the transmitter is remote, a means ofremote control is highly desirable. A calibratedposition servo loop or a rate servo loop can beused to control a motor that mechanicallytunes the oscillator. For the case of oscillatorsthat can be electrically tuned, an adjustable,regulated power supply is required.

(2) Frequency stability. Since the antennasand their associated radio-frequency circuitryare highly frequency sensitive, it is necessarythat the signal-source frequency remain con-

stant over the measurement period, which maybe in excess of 30 minutes. Frequency varia-tions of the order of 0.01 to 0.1 percent areusually acceptable. But for more precisemeasurements, a frequency stability of onepart in lo6 or better may be required.

The following are several methods of fre-quency stabilization in common use:

(a) control of the operating environment,such as by immersion of the oscillator in atemperature-controlled oil-bath

(b) coupling of the oscillator to a high-Qcavity

33

IEEEStd 149-1979

(c) electronically referencing the oscillatorto a microwave cavity (the Pound system) [ 181

(d) electronically referencing the frequencyto a stable source (automatic frequency control)

(e) phase locking the frequency to a stablesource (automatic phase control)

The particular scheme used depends upon thetype of oscillator used.

(3) Spectral purity. Some types of oscil-lators are rich in harmonics which, if trans-mitted, would cont,aminate the desired signal.In some cases spurious or nonharmonicallyrelated signals are generated. To eliminatethese unwanted signals one can either usefilters or employ a receiver that will discrimi-nate between desired and undesired signals.

(4) Power level. The required power outputof the signal source for a particular measure-ment is dependent upon the source and testantenna gains, the receiver sensitivity, thetransmission loss between the two antennas,and the dynamic range required for the measure-ment. Typical detectors used in antennainstrumentation systems have sensitivities onthe order of -50 to -65 dBm. If a receiver isused, this figure may be improved as much as30 to 60 dB. Power levels provided by typicalantenna-range signal sources range from 0 to33 dBm. Since most antenna tests are ampli-tude measurements, it is important that theoutput of the signal source be relatively con-stant. For most measurements a variation off 0.05 dB, which is easy to obtain, is notsignificant; but for some cases, such asinsertion-loss or gain-transfer measurements,variations of +_ 0.01 to + 0.02 dB for shortperiods of time may be required. For thesecases an automatic power-level control shall beprovided.

(5) Modulation. For some systems ampli-tude modulation is required; hence the signalsources should have that capability. There arecases where special pulse shaping is requiredto reduce the distortion of the pulse spectrum.

5.4 Receiving Systems. The receiving subsys-tem used in the an tennaampl i tude -pa t t e rnmeasurement system may be simply a bolo-

I )- 1

t

1

1!

34

ANTENNA-RANGE

meter detector (usually mounted directly onthe test antenna or, in the case of a scale model,inside the model) and its associated amplifier,the output of which supplies the signal to therecorder. With this system the transmitter isusually modulated. Also the dynamic range ofthe system is limited to the range over whichthe bolometer has a square-law characteristic,usually about 40 dB. The simple detector-amplifier receivingsystem finds wide applicationfor antenna ranges where extensive scale-modelwork is done or where quasi-isotropic antennasare tested. With this system the detected signalcan be fed to an amplifier by means of a high-impedance cable, thus reducing the reflectionsfrom the cable connecting the test antenna tothe console (see 7.1).

The simple bolometer-amplifier system lacksthe dynamic range, sensitivity, and frequencydiscrimination for most antenna measure-ments. This has led to the development ofsuperheterodyne receivers specifically for usein antenna measurements. These receivers aretypically designed to operate over an extremelywide range of frequencies, from a few megahertzto above 100 GHz. The harmonics of the localoscillator are used for operation in the milli-meter and submillimeter range and, for thecase of low-frequency reception, the receivedsignal is upconverted to a frequency withinthe range of the receiver [ 1, pp 15.38-15.491.

It is desirable that the receiver be physicallylocated in the master console, whereas theantenna under test may be remotely located.To avoid the degradation of excessive cableattenuation at lower frequencies and the in-convenience of long runs of waveguide athigher frequencies, the first mixer is usuallymounted directly on the terminals of theantenna under test. A single coaxial cablecarries the local oscillator power to the mixerand provides the return of the intermediate-frequency signal from the mixer to the receiver.This places a limitation on the allowable cableattenuation for proper mixing, usually about15 dB.

The receivers are often equipped for bothmanual and motor-driven tuning. The motorcontrol can be used for a mechanical automaticfrequency control and can be designed in such

INSTRUMENTATION

a manner that the receiver can be used forsivept-frequency measurements.

Antenna-pattern recorders are typically de-signed to operate at a fixed 1 kHz audio carrierfrequency; for these cases the output of thereceiver shall be 1 kHz. Antenna-pattern-rangereceivers are usually designed for continuous-wave reception, and the 1 kHz amplitude modu-lation is introduced in the receiver. The needfor remote adjustment of the modulation fre-quency of the signal source is thus eliminated.

A bolometer is most often used for detectingthe 1 kHz modulation of the intermediate-frequency carrier. The dynamic range of thereceiving system is limited by the square-lawrange of the bolometer (about 40 dB), unlesssome means of compensation is provided. Com-pensation can be accomplished by the introduc-tion of a bias voltage in the intermediate-frequency amplifier to vary its gain by the useof an auxiliary potentiometer connected tothe pen drive. The bias voltage can be con-trolled so that the receiver gain increases asthe recorder pen moves downscale. The systemdoes require a set of adjustments for properoperation, with which the dynamic range canbe increased to about 60 dB.

If the signal source is not sufficiently fre-quency stabilized to eliminate all significantfrequency drift, then an automatic frequencycontrol must be employed. The automaticfrequency control may also be needed tocompensate for local oscillator drift. It shallbe capable of operation over the dynamicrange of the input signal, that is, more than40 dB.

A useful circuit which is sometimes incorpo-rated into measurement receivers is one thatcompensates for small transmitter power-levelvariations. This circuit requires that the signalfrom the transmitter be monitored continuouslyduring the measurement.

If greater sensitivity, precision, and dynamicrange are required, the double-conversionphase-locked receiver may be employed. This type ofreceiver produces the required 1 kHz outputsignal by conversion rather than by square-lawdetection. Heterodyning to such a low secondintermediate frequency is made possible by

IEEEStd 149-1979

the use of phase-lock techniques. A block dia-gram of a single-channel double-conversionheterodyne circuit is shown in Fig 14 [ 1,pp 9.9-9.131. The phase-lock circuit causes thefirst intermediate frequency to be synchronous(locked) to the frequency of a highly stablecrystal oscillator. Because of the purity of thecrystal-oscillator spectrum, the intermediate-frequency conta ins a lmost no frequency-modulation components within the passbandof the phase-lock loop. This permits theintermediate-frequency signal to be convertedto 1 kHz by means of a second crystal oscil-lator, which is in turn phase-locked so that itsf requency dif fers f rom the intermediatefrequency by precisely 1 kHz. The bandwidthof the first intermediate-frequency amplifieris typically of the order of 10 MHz, whilethat of the second intermediate-frequencyamplifier is of the order of 100 Hz.

The noise bandwidth of the double-conversionsystem is approximately equal to twice that ofthe output bandwidth of the receiver, that is,usua!ly 200 Hz. This is in contrast to that ofthe conventional receiver employing square-lawdetection. The noise bandwidth of the latteris approximately dm; where B1 is thebandwidth of the intermediate-frequency amp-lifier and B,, the bandwidth of the postdetec-tion amplifier. Typically B, and B, are chosento be 1 MHz and 100 Hz, respectively. A con-siderable improvement in sensitivity is obtainedby using the double-conversion phase-lockedreceiver rather than the conventional receiver.

An additional advantage of the double-conversion phase-locked receiver is the factthat the change in output voltage is linearlyproportional to the change in input voltagerather than input power. In receivers withsquare-law detectors a 10 0OO:l change in out-put voltage is required to represent a 1OO:lchange in input voltage. The double-conversionphase-locked receiver requires only a 1OO:loutput ratio to represent a 1OO:l input ratio.It is quite easy to achieve a 1OOO:l outputratio, giving a 60 dB dynamic range. Althoughdynamic ranges of 80 dB can be achieved withextreme care in design, a 60 dB dynamic rangeappears to satisfy most measurement problems.

35


CHANNEL ARF IN P U T IF

IF AMPLIFIERMIXER

IF AMPLIFIER

m

45MHZ IkHz _

AMPLITUDEOUTPUT

MIXER

I I

Fig 14Basic Heterodyne Receiving System Using Double Conversion and Phase Locking

For some antenna-measurement applicationsthe rate al which data can be taken is of im-portance. While often the data rate is limitedby the rate at which the antenna under test (orthe source antenna in the case of a movableline-of-sight range) can be physically moved,there are situations where the data rate is limitedby the bandwidth of the output circuit of thereceiver. For these cases, in order to increasethe data rate, the output bandwidth has to beincreased. This will also reduce the sensitivity,and hence the dynamic range, of the systemso that there is a trade-off between increaseddata rate and system sensitivity. Usually inorder to increase the output bandwidth theoutput frequency shall also be increased. Sincethe sensitivity of bolometer detectors decreasesrapidly with increased modulation frequency,it is impractical to increase the output frequencyof a conventional receiver. With the double-conversion phase-locked receiver a detector isnot used, so one has a choice of output fre-quencies. For example, if the output frequencyis such that the output bandwidth can be in-

creased by a factor of 10, then for this receiverthe data rate will increase by a factor of 10with a corresponding decrease in systemsensitivity of 10 dB.

The phase-locked receiver requires a referencesignal that can be obtained either from areference antenna or, where distances areshort, by a direct sample from the signalsource.

If phase measurements are also required, thena dual-channel double-conversion phased-lockedreceiver may be used [ 1, pp 9.9-9.131, [19].A block diagram is shown in Fig 15. The circuitis similar to that shown in Fig 14 except that ithas two signal channels, designated A and B,which are fed by the same first and second localoscillators. The output voltage of each channelis directly proportional to the input microwavevoltage of that channel, and the signal level ineach channel can vary independently over thedynamic range of the system.

One of the primary areas of concern in a sys-

36

INSTRUMENTATION

IEEEStd 149-1979

IF AMPLIFIER IF AMPLIFIER

* 45MHz

CHANNEL A‘ I P H A S E , 45MHz

I -

Tlg_:

L

I ’

REFERENCEC H A N N E L

M I X E RA 2!! r - - - - l

45 MHz

RF INPUT

A ,ry--j~~~E

--Lcf-=CHANNEL B IF AMPLIFIER IF AMF‘LIFIER O U T

Fig 15Dual-Channel Heterodyne Receiving System for Phase Measurements

tern of this nature is that of eliminating inter-channel interference. By the use of isolators,proper shielding, and careful design, the inter-channel isolation can be maintained greaterthan 90 dB. An interchannel isolation of thismagnitude results in linearity errors over a60 dB dynamic range of less than ? 0.25 dB.

After the heterodyning of the two inputsignals to a final frequency of 1 kHz, phasemeasurement becomes essentially a time mea-surement. Fig 15 shows a second output fromeach of the 1 kHz intermediate-frequencyamplifiers feeding an axis-crossing sensor. AS

the name implies, each axis-crossing sensorsenses the time of the axis crossover of the1 kHz signal. Specifically it senses each positiveaxis crossing. At that instant each axis-crossingsensor provides a narrow pulse output. Measure-ment of the time between axis crossings ofthe signal in channel A and that in channel Bcan be directly translated to degrees at 1 kHz,

and consequently to the input radio-frequencysignal. With a- system of this type, amplitudevariations of the signals in the two signal chan-nels and the phase difference between themcan be measured simultaneously. Amplitudeand phase readouts can be provided in eitheranalog or digital form as required.

If the receiver is to be used in an automatedsystem, it is highly desirable that it be com-puter compatible, that is, its output should besuch that the receiver can be interfaced directlywith a computer. Also, an agile local oscillatoris required if the operating frequency has to bechanged during the course of the measurement.

5.5 Positioning Systems5.5.1 Antenna Positioners. For either the

fixed- or the movable-line-of-sight methods(see 3.3) two orthogonal axes are required iftrue 8 and 4 cuts are to be made. These axesaxe designated as the 8 rotational axis and the $rotational axis and are depicted in Fig 16. The

37


Q, R O T A T I O N A L

6 R O T A T I O N A L

A X I S

A X I S

Fig 16The Two Orthogonal Axes of Rotation Required by an Antenna Positioner

Using Spherical Coordinates

P O L A R I Z A T I O NROTATOR \

Fig 17Positioner Configuration in which the Source Antenna Is Supported by a

Gantry that Provides the 19 Rotation

38

INSTRUMENTATION

0 ais, which permits cuts in $I with 8 as theparameter, is the 02 axis of the coordinatesystem. The 8 axis, which provides for 8 cutswith $I as the parameter, is coincident with theline OA drawn through the origin normal tothe line OS, where OS defines the direction tothe source antenna. These axes are employedin all spherical-coordinate positioner configura-tions [l, pp 5.12-5.241.

A possible positioner configuration for themovable-line-of-sight system is shown in Fig 17.It consists of a gantry which provides the 0rotation and an azimuth positioner whichprovides the 4 rotation. This scheme is ofrestricted utility, since the distance betweenthe two antennas is limited. It is useful fortesting small antennas such as primary feedsfor reflector antennas. An alternate approachto the use of a gantry is to mount the sourceantenna on a carriage that moves along a fixedsemicircular arch centered upon a turntable onwhich the test antenna is mounted. By controlof the movements of the carriage and the turn-table both 8 and I$ cuts can be obtained. Archeswith radii as large as 19.8 meters have beenconstructed [20]. This type of range is pri-marily used for testing vehicle-mounted VHFand UHF antennas, as well as scaled models ofhigh-frequency antennas (see 7.1) [20].

In the case of the fixed-line-of-sight systemthe 0 and 4 rotations are provided completelyby the positioner for the test antenna. Twopositioners which meet the requirements forthis system are the azimuth-over-elevationand the elevation-over-azimuth types. Thesepositioners and their associated coordinatesystems are shown in Fig 18. An alternateexample of an elevation-over-azimuth positioneris the model tower (see Fig 19). Greaterflexibility can be obtained by the use of addi-tional axes; for example, azimuth-over-elevation-over-azimuth positioners are frequently em-ployed.

It is desirable that the operational coordi-nate system of the test antenna be coincidentwith that of the positioner. Otherwise theinterpretation of the measured data and theevaluation of errors become difficult. More-over, the physical characteristics of the test

IEEEStd 149-1979

antenna and the general shape of its radiationpattern are important consider’ations in theselection of a positioner. Indeed it is oftennecessary that a positioner be designed ex-plicitly to satisfy a specific measurementrequirement where nonstandard motions arerequired [ 31.

Specially designed positioners may be requiredwhen the radiation pattern of the test antennais essentially isotropic. For this case the posi-tioner has to be designed so that it does not sig-nificantly alter the radiation pattern of theantenna under test. One method is to suspendthe antenna on the end of a long nonmetallicboom. Another method is to m.ount the antennaon a plastic foam tower. For either of thesemethods, obtaining the motions required for acomplete set of radiation patterns is difficult.

If the measurement requires the use oforthogonal polarizations, then the sourceantenna sholuld be provided with a means tochange its polarization. Source antennas canbe designed in such a manner that their polari-zation can be electrically switched betweentwo orthogonal linear or circular polariza-tions. For the case of orthogonal linear polari-zations, a more common method is to usea source antenna positioner that is capable ofrotating a linearly polarized source antennathrough at least 180°, although for somemethods of polarization measurement thedirection of polarization shall be rotatedcontinuously through 360” during the mea-surement (see 11.2).

5.5.2 Antenna-Positioner Errors. The errorsintroduced in the measurement of radiationpatterns by positioners are angle errors orpointing errors. The following is an outline oferrors associated with positioners [ 1, pp. 5.32-5.411 :

(1) Geometric error(a) Coordinate-axis-alignment error. Im-

proper alignment of the coordinate system ofthe test antenna with that of the antennapositioner.

(b) Orthogonality error. Nonorthogonalityof the two rotation axes of the antenna posi-tioner.

39

IEEEsta 149-1979 ANTENNA-RANGE

c#a R O T A T I O N A L A X I S

&I R O T A T I O N A L A X I S

z0 ROTATiONAL A X I S

8 R O T A T I O N A LA X I S _

8 R O T A T I O N A L A X I S

+ ROTATICYALA X I S

8 R O T A T I O N A L AXIS4

+ ROTATIOF!A: IX:S

(4

Azimuth-over-Elevation Positioner.

b)

Elevation-over-Azimuth Positioner

Fig 18Two Standard Positioner Configurations and their

Associated Spherical Coordinate Systems

(c) Collimation error. Nonorthogonality of tioner. However, for well-designed positionersthe 19 axis with the line from the origin of the these errors are generally very small, and theirpositioner coordinate system to the source magnitudes can usually be determined from theantenna (line OS in Fig 16) manufacturer’s information. For errors due to

(d) Axis runout and axis wobble the installation of the positioner and the(2) Shaft-position error. Errors in the deter- mounting of the test antenna, range personnel

mination of the shaft-position angle do have some control. This is especially true in(3) Deflection errors. Structural changes in the case of mounting the test antenna onto the

the positioner because of thermal expansion positioner. Care should be exercised in theand contraction and changes in the forces design of the mount so that the test antenna’sapplied to the positioner coordinate system coincides with that of the

Some of the geometric errors are inherent positioner.in the design and construction of the posi- Shaft-position-angle information can be ob-

40

INSTRUMENTATION

IEEEStd 149-1979

8 ROTATIONAL AXIS

Fig 19Model Tower and Its Associated Spherical Coordinate System

tained by the use of synchro systems, potenti-ometers, or digital encoders. Typically com-mercial positioners use dual synchro systemswith the transmitters geared at ratios of 1:land 36:l with respect to each axis. The read-out system may be of either an analog type ora digital type. The analog type consists of asynchro receiver indicator, whereas the digitaltype requires an analog-to-digital converter.Readout errors of approximately 0.05’ to lessthan 0.01” are typical, depending upon theunit considered. Direct-drive digital encodersare more accurate than the geared synchrosystem and are used where high precision is

required. High-accuracy digital encoders mayeither be of the multipole-resolver type or ofthe optical type.

Deflection errors due to solar heating can bereduced by solar reflectors, insulating shields,reflecting paint, and barriers. Those due tochanges in bending moments applied at theantenna mounting surfaces can frequently bereduced by incorporating counter balanceweights in the design of the antenna mountin such a way that the change in bendingmoment is reduced.5.6 Antenna-Pattern Recorder. The antenna-pattern recorder provides a means of obtaining

41

IEEEStd 149-1979

a visual display of the antenna pattern. It isused to plot the relative signal strength receivedby the test antenna as a function of the angularposition of the antenna. The signal to be plottedis obtained from the output of a receiver ordirectly from a microwave detector, dependingupon the type of receiving system used. Theposition information is normally obtained fromsynchro transmitters or digital encoders gearedto the positioner axes. In addition to the mea-surement of signal strength, the relative phaseangle between two signals can be recorded if thereceiving system provides a dc output with anamplitude in proportion to the radio-frequencyphase angle.

Typical antenna-pattern recorders are electro-mechanical devices employing servo systems todrive the recorder axes. A chart servo systemusually positions the recording paper as a func-tion of the angular position of the antenna. Apen servo system positions a recording pen inresponse to the amplitude of the input signal.Ink-writing systems are mostly used in pref-erence to electric, thermal, pressure-sensitive,or photographic systems because of the highquality, high writing speed, reproducibility,economy, and simplicity of an ink system.

The antenna pattern may be recorded ineither polar or rectangular form. The polarform is often preferred for plotting patternsof antennas that are not highly directional.The polar format is particularly useful forvisualizing the power distribution in space.Such a plot is illustrated in Fig 20(a).

In the rectangular format [Fig 20(b)] thesignal amplitude is the y axis (ordinate) andthe position angle is the x axis (abscissa). Therectangular format permits narrow beam pat-terns to be recorded in finer detail because thepattern does not become crowded in regionsof relatively low gain as it does in a polar graph.To provide adequate resolution in a rectangu-lar display of patterns of different beamwidths,selectable chart scales are required.

An expanded rectangular-coordinate chartscale is obtained by deriving the angle infor-mation from a synchro transmitter geared tothe positioner axis, so that the synchro shaftrotates a number of times for one revolution of

ANTENNA-RANGE

the positioner turntable. A 36:l synchro ratiois commonly employed, resulting in one turnof the synchro shaft corresponding to 10’ ofpositioner rotation. This ratio is especiallyconvenient for a readout of the positionerangles with devices such as dual-pointer synchroindicators, since the 36:l ratio is compatiblewith the decimal numbering system.

NOTE: A ratio such as 16:l or 32:l is used in dual-speed synchro systems in which position angles areread as binary numbers.

With dual-speed 1: 1 and 36:l synchro trans-mitters the recorder chart scale (or one chartcycle) is 10’ of 360”. An intermediate 60’chart cycle can also be provided by gearinga synchro to the chart axis of the recorder ata 6:l ratio. Thus six rotations of the 36:lsynchro transmitter are required for one chartcycle.

X rectangular-coordinate recorder in whichthe chart paper is driven by the chart servo isreferred to as a strip-chart recorder. If the penis driven in both the x and the y coordinates,the recorder is referred to as an x-y recorder.Usually polar recordings are obtained by drivinga turntable with the chart servo. However, anx-y recorder can be used to generate polarplots by electronically performing the coordi-nate conversion. For practical reasons thetransformation is made using 1 :l synchroinformation only.

Antenna pattern recorders usually providefor a selection of amplitude functions. The pendeflection may be directly proportional to theamplitude of the input signal to the recorder,or the response may be proportional to thesquare, square root, or logarithm of the input.If the proper pen function is selected for theparticular type of detector used, either a linearor a square-law detector, the recording may bepresented as a function of the power, radio-frequency voltage, or relative signal level indecibels. This leads to the following patternformats.

Relative power patterns illustrate the varia-tion of power density at a fixed distance fromthe antenna as a function of angular coordi-nates [Fig 21(a)]. Power patterns are mostuseful in assessing small power variations be-

42

DESIGNIEEE

Std 149-1979

,

8

z IO2 2

53

4

w 6

$ 8

“, 2 0

$ 2W 4

’-t

6

ii 8

IY 3 0

2

46

8

40ANGLE

b)

Fig 20Polar and Rectangular Logarithmic Plots of a Normalized Radiation Pattern

(a) Polar Plot. (b) Rectangular Plot

(4

43

IEEEStd 149-1979

ANTENNA-RANGE

(a) (b)

(4

Fig 21Power, Field, and Decibel Plots of the Same Antenna Pattern

(a) Relative Power Pattern. (b) Relative Field Pattern. (c) Decibel Pattern.

44

INSTRUMENTATION

tween 100 and 10 percent of the peak value.Power variations in the range below 10 percentof full scale are difficult to resolve withoutchanging gain levels. Power patterns are used incalculating directivity, and sometimes inmaking measurements to determine half-Dowerbeamwidths of antennas because of the increasedresolution afforded at the higher signal levels.They would not be useful in measurementproblems where low side lobe levels and highfront-to-back ratios are of interest.

Relative field patterns show the variation ofthe electric field intensity at a fixed distancefrom the antenna as a function of angularcoordinates [Fig 21(b)]. Field patterns pro-vide greater side lobe resolution while stilldisplaying small variations in field level nearthe maximum. The half-power level on thisformat is 0.707 times the full-scale value.

If the recorder scales are linearly calibratedin decibels, a logarithmic relative gain patternresults. This format is particularly useful sinceantenna gain is usually specified in decibels.The decibel scale provides constant resolutionover the entire display range, and a widerdynamic range can be displayed than withother formats. Comparison of patterns is alsoeasier with a decibel scale, since a differencein system gain is equivalent to a constantoffset of the recorded data. A recording rangeof 40 dB is most often used, because this rangeis usually sufficient to show the side lobe levelsof interest, and it generally provides sufficientresolution for an examination of the main lobestructure [Fig 21(c)]. However, to achieve a 40dB dynamic range when operating from asquare-law detector, the recorder shall have an80 dB usable dynamic range. In order for therecorder noise to be small relative to the mini-mum input signal level, the noise level shouldbe approximately 110 to 120 dB below thefull-scale input.

There are other formats for the presentationof antenna patterns which have found wideuse. A complete radiation pattern can be pre-sented on a single page as a contour plot b yconnecting points of equal signal levels withlines to form isolevel contours in the manner

IEEEStd 149-1979

employed for topographic contour maps. Keylevels are numerically identified to permitinterpretation of the graph. One such systememploys different colors to highlight thecontours. These approaches usually requirethe use of computer graphics.

An alternate approach, which has found wideacceptance, is the radiation distribution tablewhich provides the same basic information asthe contour plot, but accomplishes it in asimpler manner. A portion of a radiationdistribution table is shown in Fig 22. Signallevels are printed numerically in decibels atpreselected intervals of 8 and 4. The graphshown was made by scanning in @ and steppingin 8 after each $J cut. In the figure the angularincrements are 0.5’ in 8 and 4, and the dynamicrange is 40 dB. Signal levels are sensed by anencoder coupled to the servo-driven logarithmicpotentiometer in an antenna-pattern recorderand printed on the data form.

The contour effect is obtained by printingthe even values of signal level and omitting theodd values or by underlining either the odd orthe even numbers.

5.7 Data-Processing and Control Computers.Al-though not shown in the block diagram of Fig13, an on-line instrumentation minicomputerprovides a natural solution to the data-gathering,control, and data-processing requirements of anautomatic antenna-measurement system. Instru-mentation computers can be equipped with avariety of input-output devices, depending onthe requirements of the particular measurementprogram. Magnetic tape, teletype, and paper-tape input-output units are commonly em-ployed for programming the computer andrecording and storing data. The computer alsobecomes a source of stored routines that canbe called up to meet the basic demands of thesystem.

In addition to solving the on-line data-processing and control requirements, the com-puter can be programmed for analysis andreduction of the measured data. For example,the polarization properties of the antenna can

45


I#I . DEGREESn R ::

Fig 22Radiation Distribution Table Recorded by Scanning in @ and Stepping in 8

Data Are Presented at 0.5’ Increments over a 40 dB Dynamic Range

46

EVALUATION

be calculated as well as its directivity. Com-puter plotters can be employed to provide avariety of visual displays of antenna patterns.Not only can contour plots be produced, butalso three-dimensional presentations. Forlengthy programs or for programming con-venience, the recorded data can be processedby a larger central computer at the user’sfacility.

6. Antenna-Range Evaluation

6.1 General. Once an antenna-pattern rangehas been designed, constructed, and instru-mented, it is necessary to experimentallydetermine the state of the illuminating electro-magnetic field over the test region, that is, theregion where the test antenna is to be mounted.Indeed, a continuing program of experimentalevaluation is usually required to determinewhether or not the errors experienced will beless than those required for the measurementsto be performed using the range. The extentof the experimental evaluation depends uponthe desired accuracies of the measurements tobe made on the range and the possible needfor official documentation of these accuracies.

The illuminating field over the test regionwill deviate from that predicted from calcula-tions based upon a highly ideal range geometrybecause of reflections from various mountingstructures, cables, obstacles on or near therange surface, and from irregularities in therange surface itself. In addition radio-frequencyinterference is often a cause of difficulty.

It is convenient to investigate separately thatPM of the field which is incident from thegeneral direction of the source antenna andthat which arrives from wide angles with respectto a line drawn from the center of the testregion toward the phase center of the sourceantenna. The former is often referred to asthe near-axis incident field, and the latter asthe wideangle incident field. The near-axisincident field may be determined from a field-probe measurement over a plane perpendicularto the range axis and coincident with the ex-pected location of the test antenna. This planeis called the test aperture.

IEEEStd 149-1979

6.2 Field-Probe Measurements over Test Aper-ture. Prior to making field-probe measure-ments, it is necessary to align the axes of thepositioner. For example, the vertical axis (r$axis in the case of an azimuth-over-elevationpositioner, or the 8 axis for an elevation-over-azimuth positioner, including the modeltower, can be aligned by use of a clinometer ora precision level. In the case of the model towera plumb bob can also be used. The horizontalaxis can be aligned optically. For the modeltower a scope can be mounted in the towerhead. In this way the scope’s optical line ofsight corresponds to the horizontal axis (@ axis)of the model tower, and thus alignment can bechecked. Once the axes have been aligned, thefield probe can be mounted. The field probeconsists of an antenna, usually a pyramidalhorn or a log periodic dipole antenna, mountedupon a carriage or track such that it can bemoved along an I-beam support. The entireunit is mounted on an appropriate positioner insuch a manner that the incident field may besampled as a function of position over the testaperture. A sketch of a typical field probe isshown in Fig 23 [l, pp 14.42-14.631. Themotion of the probe antenna is remotely con-trolled, and by the use of a synchro system orpotentiometers, which indicate the instanta-neous position of the probe antenna, a continu-ous automatic plot of the received field ampli-tude as a function of position can be made.These data are usually presented as decibelplots.

The probe antenna should be moderatelydirective in order to discriminate against wide-angle reflections and also reflections from theprobe-mounting structure. As an added precau-tion an absorbing baffle should be placedbehind the probe antenna. The E- and H-planebeamwidths of the probe antenna should besuch that a major portion of the range surfaceis contained within its half-power beamwidth.A useful criterion which ensures that thiscondition is met is given by

19s 2 2 tan-’

47


A L U M I N U MB E A MC A R R I A G E -

DRIVE MOTOR ANDS Y N C H R O P A C K A G E

\

UPPER AZIMUTHT U R N T A B L E

H O R NR O T A T O R

P R O B E A N T E N N AA N D M I X E R

P R O B E C A R R I A G E

- RF CABLE

- CONSTANT TENSIONC A B L E

CONSTANT TENSIONCABLE TAKE-UPR E E L

Fig 23Typical Field-Probe Mechanism

where h, is the height of the test antenna, Ris the range length, and 8, is the 3 dB beam-width of the probe antenna.

A means of rotating the probe antenna aboutits axis shall also be provided so that twoorthogonal polarizations, vertical and hori-zontal, may be measured. This is necessarybecause the range might be used with eitherpolarization, and hence the incident fieldshould be probed for both cases. It is goodpractice to measure also the cross-polarizedcomponents for both orientations of polariza-tion of the source antenna, because not onlymay the source antenna itself generate across-polarized component, but some depolari-zation usually accompanies the reflection ofwaves from irregular or curved objects.

One of the principal objectives of the probemeasurement is to determine the source of thereflections so that remedial action can be taken.If there is only one principal source of reflec-

48

tion contributing to the incident field, it is arather simple matter to locate it from themeasured data. Suppose that the direct wave isincident upon the aperture from a directionperpendicular to the plane of the aperture andthe reflected wave arrives at an angle 0 withrespect to that direction. From Fig 24 it is seenthat an interference pattern over the testaperture results. The spatial period P of theresultant waveform is given by

P=&

If the probe antenna were moved along a linein the test aperture which is formed by theintersection of a plane containing the direc-tions of propagation of both the direct and thereflected waves and the plane containing thetest aperture, then the measured field shouldbe that depicted in Fig 24. However, if the

EVALUATIONIEEE

Std 149-1979

Fig 24Spatial Interference Pattern Due to a Reflected Wave

probe were rotated to any other orientation inthe test aperture, then the spatial period wouldincrease and have a value given by

” = sin Bhcos o

where QI is the angle between the probe pathwhich yielded the result shown in Fig 24 andthe new orientation. From these results it isseen that by probing the field over radial linescentered on the test aperture the directionfrom which the reflected signal arrives can bedetermined. Furthermore the peak-to-peakamplitude of the interference pattern yields ameasure of the relative amplitude of thereflected wave E, to the direct wave ED. Thisratio, expressed in decibels, is given by

ER- (dB) = 20 log

- 1 + antilog (a/20)

ED 1 + antilog (a/20)’1where (T is the difference in decibels betweenthe maxima and minima of the measuredpattern. The ratio ER/ED (dB) is plotted as afunction of u in Fig 25. Generally there areseveral reflected waves, with the reflected wavefrom the ground being the predominant one.For this case the measured field is a compositespatial distribution. Usually, however, thePrincipal sources of reflection can be locatedfrom the data.

49

6.3 Incident-Field Measurements Near theRange Axis on an Elevated Range. For elevatedranges, if the amplitude distribution has exces-sive variations, then some means of eitherabsorbing the reflected energy or redirectingit from the test aperture shall be employed.In the case of reflection from the range, surfacediffraction fences usually are very effective.The field probe can be used in the adjustmentof the position, size, and orientation of thefences.

Another important use of the field probe isin the alignment of the source antenna. Sincethe source antenna is directive, any misalign-ment might yield an excessive asymmetricalamplitude taper in the illumination field overthe test aperture. The alignment procedureshall be performed in both the azimuthal andthe elevation planes, which corresponds tohorizontal and vertical motions of the field-probe antennas, respectively. A convenientmethod of alignment is as follows. First, froma horizontal traverse of the probe antenna twoequal power points are located for reference.The source antenna is then adjusted until thegeometrical mean of these two equal-powerpoints coincides with the center of the testaperture. Next the source antenna is rotated180” about its axis and again checked for align-ment. If the source antenna is still aligned, itindicates that the beam axis of the sourceantenna is coincident with the roll axis of thesource antenna’s positioner. If it is not, then


1.0987

6

ii= 01aW

9-I 8

k 7

z 6

0.01987

6

5-75 - 6 5 -55 - 4 5 - 3 5 - 2 5

$ (dB)

Fig 25Amplitude of Spatial Interference Pattern Versus Ratio of Reflected-Signal to Direct-Signal

Strengths for an Aperture-Probe Cut in the Plane of ER and ED

5 0

EVALUATION

Corrective measures shall be taken, that is,the source antenna shall be reoriented withrespect to the positioner mounting surface or,perhaps, in the case of a refleCtOr antenna its

primary antenna may need repositioning. Thenthe alignment procedure is repeated until thebeam of the source antenna is symmetric aboutthe roll axis of its positioner and at the sametime is properly aligned with respect to thetest aperture. This entire procedure is repeatedfor the vertical plane.

Once the effect of reflections has been re-duced to the point that the amplitude distribu-tion over the test aperture is found to besatisfactory and the alignment of the sourceantenna has been achieved, it is desirable thatthe relative phase over the test aperture bemeasured. In the absence of reflections, thephase distribution over the test aperture isprimarily a function of the spacing betweensource and test antennas. Therefore, for theelevated range, this measurement is one ofprecaution rather than necessity.

The one remaining characteristic of theincident field which may be measured is itspolarization. If the field-probe antenna isequipped with a rotator as is the one depictedin Fig 23, then polarization patterns can bemeasured at various positions in the testaperture (see 11.2.2). It should be pointed outthat the alignment of the source antenna shallbe completed before this measurement can bemade.

6.4 Incident-Field Measurements Near theRange Axis on a Ground-Reflection Range. Un-like the elevated range, the ground-reflectionrange is designed to utilize the reflection fromthe range surface in order to produce a broadinterference pattern in the elevation plane. Thesource antenna shall be positioned so that thebroadest maximum of this pattern is centeredupon the test aperture. This is accomplished byfirst locating the source antenna at the positionPredicted under the assumption that the reflec-tion coefficient of the range surface is equal to

minus one, that is, ht = $ . Then field-probe

measurements are made sling the intersectionOf the elevation plane and the test aperture.

IEEEStd 149-1979

The source-antenna height is adjusted until thefield in the vertical direction is symmetricalabout the center of the test aperture. Thisshould be done for both polarizations since theoptimum height for each polarization can bedifferent. The orientation of the source antennain the elevation plane usually does not have tobe corrected in the case of ground-reflectionranges since a small change in the source-antenna-beam pointing direction in elevationhas little effect on the desired interference pat-tern produced at the test aperture. In the hori-zontal plane, alignment can be accomplishedby use of the same procedure as outlined forthe elevated range.

In addition the test aperture should beprobed and, as in the case of the elevated range,sources of extraneous reflections should belocated. If necessary the reflecting objectsshould be removed, or absorbing baffles andreflecting screens used, to minimize the re-flected energy reaching the test aperture.

For polarization measurements it is neces-sary to orient the field probe so that the linearlypolarized probe antenna is pointing toward thephase center of the array formed by the sourceantenna and its image (refe’r to Fig 8). Aspreviously pointed out, this is not at the rangesurface but rather at an approximate heightgiven by

where 1 r 1 is the magnitude of the reflectioncoefficient of the range surface. /r 1 can betaken to be simply the amplitude ratio of thespecularly reflected wave from the range sur-face to the direct path wave as taken fromFig 25. Once this orientation has been achieved,then polarization patterns at various positionsin the test aperture can be made.

If the ground-reflection range is to be usedfor circular polarization, then it is highlydesirable that the polarization of the sourceantenna be adjustable so that precision adjust-ments in the polarization of the incident fieldcan be made. By varying the relative ampli-

51

IEEEStd 149-1979

tude and phase of the vertical and horizontalfield components of the source antenna, it ispossible to compensate for the effects of un-equal reflection coefficients exhibited by therange surface for the two field components.By transmitting the proper elliptical polariza-tion, the axial ratio of the incident field canbe adjusted to less than 0.1 dB at any givenposition in the test aperture. It should bepointed out that the characteristics of therange surface may change as a function of timeand frequency of operation.

6.5 Wide-Angle Incident-Field Measurements6.5.1 General. The two most commonly

used techniques for the evaluation of wide-angle fields are the antenna-pattern-comparisonmethod and the longitudinal-field-probemethod. These methods can be used on eitherthe elevated range of the ground-reflectionrange.

6.5.2 Antenna-Pattern-Comparison Method.The antenna-pattern-comparison method i sbased upon the premise that in the absence ofany reflected or extraneous signals the mea-sured azimuthal antenna patterns of a testantenna will be unchanged with small changesin the test antenna’s position with respect tothe source antenna. If, on the other hand,antenna patterns are measured for several dif-ferent positions of the test antenna and thepatterns exhibit changes from position to posi-tion, then this indicates the presence ofreflected or extraneous signals.

To illustrate the effect that reflected wavesmay have on the measured pattern of a direc-tive antenna, consider the situation depictedin Fig 26. A reflected wave is incident from adirection 8 degrees from the test antenna’sbeam axis. Usually the level of wide-anglereflected waves is at least 30 dB below thelevel of the direct wave. When the test antennais oriented so that its major lobe is pointingtoward the source antenna, the reflected waveis received on a side lobe. The effect upon themeasured level of the major lobe is negligible.However, if the antenna is rotated such thatthe major lobe is pointing toward the directionof the incoming reflected wave and a side lobeis pointing toward the source antenna, then the

ANTENNA-RANGE

apparent level of the side lobe can deviatesignificantly from that obtained in the absenceof the reflected wave. The actual deviationdepends upon the relative amplitudes andphases of the direct and reflected waves. Thusif the azimuthal pattern of the test antenna ismeasured at several positions along the rangeaxis, variations will occur in the resultingpatterns because of the change in the relativepath lengths of the two waves.

The antenna-pattern-comparison method thenconsists of recording the azimuthal patterns ofa test antenna for enough different positionsalong the range axis so that the maximumexcursions of the side lobe levels are obtained.It is convenient to record all the patterns on asingle chart from which the apparent directionof the incoming wave and its relative level canbe determined. An example of such a mea-surement is shown in Fig 27. Note that there isa variation of approximately 12 dB in thepatterns at an azimuth angle of 120’. If thedirect and major reflected waves were bothreceived on the side lobe, this would mean thatthe reflected wave is 4.5 dB below the levelof the direct wave. However, the reflected wavewas received on the main beam and the directwave on the side lobe. Since the variationoccurred approximately 30 dB below the peakof the main beam, the reflected wave is at least-34.5 dB relative to the level of the directwave. The graph shown in Fig 28 is useful inthe determination of the reflected signal levelsfrom measured data. Of course it takes manymeasurements to ensure that the greatest vari-ation is obtained; indeed, it is possible to missthe worst case conditions.

Another useful antenna-pattern-comparisonmeasurement can be made by measuring twoazimuthal patterns about the same center ofrotation with the test antenna rotated 180’about its beam axis between cuts. Also forconvenience, the direction of travel of theabscissa on the chart recorder can be reversedand the synchros appropriately adjusted so that,in the absence of reflected or extraneoussignals, the recorded pattern should be identicalwith that of the first cut. If there are reflec-tions, the patterns will not be identical unlessthe reflections are perfectly symmetrical about

52

EVALUATION

IEEEStd 149-1979

%

ER

\\

/”\e \

rJq tED ___--!&

(a)

it<]“;-__

l . . . .

2 NPOSITIONS

0 . . . l

2 NPOSITIONS

(b)

Fig 26Illustration of how the Side-Lobe Level of the Test Antenna Is Affected During

Antenna-Pattern-Comparison Measurement(a) Test Antenna Pointing Toward Source.(b) Side Lobe Pointing Toward Source

the range axis. From these data one can deducethe apparent direction from which the re-flected signal is incident upon the test antenna.

On some ranges, particularly where modeltowers are used, it might be desirable to makethe antenna-pattern comparison by takingidentical conical cuts at symmetric azimuth-Pointing directions with respect to the rangeaxis. For example, the azimuth positioner can

be rotated to an angle of 8 degrees, and theconical cut can be made by rotating the headof the model tower. Then the azimuth posi-tioner can be rotated to -8 degrees. It will benecessary to also rotate the test antenna 180’about the head axis to achieve the same startingpoint. After appropriately resetting thesynchros, the pattern is repeated and a com-parison is made.

53


3 = A Z I M U T H A L P A T T E R N A N G L E (360”)

Fig 27Azimuthal Pattern Comparisons (360” Cuts) for Incremental Longitudinal

Displacements of the Center of Rotation

6.5.3 Longitudinal-Field-Probe Methods. Analternate method of determining the level anddirection of reflected waves is by use of thelongitudinal-field-probe method. Unlike thefield probe previously described, which waspointed in a direction transverse to its path oftravel, the longitudinal field probe is orientedwith its beam axis along its direction of travel.The length of travel of the probe antennashould be great enough to detect maximumand minimum points of the interferencepattern formed by the direct and reflectedwaves. The approximate longitudinal distanceP, between the points of constructive inter-ference between the direct and reflected wavesis given by

P, =h

2 sin* (o/2)

where 8 is the angle between the direction to

54

the source antenna and the direction fromwhich the reflected wave is approaching [l,pp 14.63-14.661. Thus by measur ing theperiod of the amplitude variation of thereceived signal, the approximate direction tothe source of the reflected signal can be deter-mined for one reflected wave.

The level of the reflected signal relative tothe direct wave can be approximately deter-mined from the peak-to-peak variation in themeasured data. The radiation pattern of theprobe antenna shall be taken into accountwhen the data are analyzed. To obtain a morecomplete determination of reflected or ex-traneous signals it is usually necessary to makelongitudinal probe measurements along severaldifferent azimuthal directions.

6.6 Evaluation of Anechoic Chambers. Ane-choic chambers are designed in such a mannerthat the uniformity of the illuminating field

EVALUATIONIEEE

Std 149-1979

.,.0 - 5 -IO -15 - 2 0 - 2 5 - 3 0 - 3 5 - 4 0 - 4 5 - 5 0 - 5 5 - 6 0 - 6 5 - 7 0

PATTERN LEVEL CORRESPONDING TO ZERO REFLECTION (dB)

XOTE: The curvesupward. All curves

plotted as straight lines below the 5.7 dB ordinate value are actually veryare correct as plotted above the 5.7 dB ordinate value.

slightly concave

Fig 28Amplitude of Spatial Interference Pattern for a Given Reflectivity Level

and Antenna-Pattern Level

over a given region within the chamber meets the source antenna, that is, by specifying thecertain specifications. This region is usually a,mplitude and phase variations of the directreferred to as the quiet zone and is the region wave from the center of the quiet zone to thein which the test antenna is to be placed. The edge of the zone. Then the “quietness” of theactual size of the quiet zone may, for example, zone is dependent upon the magnitudes of thebe specified in terms of the direct wave from reflections from the walls, floor, and ceiling

55

IEEEStd 149-1979 SPECIAL MEASUREMENT

of the chamber. There is no standardized entire rear portion of the chamber, including .;’

figure of merit for anechoic chambers. What is the critical points such as corners can bedone is to establish the ratio of an “equivalent” evaluated.reflected wave, that is, the aggregate effect of As previously mentioned, the results of theall reflected waves incident upon the probe measurements depend upon the radiationantenna used to test the chamber, to the direct pattern of the probe antenna. One way towave. This means that the directivity of the avoid this effect is by the use of an isotropicprobe antenna will certainly affect the results probe antenna. For example, the probe can beobtained [ 211. a tridipole antenna consisting of three mutually

When an anechoic chamber is to be used for orthogonal dipoles which are designed in suchantenna-pattern measurements, usually the a manner that the antenna pattern is essentiallychamber is evaluated by means of the free-space isotropic. Such antennas have been designedvoltage-standing-wave-ratio (VSWR) method which respond equally. well (within + 1 dB) to[21], [22]. Th is method is similar in principle signals arriving from any direction and withto the field-probe method previously described, arbitrary polarizations. Of course great careexcept that instead of fixing the orientation shall be exercised in the design of the carriageof the probe antenna normal to the direction used to position the probe in order to avoid theof motion, it is made adjustable as shown in introduction of additional reflections in theFig 29. This allows for a more directive an- chamber. The advantage of this type of probetenna to be chosen for the probe. Usually a is that complete reflectivity information canstandard-gain antenna is used for this purpose. be obtained with only three orthogonal scansFor a given orientation the probe antenna is per frequency required. In general this systemmoved continuously along a transverse line, yields higher reflectivity levels than that ob-although other directions of travel may be tained with a directional antenna since thechosen. The linear motion of the antenna is equivalent reflected wave is composed of allcoupled to the recorder so that the inter- the reflected waves from all surfaces.ference pattern may be recorded. The graph of This technique is most useful when theFig 28 can be used to determine the ratio of antennas to be tested in the chamber are quasi-the reflected to the direct wave, provided the isotropic. If on the other hand the test antennascomponent of the direct wave received is are moderately or highly directive, reflectionslarger than that of the reflected wave [21]. from the rear of the chamber will be suppres-This entire procedure is repeated for various sed. For these cases the use of a directive probeorientations of the probe antenna. The orienta- for the free-space VSWR measurement may betion for which the maximum ratio is obtained more appropriate.represents the direction from which the equiva- The antenna-pattern-comparison method haslent reflected wave approaches. Measurements been used for evaluating anechoic chambers.should be made over several transverse direc- This method is not recommended as the pri-tions, including one horizontal and one vertical. mary technique for evaluating a chamber sinceAlso all measurements should be made with it is difficult to determine the maximumboth vertical and horizontal polarization. reflectivity levels [ 211 .

The wide-angle incident field, including thereflected signal from the rear of a chamber, canbe determined from a longitudinal probe(see 6.5). For this measurement the probe car- 7. Special Measurement Techniquesriage is oriented parallel to the axis of thechamber, thus providing the longitudinal travel 7.1 Modeling Techniques. Scale-modeling tech-for the probe. Again, free-space VSWR measure- niques are often used when the measurementments are made as a function of the angle of of an antenna in its operational environment isorientation of the probe antenna with respect impractical. This situation frequently existsto its direction of travel. In this manner the for antennas located on large supporting

56

TECHNIQUES

IEEEStd 149-1979

Fig 29Geometry for Free-Space VSWR Method

structures, such as ships, aircraft, and largeman-made satellites, which influence the prop-erties of the antenna. In moving systems orchanging environments, instability of the sup-porting vehicle or surrounding medium maycompel an unreasonable amount of experi-mental data, which require statistical treatment.-Another major situation, for which modeling isoften used, is in development work requiringsuccessive modifications which are costly dueto the large size or, occasionally, the very smallsize of the final antenna system. Thus the mainmotives for modeling are to obtain greaterexperimental control over the measurements,or to obtain an economic advantage in themeasurement program.

Generally the model is reduced in size fromthe full-scale antenna; but whether reduced orincreased, the requirements [ 23]-[ 251 forexact simulation by the model are as follows:

(1) Linear dimensions of the model shall bel/n times those of the full-scale antenna.

(2)’ Operating frequency and conductivity ofthe materials used in the model shall be IZ timesthose of the full-scale antenna.

(3) Permittivity and permeability of tl.materials used in the model shall have the same

values at the scaled frequency as at the originalfrequency.

In the preceding, n is an arbitrary number(generally, but not necessarily, greater thanunity) which determines the scale of the model.Also in the preceding, the imaginary parts ofthe complex permittivity and complex perme-ability are included in the expression for con-ductivity [26]. More general forms of scaling,which permit additional parameters to bechanged, are occasionally desired; these aredescribed in the literature [ 27 J .

In a practical model it is usually not feasibleto exactly satisfy the full set of requirements inthe preceding list. However, for antennas thatare not highly resonant the error will usuallybe small if good conductors such as copper oraluminum are used to simulate good conduc-tors, and if low-loss dielectric materials ofidentical electric permittivity and magneticpermeability are used to simulate low-lossdielectrics. The principal problem is presentedby poor conductors or lossy dielectrics. In suchcases it is not always possible to obtain materialsthat satisfy the scale-modeling requirements.

In constructing a scale model, not only doesthe antenna need to be simulated, but also

57

IEEEStd 119-1979

those portions of the surrounding structuresand environment which have an appreciableeffect on the properties of the antenna. Inmany cases it is difficult to construct simulatedsurroundings due to the complexity of theelectromagnetic environment. For example,when the antenna interacts with the earth, ithas often proven to be impractical to ac-curately scale the highly variable and sometimesunknown properties of the soil. In such casessimplified models are employed. Good judg-ment is required in determining the extent towhich one simulates the antenna’s surround-ings.

It is usually necessary to devise a means bywhich the scale model of the antenna alone canbe tested independent of its environment todetermine if it has the same electrical charac-teristics as the full scale antenna. One way ofaccomplishing this check is to build asimplifiedenvironment, such as a flat circular groundplane, to test the salient characteristics (forinstance, radiation patterns) for the full-scaleantenna. The simplified environment and an-tenna are then scaled.

If the results obtained using the scaled modelof the antenna sufficiently duplicate those ofthe full scale antenna then the scaled modelcan be used in the scale model of the antennaenvironment.NOTE: Usually a choke is required along the edge ofthe modelled circular ground plane to prevent currentsfrom being excited on the rear side of the plate.

In the case of an input-impedance and patternmeasurement using a scale model, all nearbyantennas shall be included in the model andterminated in the appropriate impedances.Since the other antennas may operate at fre-quencies different from that of the modeledantenna, the impedances at the operating fre-quencies shall be scaled correctly. To obtainthis information, the matching section may beterminated in a network equivalent [28] to theimpedance measured on the scale model.

Special measurements on antennas which areelectrically small, as is common at low frequen-cies, can often be made by quasistatic methodsusing electrostatic cages [29]. Here the chargeinduced by a known field strength is measured,allowing an equivalent area to be determined.

58

SPECIAL MEASUREMENT

While the radiation pattern is usually theantenna characteristic of most concern in modelmeasurements, other antenna properties ofinterest can also be reliably reproduced. If theexact scaling procedure described is followedthroughout the antenna system, all fields arereproduced exactly in shape, both externallyand within the feed line. Thus power gain,directivity, radiation efficiency, input imped-ance, mutual impedance, boresight error, andin general all properties dependent only on fieldratios are preserved. If the modified scalingprocedure is followed, efficiency and hencepower gain will not be reproduced, but theremaining properties will be reproduced accur-ately enough for most purposes, providing thatthe antenna does not have extremes of current,charge concentration or mismatch. Certainantenna characteristics, such as the power levelfor high-voltage breakdown (see Section 18)and the noise temperature, (see 12.43) cannotbe scaled because of the frequency-dependentnature of the mechanisms involved.

In the measurement of radiation patterns thefeed cable may perturb the measured quantitiesappreciably. In that event the scale-modelantenna may transmit from a battery-operatedtransmitter, or, alternately, the scale-modelantenna may contain a receiver from which thedemodulated signal can be removed by meansof high-resistance wire leads which minimizethe disturbance of the radio-frequency field[30]. Another approach is to use a semicon-ductor laser and transmit the signal via anoptical fiber. The support for the model shouldreceive particular attention if the pattern nullstructure is to be accurately determined.

In addition to the special problems of scale-model measurements referred to in this section,there are general procedures and precautions to

be employed during any of the measurements.These procedures are essentially the same asthose described in Section 8.

7.2 Antenna-Focusing Technique. For sometest situations it is difficult or impractical tomeasure antenna characteristics using far-field-range techniques. A technique that canbe used to measure the far-field antenna pat-tern at reduced ranges is that of focusing the

TECHNIQUES

test antenna at the range at which the measure-ment will be made [16], [ 311. This approachis limited to those test antennas which areprovided with a means of changing their focusfrom infinity to a finite distance. This canusually be accomplished with phased arraysand reflector-type antennas. Considerable workhas been done with paraboloidal reflectorantennas, since they are the most common ofthe large antennas. The geometry for the useof ray optics in establishing the focusing of aparabola is shown in Fig 30. When the feedantenna is located at point F, the antenna isfocused at infinity. By moving the feed toposition F’, the rays are brought into a quasi-focus at F ” . It is not a true focus since thereflector shape is paraboloidal rather thanellipsoidal, which is required for two finitefoci.

The required distance E that the feed shall bemoved can be determined by ray tracing and isexpressed approximately as

E % + [f2 + (LJJ2]

where R is the distance OF”, f is the distanceOF and E is the distance FF’ as shown inFig 30. Once the feed is moved to point F’,the antenna’s radiation pattern is measured ata range R. After testing the feed is returned toits original position at point F. For ranges asshort as O2 /8h the measurement will yield afairly accurate description of the main lobe of

IEEEStd 149-1979

the far-field pattern of the antenna focused atinfinity; however, the side lobe description willbe in error. An improvement can be made bymaking a fine adjustment of the feed aboutpoint F’. One approach is to adjust the posi-tion of the feed for a maximum depth of thefirst null. This yields a better description ofthe main lobe. However, if the power gain ismeasured on the focused antenna at range R ,it will be low by several tenths of a decibel ascompared to a far-field measurement. Anotherapproach is to adjust the feed for the maxi-mum power gain in the direction of the peak ofthe main lobe of the focused test antenna asmeasured at a range R. This will yield a slightimprovement in the measurement of powergain over that which would be obtained usingthe maximum depth of the first null criterion.Despite these shortcomings this method hasbeen found to be very useful.7.3 Near-Field Probing with IMathematicalTransformation. Another near-field techniquefor the determination of antenna characteristicsis that of near-field probing with mathematicaltransformation. In this technique the complex(phase and amplitude) vector field is sampledover a well-defined surface. The measured dataare computer processed to obtain an angularspectrum of plane, cylindrical, or sphericalwaves appropriate to the measurement surfaceemployed. The angular spectrum is then cor-rected for the directive and polarization effectsof the measuring probe, and far-field parameterssuch as power gain, relative pattern, and polari-

Fig 30Geometry for Geometric Optics Approach to Focusing

59

IEEEStd 149-1979

zation are calculated from the corrected spec-trum. In the general theoretical formulation thefield of the test antenna is represented by asuperposition of basis fields [ 321, which are solu-tions to Maxwell’s equations on the surface be-ing used. These would be plane waves expressedin terms of complex exponentials for a planarsurface, cylindrical waves expressed in terms ofBessel functions and complex exponentials fora cylindrical surface, and spherical wavesexpressed in terms of associated Legendrefunctions and spherical Bessel functions for aspherical measurement surface. The expressionfor the response function of the probe in termsof the same basis fields is obtained by fittingits complex vector-receiving pattern with thecorresponding basis functions. The outputvoltage of the probe as it moves over the sur-face may now be expressed in terms of theprobe response which is known, the basis fieldsof the test antenna, and the effect of theorientation and motion of the probe.

These ideas, as well as the concepts associatedwith measurement techniques and computa-tions, can best be understood by referring tothe example of measurements on a planarsurface. Much of the work in near-field mea-surements has been done for planar measure-ments [ 32]-[ 351, and the concepts associatedwith cylindrical or spherical surfaces can beviewed as an extension of this more familiarcase. For the case of the planar surface, thefar-field response of the probe is denoted byso1 (K), where K is the x-y part of the wavevector k, K = kxex + k,e,. so1 (K) is a com-plex vector function, and its value for eachvalue of K gives the response of the antennato a plane wave incident from that direction.The magnitude of each component of so1 (K) isidentical with the far-field patterns of theprobe, and soI (K) may therefore be deter-mined from a conventional pattern measure-ment where phase as well as amplitude aremeasured. If P denotes the x-y position ofthe probe in the plane z = d, tlo(K) the trans-mitting characteristic of the test antenna to bedetermined, and y = +_ k,, then the complexsignal at the output of the probe is given by

60

SPECIAL MEASUREMENT

dk ,

NOTE: In this section the time dependence is e-iwt(see 10.1).

The factors eird and eiK’ p are the result of themotion of the probe, which in this case can beexpressed quite simply as compared to thecases of cylindrical and spherical motions. Theequation for V(P) may be inverted to obtainone equation for the two components of~,oW) givenby

so1 W).t,o W) = emiKeP dx dy

A required second equation is obtained byrepeating the measurement with a second“independent” probe, which may be obtainedby rotating a linearly polarized probe about itsaxis by 90’. From the two equations bothcomponents of tlo (K) may be determined aswell as the far-field parameters. For instance,the power gain is given by

G(K) = -4 W/W2 / t,o (K) 1’--.I- py2

where rt is the antenna reflection coefficient.The essential components of a measurement

system are shown _schematically in Fig 31. Thescanner is a large mechanical structure whichresembles an x-y recorder. It supports theprobe and moves it over the plane to performthe measurements. The scanner shall be preciselymade so that the probe maintains a planar mo-t ion (62 < X/100) and the x--y pos i t ion ofthe probe is accurately known. Data shall bemeasured at equally spaced points in both thex and the y directions over an area somewhatlarger than the area of the antenna. The sizeof the area essentially determines the maxi-mum angle to which accurate far-field datacan be calculated through the relationship

emLX-ax

z tan-l ___2d

where a, and L, are, respectively, the antennadimension and scan length in the x direction.

TECHNIQUESIEEE

Std 149-1979

REFERENCE1 SIGNAL

DIGITALPOSITION

PC DISPLAYw

VERTICALI

( S CA NN E R ) ~yuNTTcp~~Y AXIS

BCDOUTPUT DIGITAL

I- __

POSITIONERDIGITAL BCD

,,‘,“,‘d~~,“,, CRZ~IAND

_ POSITION OUTPUT r - -

HORIZONTAL DISPLAY I * ISYNCRO X AXIS I IOUTPUT ’ I

POSITIONCOMMAND I

L-lIz!t+‘I,PROBE

I I

OPTIONAL INPUT TO COUPLER

r - - - - - - - - iI I‘- I

POSITIONER MOTOR CONTROL

SCAN (X OR Yl

Fig 31BIock Diagram of Automatic Position and Measurement System

In general the spacing between data pointsshould be slightly less than h/2. For broad-beam antennas the spacing may have to be assmall as h/4, while for highly directive antennasit may increase to approximately 1 wavelength.Tests should be performed [36] on the actualantenna-probe pair to determine the optimumspacing and scan area as well as to verify thatmultiple reflections are small enough to beneglected.

The major computation is the integrationrequired to evaluate so1 (K) *to1 (K), which isa two-dimensional Fourier transform of com-plex data. The fast Fourier transform is usedto accomplish this integration with greatefficiency [ 37 3 .

Planar near-field techniques have been usedfor a variety of antennas at frequencies fromabout 1 GHz to 65 GHz. The lower frequencylimit is due primarily to the effectiveness ofabsorbing material and the beamwidths ofantennas, while the upper limit is due to theaccuracy of positioners. In general the planartechnique works very well for directive an-tennas, and the errors are largest for angles

r-r-l

( OUTPUT

L 1

close to + 90’ from the normal to the mea-surement plane. If there is significant energyat these wide angles, a prohibitively large scanarea may be required. It is for such cases thatthe cylindrical or spherical scan surfaces aremore attractive.

For the cylindrical scan surface [38] thetheory and computations are somewhat moreinvolved than for the planar case. The majorcomputations can still be performed using thefast Fourier transform, and compensation forthe directive effects of the measuring probe isstill included. X cylindrical scan surface may beobtained by a onedimensional motion of theprobe plus a rotation of the test antenna onan azimuth positioner. This reduces the com-plexity of the scanner, but will increase theamount of microwave absorber required.

When a spherical scan surface is used, themathematics become much more involved[39] -[ 411, and while the fast Fourier trans-form can still be used for a large part of thecomputations, there still remains a sizableeffort to obtain far-field parameters frommeasured data. A complete probe correction

61

IEEEStd 149-1979

can only be obtained for certain ideal or simpleprobes. In principle this scan surface is bestsuited to broad-beam antennas, and a “self-scan” is possible by mounting the test antennaon a two-axis positioner and leaving the probefixed. This is especially attractive for largeantennas already mounted on a positioner,since the scan surface could be generated with-out additional hardware.

7.4 Swept-Frequency Technique. When theantenna patterns of very broad-band antennas,such as the log-periodic types, are measured atdiscrete frequencies over their operating bands,it is possible to miss significant variations intheir amplitude patterns [42] . These variationsare typically frequency dependent and narrowband. They are often referred to as “anomalies”in the pattern. Usually it is possible to optimizethe design of these antennas in such a waythat the “anomalies” are removed [43] , [ 441.

It is necessary to employ the swept-frequencytechnique in order to determine the frequenciesat which the “anomalies” occur. Principalplane cuts are usually all that are required, andfor linear polarization these would be in theE plane and the H plane of the test antenna.For the measurement one of the angular spacecoordinates of the test antenna is fixed, whilethe other is varied incrementally over the angu-lar range of interest. For each angular incre-ment the amplitude of the received signal isrecorded continuously as the frequency of op-eration is swept over the operating band of thetest antenna. All the curves are recorded onthe same chart so that a family of curves isobtained. If the gain and amplitude patternsof the test antenna were ideal, that is, invariantunder a change in frequency, then the measuredfamily of curves should consist of approxi-mately parallel straight lines with a -6 dB peroctave slope on a log-frequency scale, aspredicted from the Friis transmission formula.To obtain such a result, the antenna rangewith its instrumentation would also have to beideal.

If the measurements are made on a goodfree-space range for which an optimum broad-

62

SPECIAL MEASUREMENT

band source antenna and a receiving systemwith an approximately flat frequency responseare chosen, then good results can be obtained.The curves shown in Fig 32 are typical results.

Frequency-dependent reflections and aninsufficiently flat frequency response of theinstrumentation are the principal sources oferror. If the response of the instrumentationis known, then corrections to the data may beaccomplished. There is an alternate approachthat may be employed. This method is basicallya swept-frequency gain-transfer measurement.A single swept-frequency measurement is madeusing a reference antenna, and the data arestored. Then all data taken on the antennaunder test are electronically compared to theresponse of the reference antenna. It is the dif-ference between the response of the referenceantenna and that of the test antenna that isrecorded. This can be accomplished by the useof a data normalizer or by use of an on-lineminicomputer. The relative amplitude patternof the test antenna can be obtained for anygiven frequency in the band by plotting therelative levels for each increment of angle as afunction of angle.

The reference antenna should have an ampli-tude pattern similar to the antenna under test,so that the response to reflections on the rangeis about the same for the reference antenna asfor the test antenna. In this way the effect ofreflections tends to be removed. Obviously, asthe test antenna is rotated away from boresight,the error due to reflections increases. Note thatit is possible to use the test antenna itself as thereference antenna. This results in the first curvebeing a straight line.

If the measurements are made in a taperedanechoic chamber, it is necessary to ensure thatthe source antenna is sufficiently close to theapex of the chamber so that deep nulls do notappear in the illumination fie!d (see 4.5.4). Itis advisable that the swept-frequency techniquebe used to check the position of the sourceantenna.

The swept-frequency technique is particularlywell suited to the measurement of cross polari-zation in the far field of an antenna [45] -[46].

TECHNIQUESIEEE

Std 149-1979

0 . 5 0 . 6 0 . 7 0 . 8 0 . 9 1 . 0

F R E Q U E N C Y (GHz)

Fig 32Amplitude Pattern of a Broad-Band Antenna Taken as a

Function of Frequency with Angular Coordinate Taken in 5” Steps.Top of Line Is Broadside Case.

The reason for this is that cross polarizationoften is generated in mechanically compli-cated support structures in which a variety ofelectrical resonances can exist, thus producingscattered radiation the amplitude of whichvaries rapidly with frequency. It should benoted that the relevant frequency range neednot be great if the antenna structure is largein wavelengths.

Power-gain measurements can also be per-formed as a function of frequency [47].Such methods are discussed in 12.2.2 and12.3.1.

7.5 Indirect Measurements of Antenna Char-acteristics. The performance of a reflectorantenna is principally determined by themechanical accuracy achieved in its construc-tion. It is, therefore, possible to measure theactual surface of a reflector and, from thatinformation, to compute the electrical per-formance of the antenna. Moreover, such atechnique can be used as a diagnostic tool forthe adjustment of the reflector surface to withinacceptable tolerance bounds. In general a mea-surement accuracy of at least one twentieth ofthe shortest wavelength of operation is desirable.

For some applications analytical photogram-metr ic t r iangula t ion [48] can be used tomeasure the surface of reflector antennas.This method utilizes two or more long-focal-length cameras that take overlapping photo-graphs of the surface to be measured. Thesurface is uniformally covered with self-adhesive photographic targets the imagesappearing on the photographic record as shownin Fig 33. A least-square triangulation .processis used in which two-dimensional measure-ments of the images of the targets are processedsimultaneously to generate a unique set ofthree-dimensional coordinates for each discretetarget. Accuracies of the order of one part in20 000 to 100 000 of the diameter of thereflector, depending upon the extent of themeasurement procedure, are attainable.

For large reflector antennas operating atmillimeter wavelengths the photogrammetrictechnique may not be sufficiently accurate toadequately predict the antenna’s performancecharacteristics. Another method that can beused is that of precise range measurements.As an example of the accuracy demanded ofsuch range measurements, a 65 meter antennadesigned to operate at a wavelength of 3.5 mm

63

IEEEStd 149-9179

E X P O S U R E S T A T I O N

(Xi,

P H O T O G R A P H I C

A D O P T E D

C O O R D I N A T E S Y S T E M

ANTENNA-RANGE 11

Fig 33Schematic Illustration of Analytical Pbotogrammetric Triangulation

1491 requires over 3000 points on the surfaceto be set to an accuracy of + 0.1 mm. Thisaccuracy may be achieved by range measure-ments from two fixed points, such as the focusand vertex of the parabolic reflector. Theserange measurements, over distances from a fewmeters up to about 60 meters, shall be maderapidly, preferably using an automated system.A modulated laser beam can be used for thispurpose [ 501. The surface of the reflector iscovered with targets (small optical corner cubes)in a manner similar to that shown in Fig 33.The laser beam is directed to the targets bymeans of programmable mirrors. The entiremeasurement procedure can be controlled by asmall digital computer. The phase of the re-turned signal is measured with respect to areference. The phase shift is proportional tothe total distance traversed. If the distance andmodulation frequency are such that the phaseis shifted more than one cycle, an ambiguitywill occur. This can be resolved by a crudeknowledge of the distance or by using a dual-frequency system. An accuracy of f 0.08 mmat ranges up to 60 meters has been achievedusing this technique [ 501. This technique hasbeen successfully used on very large reflectors[511.

For high-precision reflector surfaces it ispossible to measure very accurately the curva-ture of the surface at a point by use of a devicethat contacts the surface at three points andhas a precise depth transducer at its center(located over the point to be measured). Themethod can be illustrated by reference toFig 34. The curvature K of the curve S is givenby

depKc-

dS

where 19p is the angle that the tangent at pointP makes with the x axis, and S is a length ofarc. The angle 19p can be obtained by integra-tion, that is,

OP = J,+ KdS

Since sin 19p is the derivative of the y coordi-nate with respect to the arc S, then

Yn = J,,” sin 8, dS

64

OPERATIONIEEE

Std 149-1979

Fig 34Geometry Used to Relate the Coordinates of a Point on the Surface of a

Reflector to the Measured Curvature

Thus performing the two integrations onecan obtain the coordinates of each pointalong the curve.

By making the three contact points wheels,the device can be rolled along various radii ofthe reflector, continuously sampling the curva-ture. An accuracy of 0.05 mm has been achievedby use of this method [52] .

After the antenna has been set as accuratelyas possible, it may deform as it moves intoposition or as the ambient temperature changes.A knowledge of these deformations permitspointing corrections to be made and may alsosuggest ways of improving the performanceof the antenna. Various instruments havebeen constructed for measuring these deforma-tions [ 53]-[ 551.

8. Antenna-Range Operation

It is highly desirable that a set of standardoperating procedures be developed for an an-tenna range. These procedures should be docu-mented in an operations manual. The contentsof the manual will be determined to a largedegree by the mission of the range. Certainlya range developed for a single function requiresmuch less documentation than does a multi-function range. The manual should be com-plete enough so that new personnel can operatethe range with a minimum of assistance.

The manual might include but not be limitedto the following

(1) General information. General informa-tion concerning the methods for antenna

testing, description of the range, coordinatesystems, and lists of references includingappropriate customer specifications and profes-sional standards should be contained in thissection.

(2) Range-ecalua &on techniques. Standardprocedures for the evaluation of the antennarange should be developed and documented(see Section 6). Included in the documentationmight be a general description of the recom-mended methods, the instrumentation required,detailed procedures, sample data, and methodsof data interpretation. Sometimes this materialis contained in a separate document and issimply referenced in the operations manual forthe range. An important part of the evaluationof the range, which should be included in themanual, is the alignment of the axes of thepositioner and the boresighting of the sourceantenna as discussed in 6.2-6.4.

If the data indicate excessive reflections, thensuggestions for the corrective action to betaken should be given. For elevation ranges theadjustment of the position and height of dif-fraction fences might be required, or if a singlereflecting object is located, then it should beremoved, or steps taken to redirect the re-flected energy away from the test antenna.

(3) Standard measurement setups. For eachtype of measurement that is repeatedly madeon the range a standard measurement setupshould be specified. For example, in the caseof antenna-pattern measurements, the docu-mentation should include a procedure for theselection of the positioner, transmitter, sourceantenna, receiver, recorder, and data-processing

65

IEEEStd 149-1979

equipment. In addition it should provide ameans of determining the required spacingbetween source and test antennas and theirheights. If scale models are used, there shouldbe a procedure for the selection of the scalefactor and construction technique (see 7.1).When vehicles, buildings, or other structureson which the antenna is mounted are to bemodeled, it is necessary to decide the extentof the modeling. For example, if a VHF an-tenna is located near the cockpit of an air-craft, the interior of the cockpit may have tobe modeled quite accurately; whereas, if theantenna were mounted on the tail section ofthe aircraft, the cockpit area probably wouldnot have to be modeled so accurately.

If other antennas are located on the structure,they must also be included in the model andappropriately electrically terminated, a factthat should be included in the manual, as wellas the method of mounting the model on thepositioner. The method of feeding the modelantenna should also be discussed.

(4) Standard operating procedures. A step-by-step procedure for equipment operationshould be provided including warm-up timesand any calibration of components that isrequired. The following considerations shouldbe discussed. The stability of the equipmentought to be checked prior to making anymeasurements. Repeatability of measurementsis most important. For example, if it is relativeamplitude measurements that are to be madeand a polar recorder is employed, then a e-cutand an @-cut should be recorded and for eachcase allowed to retrace four or five times. Eachretrace should fall exactly on top of the pre-vious patterns. If they do not, corrective actionshould be taken.

If the antenna under test is known to have asymmetrical pattern, then its symmetry shouldbe tested at the outset. This can be accom-plished by recording a pattern about the planeof symmetry. Again corrective action can betaken if the resulting pattern so indicates.

A cursory check of the pattern range usingthe antenna-pattern-comparison method (see6.5) should be included in the procedure. Forthe case of a production range which has highusage, a set of two patterns may be all that-is

66

ON SITE MEASUREMENTS OF

required. For example, if a model tower isused, a 90” conical pattern recorded on theright side of the tower should be identical toone taken on the left side.

The operating procedure should indicatewhat constitutes a complete set of data. Forexample, when conical cuts are made forantenna-pattern measurements, what will bethe maximum number of increments of 8 to betaken? Perhaps both a complete set and apartial set will be defined.

There shouid be a continuous check of theequipment and range performance during therecording of a complete set of data. It is goodpractice to monitor the transmitted signalduring the measurement of a pattern set.

It is also good practice for a set of crossoverpatterns to accompany each set of conicalpatterns. These are obtained for a set ofconical patterns by recording an extra set of8 cuts with Q = 0” and 4 = 90” as referencepatterns. For each conical pattern (Q cuts)there will be two crossover points on each ofthe reference patterns. As each conical cut isrecorded, the appropriate crossover points aremarked with a dot or a small X on the refer-ence patterns. These points should fall on thereference patterns. If they deviate from thereference pattern by a specified amount, cor-rective action should be taken.

If the range is equipped with an automaticintegrator, positioner programmer, or anydata-processing equipment, then the appropriatecalibration procedures should be detailed aswell as the operating instructions.

A method of identification of patterns isessential to the operation of an antenna rangeand should be described in the operationmanual. The conditions of the measurementshould be written on each pattern. Usually foreach set of patterns a cover sheet is preparedwhich contains the general information and asketch of the antenna under test along with acode that designates the particular patternset. For this case only the polarization, codenumber, and cut designation will be requiredon the individual patterns. Sometimes it isdesirable to make a rubber stamp to indicatewhat information is required. and perhaps toindicate the orientation of the test antenna.

AMPLITUDE PATTERNS

Any special conditions should be noted eitheron the patterns or on the cover sheet. In anautomated system this can be accomplished inthe software by requiring the operator to inputthe required information.

9. On-Site Measurements ofAmplitude Patterns

The measurement of the complete two-dimensional radiation pattern... of a full-scaleantenna located on its ultimate site may belaborious and an expensive task. It is necessaryto make such measurements when the antennaradiation is significantly affected by the siteon which the antenna is located or when itsconstruction and,‘or assembly are practical onlyif carried out at the site. Furthermore, on-sitemeasurements of an antenna may have theadditional usefulness as a conclusive demonstra-tion that the performance of the antenna is asdesired and that it interacts with its environ-ment in a predicted manner.

A wide variety of techniques have beenemployed for the on-site measurements of theamplitude patterns of an antenna. The pro-cedures outlined in the following discussionare commonly used and illustrate some of theproblems involved. It should be emphasized,however, that each on-site antenna measure-ment has special considerations and techniquesthat are not necessarily discussed here.

Fig 35 indicates the essential parts of thetypical measurement system. A distant sourceis carried by a vehicle, which is maneuveredthrough the space surrounding the test antennato produce a source of an essentially planewave incident on the antenna from all direc-tions of interest. The direction to the source,with respect to a reference direction at theantenna, is obtained from a tracking device.This information provides the angular data toa recording device. The amplitude of the signalreceived by the antenna provides the amplitudedata to the recording device. These data maythen be processed to present the antennapatterns in the desired form.

Airborne vehicles, such as conventional air-planes. helicopters. blimps, and free and captive

IEEEStd 149-1979

balloons, have been employed to carry thesource. The source should be in the far-fieldregion of the antenna system being measured.If this is not possible to achieve with airbornevehicles, man-made earth-orbiting satellites[ 561, the sun [ 571, and radio stars [ 581, havebeen used.

If the attitude of the source antenna relativeto the antenna under test changes, a change inthe received signal is likely to occur. To mini-mize this possibility, the source antenna shouldbe oriented so that the peak of its beam is inthe direction of the antenna being measured,and the useful portion of the source patternshould be as uniform as possible. In addition,when the source antenna cannot change thedirection of its beam, the course flown by theaircraft should be chosen to minimize changesin attitude. A favorable course for this purposelies along a circle centered on the antennabeing measured and contained in a planeperpendicular to a vertical axis through theantenna. It is equally important that the designand placement of the source antenna includethe effects of its environment. These environ-mental effects are greatly dependent on therelative size of the aircraft and the operatingwavelength. The source antennas may beseparated into two classes according to thisdistinction.

In the lower frequency class (HF and VHF)the selection of a source antenna may dependon polarization. For horizontal polarization asleeve-dipole antenna trailed behind the aircrafthas proved satisfactory. This antenna is fabri-cated from a length of standard coaxial cable byremoval of the shielding braid for a distance ofa quarter wavelength, and it can be supportedin a horizontal position by a miniature para-chute. For vertical polarization useful resultshave been obtained with a monopole wherethe aircraft serves as the “ground.” Since itsradiation pattern has a null in the verticaldirection, measurements made with the sourceantenna near the antenna’s zenith should takethe inherent pattern variations into account.For techniques where the polarization has tobe changed during the course of the measure-ment, rotatable ferrite-loaded dipoles, whichcan be mounted on the side of the aircraft,

67

IEEEStd 149-1979 ON SITE MEASUREMENTS OF

POSSIBLEREFERENCEANTENNA

ANTENNAUNDER

-TEST

SOURCEANTENNA

TRACKlNG

TO RECORDER TO RECORDER

Fig 35System for On-Site Measurements of Amplitude Patterns

have been found useful [59] . Careful calibra-tion is needed due to the asymmetry of theaircraft structure with respect to the antenna.

In the higher frequency class (microwaveregion) wing-tip antennas have been successfullyutilized. This placement minimizes excitationof the airplane structure and permits a sourceradiation pattern that is reasonably close to thepattern of the isolated source antenna.

As indicated in Fig 35, a tracking device isrequired to establish the relative direction tothe source. In particular, the direction of theaircraft with respect to the tracker is mea-sured. The direction of the aircraft relativeto the antenna under test shall then be deter-mined by taking into account the parallaxerror introduced because of the known separa-tion between the antenna under test and thetracker. In addition to the source direction, itmay be necessary for the tracking system todetermine the range to the source. This infor-mation may be needed either to compute theparallax correction, or to correct for the changein the incident power flux density causedwhen the aircraft does not fly in a perfectcircle about the antenna under measurement.

Two types of tracking instruments are incommon use, optical and radar trackers. Animportant distinction between the two is thatthe ordinary optical tracker furnishes onlythe direction of the source, while radar

furnishes direction and range. A laser trackerwould furnish range also. When an opticalsystem is employed, the range of the sourcecan be calculated from the aircraft attitudemeasured by an instrument in the aircraftand transmitted to the test site.

The system described in the preceding dis-cussion may be adequate for measuring the on-site amplitude patterns of an antenna. However,there are occasions when this system is notsufficiently accurate, because the amplitude orpolarization of the wave radiated by the sourcetoward the antenna under test is insufficientlystable. This variability is particularly likely atmicrowave frequencies. In these cases a refer-ence antenna (Fig 35) may be placed close tothe antenna under test to measure the apparentvariation in the strength of the source. Thesedata can be used to normalize the signalreceived by the antenna under test and henceremove the apparent variations in the sourcesignal. In addition to the shape of the amplitudepattern, the reference antenna also may beused in a measurement of power gain, as dis-cussed in 12.3. Finally the reference antennacan sometimes be designed to determine thepolarization of a set of source antennas, asdiscussed in 11.2. This feature, in turn, maypermit the polarization of the antenna undertest to be determined. It should be recognized,however, that unless the reference antenna

68

AMPLITUDE PATTERNS

system is carefully designed, it may introduceerrors as great as those it is intended to cancel.For example, its received signal may varybecause of site reflections in addition to sourcevariations.

At microwave frequencies it is desirable tomake the response of the reference antennasubstantially independent of the ground andsurrounding structures. Such independencerequires the use of a reference antenna havinga narrow beam and low minor lobes and mayalso require treatment of the ground and otherstructures near the reference antenna, whichmay be satisfied by slaving the reference an-tenna to the tracking device. At frequencieswell below the microwave range, the pattern ofthe reference antenna may be so broad thatground reflections produce significant varia-tions in the received reference signal. If thepattern of the reference antenna system can beaccurately determined, then it is still possibleto employ its output as a means for correctingvariations in the wave radiated by the sourceduring measurement of the patterns of theantenna under test. However, at the lowerfrequencies such as HF it is customary to usethe reference antenna only during the measure-ment of power gain at one angle, as describedin 12.3.3. Regardless of the frequency beingemployed, the reference antenna system shouldbe designed and located so that it has a negli-gible effect on the patterns of the antennaunder test.

The process of pattern measurement andrecording may involve either a point-by-pointor a continuous method; the latter is preferable.Commercially available continuously recordingequipment can often be adapted to introduceautomatically the various corrections that areneeded in both the signal and the angle inputs.The data may also be stored on magnetic tapeand processed by computer. The antenna pat-terns can be presented in different forms asdiscussed in Section 5. The most comprehen-sive form is that of the contour plot.

10. Phase

10.1 General. The radiation pattern of anantenna is completely described by the magni-

I E E EStd 149-1979

tude and phase of the radiated field componentsin two particular orthogonal polarizations.Most commonly only the magnitude of aspecified component of the field is measured,namely, that for which the antenna is designed.However, the phase of this component mayalso be of importance. Furthermore, for acomplete description of the field, the magni-tude and phase of the cross-polarized com-ponent shall also be measured. The variationof the phase of the far field provides necessaryinformation for focused reflectors and beams,tracking antennas, and phase interferometers.The variation of the phase and magnitude ofthe near field may be necessary to permit anaccurate prediction of the far field patterns(see 7.3) or in other cases to interpret thecharacteristics of an antenna.

A single-frequency field component of radianfrequency w can be represented as a scalarfunction of time by

E(t) = E, cos (wt+ $) = Re (E,-,ei$ eiwt)

where + is a real number and EO is a positivereal number.

NOTE: There are two generally accepted conventionsin the writing of the complex time factor, ejwt andeWiwt. For a positively traveling wave the phase factorwill be e-jhx when ejwt is used. The factor ejwt isused in this standard except in 7.3. To change anyformula replace -i by +j.

The phase at time t of e (t) is the angle of thecomplex number E,e i($ +wt), or

p h a s e 6 (t) = IJJ +wt

If the phase of e is used without specifyingthe time, it is implied that t = 0, and

phase e”=$

If the field /? propagates along the x axiswith a velocity IJ = w/h, where k is the wavenumber, then

e (t, x) = E, cos (at - kx + $)

At point X, as a function of t, & is representedby the curve e (t, 0) shifted in the positive di-

69

IEEEStd 149-1979

reaction by x/u (see Fig 36). This is consideredas a phase delay, lag, or retardation. The phaseis decreased by amount hx in the propagationfrom 0 to x.

In contrast to the scalar case, when dealingwith a vector field it is important to realizethat, while there is a “natural” or canonicalway of defining the phase of a linearly polarizedfield vector, this is not true for ellipticallyor circularly polarized field vectors. Forexample, in the linearly polarized case it isnatural to express the field & (t) as the productof a scalar function g (t) and a unit realvector u which indicates the direction ofpolarization. The phase of E (t) is the phaseangle ofE,& . If the polarization is not linear,this is no longer possible. A complex vector u^ ofmagnitude 1 (f? * *u^ = 1) can be used, but it isnot uniquely defined by the polarization alone:the vector u^ can be replaced by u^ ’ =h da,which represents the same polarization, but thephase relative to c ’ differs from that relative tofi by 0. This concept will be expanded in thenext section.

10.2 Phase Patterns. A specified componentof the far field produced by an antenna can beexpressed in the form

where E and $ represent the (8, 4) dependenceof the magnitude and phase, respectively, ofthe specified component, and (r, 19, 0) arespherical coordinates of the observation point.The vector u^ is a vector of magnitude unity(unitary vector), which indicates the polariza-tion of the specified component. For a linearlypolarized component the vector u^ is taken asreal; for instance, it may be the unit vector u,(or u6) in the direction of increasing 8 (or 0).Then the phase J/ is defined unambiguously.However, for an elliptically or circularly polar-ized component the vector u^ is complex, aspreviously stated, and is not uniquely definedby its polarization and the normalizing condi-tion. Multiplication by e j”! preserves these con-ditions. This changes $ accordingly. Thereforea definite convention shall be used to specify1; and, when discussing a pattern, this conven-tion shall be specified for every direction (8.0)of interest.

10.3 Antenna Phase Center. For many appli-cations it is desirable to assign to the antennaa specific reference point, the phase center,from which radiation may be said to emanate.

Fig 36Phase Shift of Single-Frequency Field Propagating in the x Direction. The Phase of & (t,~)

Lags the Phase of @ (t, 0, ) by kx

70

PATTERNSIEEE

Std 119-1979

Knowledge of the location of this referencepoint is highly desirable and sometimes aprerequisite to the successful design of an-tennas such as phased arrays and primary feedsfor reflectors. In addition, the design of track-ing and homing systems on aerospace vehiclesand of rendezvous radar systems may requirethe precise determination of the phase char-acteristics and phase reference point of theelements of these systems.

If there exists an origin of coordinates suchthat, for any given frequency, the function $in the preceding equation is independent of19 and 4, this origin will be the center ofspherical wave fronts or equiphase surfaces.It is the reference point or the phase center ofthat component of the radiated field, and thepattern is described for the specified polariza-tion by the single real function P(B, $I). Thisoccurs for linear antennas and arrays that haveodd or even crossing symmetry, that is, suchthat the current satisfies the condition [60]

1(-x) = *r*(X)

For most antennas a “true” phase center,valid for all directions does not exist. However,one may sometimes find a reference point suchthat + becomes constant over a range of direc-tions of interest, for instance, over a portion ofthe main beam of the antenna. This point hasalso been called a phase center, or an apparentphase center [60]. It is particularly importantto determine such a point for the primary feedof a reflector since it allows one to use geo-metrical optics to properly place the feed withrespect to the reflector and to compute theradiation pattern.

Some antennas, such as the conical log-spiral,have a natural axis of symmetry. It is con-venient to take this axis as the polar axis of thesystem of spherical coordinates. The phasepattern may then be separated into the product

$Jte, 4) = o(e) (P (4)

For a particular choice of the origin 0 on theaxis it may happen that, at least over somerange of interest, the phase $ is independentof 8 and reduces to ~11 (4). This point 0 can be

considered as a phase center for the 0 coordi-nate. It can be found by the same methods asthe ordinary phase center by observing a cutof the phase pattern for a constant value of+ (see 10.4).

The theoretical calculation of the apparentphase center is usually laborious and limited tothose antennas for which the complete expres-sion for the far field is known [61]-[64]. Indoing so, the most straightforward approachis to calculate the radius of curvature of thisequiphase surface at the point of interest. Therays from the phase reference point to theobservation point form a pencil of lines as oneof the spherical coordinates, say 0, varies. Theevolute of the far-field equiphase contour istraced by the envelope curve of the rays. Theevolute is the locus of the phase referencepoints for varying 6. As 19 and Q are bothallowed to vary, the evolute of the equiphasesurface generates a warped surface upon whichthe phase reference points or apparent phasecenters lie. There are in general two principalphase centers on each normal.

For many applications it becomes necessaryto determine or check, by exper imentalmethods, the location of the phase centers ofthe antennas under study or construction. Theexperimental determination of a phase centercan be done by finding an equiphase surfacein the far field. If this surface is a sphere, itscenter is the phase center. If it is not, one cansometimes approximate the equiphase surfaceby a sphere over a limited range of directionsand thus obtain the corresponding apparentphase center. There are methods that dependonly upon amplitude measurements of the farfield [ 631. However, the use of amplitudepatterns to derive phase information is severelylimited by the realizable recording accuracyand not normally suited for the determinationof small phase changes.

10.4 Phase Measurements10.4.1 General. The radiation pattern of an

antenna is completely described by the magni-tude and phase of the radiated field componentsin two particular orthogonal polarizations. Inthis section emphasis is given to the aspect ofthe measurement of the phase patterns. It is

71

I E E EStd 149-1979

ANTENNA

1 REFERENCE/

/PHASE

FLEXIBLEI-

MEASUREMENT, /CAELE

CIRCUITTEST

(0)

PHASE

FIXEDANTENNA)a

M O U N T u

DISTANTSOURCE

s-0%

ANTENNAUNDER TEST

Fig 37Arrangements for Measuring Phase Patterns

(a) Near-Field, or Short-Distance, Pattern Phase.(b) Far-Field Pattern Phase

understood that in an actual measurement bothphase and ampli tude would be recordedsimultaneously.

Since phase is a relative quantity, a referencesignal shall be provided at all times for com-parison.

NOTE: Because of the periodic nature of angles, phaseis only defined up to multiples of 2n rad.

For measurements made at short distances theantenna under test may be used as the trans-mitting antenna and a simple receiving antennaor probe may be used to sample the radiatedfield [65], as shown in Fig 37(a). A referencesignal may be coupled out from the transmis-sion line leading to the test antenna and com-pared with the received signal in a suitablecircuit. For measurements at distances toogreat to permit direct comparison betweenthe sampled signal and this type of referencesignal, the arrangement of Fig 37(b) may be

used. The signal from a distant source isreceived simultaneously by the antenna undertest and by a fixed reference antenna. Tomeasure a phase pattern the antenna under testis rotated in the usual manner as for measuringradiation patterns. The arrangement of Fig 38can be used for the measurement of the relativephase between two ports of amultiportantennaas a function of the angle around the antenna.

The ‘most comprehensive information aboutthe phase characteristics of an antenna can begained by recording the complete phase pat-terns, preferably with an automatic system.From these phase patterns contours of con-stant phase and the phase center or referencepoint can be determined (see Section 5).

A consideration of the typical characteristicsof the phase pattern of antennas with a well-defined phase center is useful in the interpreta-tion of patterns from an antenna with unknowncharacteristics [60] . Let us assume that a

72

MEASUREMENTSIEEE

Std 149-1979

ROTARY

DISTANTSOURCE

b-Q

ANTENNAUNDER TEST

Fig 38Phase Measurement Between Two Ports of a Multiport Antenna

given antenna has a phase center on the axisof the structure and that the antenna is posi-tioned in a r,0,$ coordinate system with8 = 0” as the antenna axis. If a phase pattern isrecorded for this antenna about some origin,or center of rotation, which coincides withthe phase center, this pattern over the mainbeam will, by definition, be a constant. If theamplitude pattern has side lobes, there will bea 180’ phase reversal at each null, and thephase pattern will exhibit an abrupt transition.If a phase pattern is recorded as this antennais rotated about some origin along the axis ofthe structure, but displaced a distance r’ fromthe phase center, the phase of the field of theantenna will be modified by a cosinusoidalfunction of 8. When t-’ is very small comparedto the distance to the point of observation, andthe pattern is normalized to the phase of thesignal received when the phase center is atposition A in Fig 39(a), the resultant phasepattern for an antenna with only one radiatedbeam will be given by

* = kr’(l- cos 6)

Normalization to the phase of the signal receivedwhen the phase center is at position B in Fig39(b) will produce the image of the firstpattern.

In theory the position of the phase centercan be calculated from one such phase pattern.If a change in phase @ is measured as the an-tenna is rotated from 0 = 0” to 8 = 0,, thephase is displaced from the center of rotation

by the distance

rf=& (i+)

In practice it may be necessary, because ofboth pattern and experimental anomalies, torecord several patterns as the antenna is re-positioned along its axis to bracket the curvefor r’ = 0.

If the apparent phase center is displaced fromthe axis of the antenna by d, as indicated inFig 39(c) and (d), the phase change withrotation will be

\zIX kr” [l-cos(8+arctan $)]

w h e r e (r”)’ = (r’)* + d*. If the minimum de-tectable phase change, that is, the resolution ofthe phase-detection system, is 0.5” and a phasecomparison is made over a range of 0 = + 10”from the axis of the antenna, the indeter-minacy in the position of the measured phasecenter may be as great as 0.1 wavelength. Inpractice this uncertainty can be minimizedby recording over a greater range of 8 if thereis an apparent phase center over this greaterrange, and/or by recording a second value for

Ir min when the phase center approaches thecenter of rotation from the opposite direction.In the ideal case the geometry is such that thetwo values of rfmin should indicate a bracket ofthe apparent phase center.

73

IEEEStd 149-1979 PHASE

CENTER OF

IROTATION/

1,r’= 0

I

2

r f 8 )IxI r”

’ I

R -iy.L--d lrkf)

r'= 0

/- i-- -kd

i

0 4

Fig 39Geometry and Phase Change as a Displaced Source Is Rotated about a Given Origin

To make these measurements, the antennaunder test is placed in the far field of a sourceof the desired polarization, on a positioner thatis capable of precise rotation about a pointand of precise translation along the axis of theantenna. To correct for and evaluate a trans-verse displacement of the phase center fromthe axis, and to correct for minor mechanicaltolerances in the positioner, it is highly desirable

to also be able to precisely translate the antennain a path that is orthogonal to its axis.

10.4.2 Instrumentation. When consideredfrom the viewpoint of measurement systemerror and instrumentation complexity, fewantenna-system measurements have been moredifficult or time consuming than those involvingthe measurement of phase of the radiated field.

74

‘cI

MEASUREMENTS

However, recent advances in phase-measurementinstrumentation, for example, the introductionof radio-frequency vector voltmeters, computer-controlled and manual radio-frequency networkanalyzers, and phase and amplitude receivers,have greatly reduced the magnitude of themeasurement problem. If it is necessary, orfor some reason desirable, that a phase-measurement system be assembled, there are,in the literature, papers concerned with theanalysis and classification of most of the mea-surement systems that have been proposedand used [19], [66] (see Section 5).

10.4.3 Sources of Error. Whether a com-mercial phase-measuring instrument or anassembled system is used, considerable atten-tion shall be paid to the numerous sources oferror. A major source of error in the measure-ment of phase is due to the interaction ofreflected waves from components which arenot matched to the waveguide or transmissionline in use [67] . In addition to possible phaseerrors introduced by nonideal components,multiple reflections between the componentsalter the relative magnitude and phase of thetraveling waves on these lines. The relativephase and magnitude of these wavefronts maycoincide with those that would be present inthe matched case, but more likely will lie some-where between extreme limits set by the mis-match. For two cascaded discontinuities, withreflection coefficients small with respect tounity, the maximum possible phase mismatcherror is approximately

arcsin Ir, / 1 r2 /

where l’i and I’* are the reflection coefficientsof the two discontinuities when viewed from acommon point between them. For example,the relative phase between the fields at any twopoints on a transmission line connecting twoconnected instruments or components withreflection coefficients such that they wouldcause voltage-standing-wave ratios of only1.3:1 and 1.5:1, respectively, could vary by asmuch as + 1.49” from that on a reflectionlessline.

The only way to reduce this possible error isto reduce the magnitude of the mismatch of

IEEEStd 149-1979

these sources. This can be done with well-matched attenuators or pads. However, such animprovement is limited by the cascaded dis-continuities due to the irreducible mismatchof the transmission-line connectors. Lappedwaveguide flanges and precision connectorsshould be used where possible. It is particularlyconvenient to use signal-flow-graph techniquesfor error analyses of these measurement sys-tems [ 681, [ 691.

Measurement systems which involve thepropagation of energy from the source to thephase-measurement receiver through two pathsimpose strict requirements on the frequencystability of the source. If these paths are notexactly equal in electrical length, a shift in thefrequency of operation from fi to fz willcause a shift in the measured phase differenceA\k between the signals on two paths ofapproximately

A@ = 360Q, - Q2 Q, -Q2

x1 - x*

degrees, where Q, and Q2 are the lengths of thetwo paths, and X, and X2 are the guide wave-lengths at f, and f2. If these paths involvecombinations of waveguides or transmissionlines and free space, the electrical lengths ofthe paths shall be calculated accordingly.

Xs an example, for measurements made at1000 MHz, with a difference in a free-spacepath length of 1 meter between two channels,a change in frequency of 1 MHz (that is, 0.1percent) will cause a phase change between thetwo channels of 1.2”. If this change in frequencyoccurs during a measurement, the change inphase appears as an error in the measurement.To minimize such errors, the two path lengthsshould be kept equal in electrical length byadding an additional transmission line to theshorter channel. This latter requirement is,of course, a necessity for swept-frequencymeasurements.

In many phase measurements, such as in themeasurement of fields around an antenna, itbecomes necessary to move a receiving probein space. This movement involves a bending orflexing of the coaxial cable leading to the

75

IEEEStd 149-1979

probe, or where waveguides are used the rota-tion of several rotary joints may be required.Most rotary joints change, to some degree, thephase of the output’ signal with rotation. Atmicrowave frequencies it is not difficult tochange the electrical length of a short sectionof coaxial cable by a degree or more when it isbent or flexed. A change in temperature willcause similar errors. This change in electricallength will appear as an error in the measure-ment of the change in phase of the fields as afunction of space. This phase shift is difficultto eliminate although there are cables which aredesigned to minimize it. Alternately it cansometimes be minimized by first selecting asection of cable that exhibits little changewith the necessary flexing. A difference cansometimes also be noted between differentpieces of the same cable. The cable should belong enough so that sharp bends are not neces-sary and the movement of any one section ofthe cable is held to a minimum. X dissipativeor lossy cable sometimes may show less phaseshift with flexing than the low-loss cables. Inwaveguide systems there is little that can bedone to eliminate this error except to calibratethe phase shift of the movable arms and rotaryjoints as a function of position in space if sucha calibration is required.

These examples should indicate to the experi-menter that accurate measurements requirecareful attention to the details of the system,and assumed accuracies of say 5’ or less shallbe viewed with caution unless all factors havebeen taken into account. One factor that issometimes overlooked is the signal-to-noiseratio in the system. A high signal-to-noise ratiois necessary to prevent the noise from con-tributing to errors in the phase measurements.

11. Polarization

11.1 General. Polarization is a property ofsingle-frequency electromagnetic radiation de-scribing the shape and orientation of the locusof the extremity of the field vectors as a func-tion of time [ 1, 3.1-3.41, [70].

NOTE: Some of the methods of analysis used forsingle-frequency fields can be extended to partiallypolarized fields [71], [72]. Random fields or randomantennas will not be considered here.

In common practice, when only plane waves orlocally plane waves are considered, it is suf-ficient to specify the polarization of theelectric field vector E. In a plane wave with aknown direction of propagation the magneticfield vector H is simply related to the E field.It can be deduced by a 90” rotation about thepropagation vector, followed by multiplicationby the intrinsic admittance Ye of the medium,

yo= 5$_IJwhere E is the permittivity and 111 is the perme-ability of the medium. In vacuum Y0 z l/377 =2.66 X 1O-3 52-l.

The far field radiated by an antenna isgenerally observed in a small region where itcan be considered as a plane wave propagatingaway from the antenna in the radial direction.The electric field is in a plane perpendicularto that direction. The locus of its extremity,is, in general, an ellipse that may degenerateinto a segment of a straight line or into a circle.Correspondingly, the polarization is calledelliptical, linear, or circular.

The sense of rotation of the extremity of Edescribing a circle or an ellipse in the plane ofpolarization (perpendicular to the direction ofpropagation) is called the sense of polarization,or handedness. This sense is called right handed(left handed) if the direction of rotation isclockwise (counterclockwise) for an observerlooking in the direction of propagation (seeFig 40). Alternative conventions are discouragedas they could be a source of confusion. If,instead of considering the field vector at apoint as a function of time, one considers thevector at a given instant of time, as a functionof distance along the direction of propagation,one might be lead to incorrect designations.

Elliptical polarization is characterized by theaxial ratio of the polarization ellipse, the senseof rotation, and the spatial orientation of theellipse with respect to a reference direction in

76

MEASUREMENTS

IEEEStd 149-1979

i e-R IG H T H A N D

C L O C K W I S E

‘ION

Fig 40Illustration of the Sense of Rotation

I -

.e-900

0: 900

Fig 41Polarization Ellipse in Relation to Antenna Coordinate System

IEEEStd 149-1979 POLARIZATION

A N T E N N A

E+ E,TransmItted Matched lncldent

E,lncldent

Fig 42Relation Between Polarization Properties of an Antenna when Transmitting and Receiving

Et - Far-Field Electric Vector of Antenna; E, - Electric vector of IncidentWave which Is Polarization Matched to Antenna; Ei - Electric Vector of

Arbitrarily Polarized Incident Wave

the plane containing the ellipse. The anglebetween the reference direct ion and themajor axis of the ellipse is called the tilt angle.

For a plane wave, the tilt angle is measuredc l o c k w i s e f r o m t h e r e f e r e n c e d i r e c t i o n w h e n

electric field vector Et i n t h e f a r f i e l d , r a d i a t e dby the antenna in that direction. Since an an-tenna usually has a spherical coordinate systemassociated with it (see 3.1), its polarization in adirection (0, @) can be illustrated with respectto the coordinate system, as shown in Fig 41.Although the reference direction for establish-ing the orientation of the polarization ellipseis arbitrary, it is common practice to take theu, axis as the reference direction [ 21.antenna-pattern-measurement situations it is

[73,475-4771. T h e h o r i z o n t a l a x i s i s u s u a l l ychosen as the reference direction. This conven-tion shall be employed throughout this standard.

When an antenna receives a plane wave com-ing from a given direction, the response (open-circuit voltage, short-circuit current, or availablepower) is maximum, for a given intensity ofthe incident plane wave, when the polarization

E, h a s t h esame axial ratio, the same sense of polarization,and the same spatial orientation of the majoraxis as that of the antenna for that direction(see Fig 42). Because the sense of polarization isrelative to the direction of propagation, and

a n dEm, the senses appear to be different whenlooked at from a single point of view. Also, in

78

MEASUREMENTS

order to be consistent with the definition ofantenna polarization, the local coordinatesystem associated with each of the waves isoriented so that one of the coordinates is inthe direction of propagation. As a result thetilt angles for the two polarization ellipses,which are measured according to a right-handrule, are different. As shown in Fig 4.2 if rt isthe tilt angle for the ellipse described by Et,then the one described by Em will be

?rn = 180° - Tt

It should be noted that 7, may also be ex-pressed as -TV plus any integral multiple of180” since all these angles correspond to thesame orientation of the major axis of thepolarization ellipse. The polarization of theincident wave, which yields the maximumresponse at the antenna terminals, as describedabove, is called the receiving polarization of theantenna.

If the incident plane wave has a polarizationthat is different from the receiving polarizationof the antenna, then a polarization loss occursdue to this mismatch. The polarization effi-ciency p is used to account for polarizationmismatch [ 70, pp 544-5491, [ 741.NOTE: This factor is also called polarization mismatchfactor [75], and polarization receiving factor (ANSI/IEEE Std 100-1977, Dict ionary of Electr ical andElectronics Terms).

It is defined as the ratio of the power actuallyreceived by the antenna divided by the powerthat would be received if a wave from the samedirection, with the same intensity and polariza-tion matched, were incident on the antenna.

The Poincare sphere [70, pp 540-5441 isa useful graphical aid for the visualization ofpolarization effects since the polarizationefficiency is uniquely determined by theseparation of two points on the sphere whichdescribe the polarizations of the incident waveand the receiving antenna, respectively.

The construction of the Poincare sphere is aconsequence of the fact that any wave can beresolved into two orthogonal components(they may be two orthogonal linear, elliptical,or circularly polarized components). The total

IEEEStd 149-1979

power in the wave can then be represented asthe sum of the powers contained in theorthogonal components:

y,E: = Y,E2, + YoE;

where Et represents the effective value of themagnitude of the electric field vector of thegiven wave, and EA and E, are those of thetwo orthogonal components. By dividing theequation by the total power Y, E t one obtains

E; +‘; =1

where EA and 2; represent the fractions ofpower in their respective polarizations. Thislatter equation leads to the construction shownin Fig 43. It is evident in Fig 43 that the posi-tion of W along the arc AWB indicates thedivision of power in the wave W between thetwo orthogonal polarizations A and B.

For example, if W represents the incidentplane wave and A the antenna’s receivingpolarization, then the angular separation 2< be-tween A and W determines the division ofpower density in the wave between the receivingpolarization and the orthogonal polarization.The polarization efficiency p is given by

p = 2 ‘A = cosz {

Note that p = 1 when A and W are coincidentand 0 when A and W are diametrically opposite.For the former case the wave is said to be co-polarized with respect to the antenna’s receiv-ing polarization and for the latter case they arecross-polarized. Fig 44 depicts the relative spa-tial orientation of the polarization ellipses oftwo cross-polarized fields. Note that for ellip-tically polarized fields a rotation of 90’ aboutthe direction of propagation alone does notproduce a cross-polarized field. The rotationshall be accompanied by a change of the senseof polarization.

The Poincare sphere is generated by rotatingthe semicircle A WB about the A-B axis withthe angular measure around the equator ofthe resulting sphere being the relative phase

IEEEStd 149-1979

a ------sin 5

C‘P ’5

cos 5

----- - 00

POLARIZATION

a 25

Fig 43IlIustration of the Division of Power Between Two

Orthogonal Elliptical Polarizations A and B

Fig 44Cross-Polarized Field Vectors

80

WIEASUREMENTSIEEE

Std 149-1979

--

135O VERTICALLINEAR

HORIZONTALLINEAR

Fig 45Poincard Sphere Representation of the Polarization of a Plane Wave W

between the two orthogonal polarizations A The complex polarization ratios are given byand B (see Fig 45). In Fig 45 the angle 2{ ofFig 43, which was defined for arbitraryelliptical components A and B, is assigned the

p^L = pLei6L

symbols 2a, 20, and 27 as the polarizations zD= pDei60A and B become

(horizontal linear)

(vertical linear)

(45’ linear)

(135’ linear)

A

PC = PCe j6 c

where

pL = tan o = Zv/EH

pD = t a n fl = Eiss /z‘$s

PC = tan7 = ER/EL

and where 6,, 6,, and 6, are the relative

EL (left-hand circular (LHC)) phases between the corresponding orthogonal

2Y components. The relative phase AC of the

ER (right-hand circular (RHC)) circularly polarized components is defined bythe angle of the instantaneous electric vectorof the right-hand circular component with

There is a one-to-one correspondence between respect to the horizontal direction at theall possible polarizations and points on the instant that the electric vector of the left-handPoincare sphere (see Fig 46). circular component is in the horizontal direc-

81


LHCPOLES REPRESENTC I R C U L A R P O L A R I Z A T I O N S

L A T I T U D EREPRESENTSA X I A L R A T I O UPPER HEMISPHERE

LEFT -HAND SENSE

E Q U A T O RREPRESENTSL I N E A RP O L A R I Z A T I O N S

dLlNEAR

LOWER HEMISPHERELONGITUDE REPRESENTS

RIGHT -HAND SENSETILT A N G L E

RHC

Fig 46Representation of Polarization on the Poincarc? Sphere

VERTICAL

6CHORIZONTAL

Fig 47Definition of Phase Reference for Orthogonal Circular Components

tion, as shown in Fig 47. Hence the twocircularly polarized components are in phasewhen the electric field vectors in these twowaves are in the same horizontal direction atthe same time. The circular polarization ratiopc is of particular interest since the axial ratior of the polarization ellipse may be expressedas

PC + 1r =-

PC - 1

Note that the sign of the denominator givesthe sense of polarization with r positive for theright-hand sense and negative for the left-handsense. Also, the tilt angle r is given by

7 = 6,/2

If the points corresponding to the receivingpolarization of the antenna A, and the polari-zation of the incident wave W are located onthe Poincare sphere, then the polarization

82

MEASUREMENTSIEEE

Std 149-1979

HORIZONTALLINEAR

45O LINEAR

P’ RHC

Fig 48Polarizations of Incident Wave W and Receiving

Antenna A,, Plotted on the Poincark Sphere

efficiency can be determined by use of theexpression

p = cos2~

where 2< is the angular separation betweenA, and p can be obtained with the use of the polari-zation ratios. For example, it can be shownthat

1 + p&p; + 2pwp, cos AP=

(1 + P’,, (1 + Pi)

w h e r e pw and pr can be any of the threepolarization ratios pi, pD, or pc, with A beingthe corresponding difference in phase angles[(sL)W - (aL)r], [(sD)W - (sD)r]y Or[(SC )W - (SC ),] . When the circular polariza-tion ratio pc is chosen, the angle A is twicethe spatial angle between the tilt angles of thepolarization ellipses of the two polarizations.This equation for p can also be written as

p+p^wp^:/*p = (l+P&)(l+P:)

where i w and p^ r are the complex polarization

ratios. This form is interesting because it issimilar to the equation for mismatch loss inlossless transmission lines with the source andload reflection coefficients being analogousto PW and pr, respectively (see 12.5).

The polarization efficiency can also be writtenin terms of the axial ratios for the two polariza-tions by the use of the relationship betweenthe circular polarization ratio PC and the axialratio :

(1+rZ,)(1+rZ,)+4rw~r+(1-r~)(l-rt)cosn

P=2(l+rk)(l + r: 1

In this formula the axial ratios are taken aspositive for right-hand polarization and nega-tive for left-hand polarization. The angle A isthe difference between the relative phases( &c)w,~ for the two polarizations W and A,.

The Poincark Sphere and the polarization boxr1, PP 3.1-3g.73, which is a graphical repre-sentation of the Stokes parameters, provide aconvenient means of converting from one setof polarization parameters to another. Thepolarization box is shown in Fig 49 which alsoindicates its relationship to the Poincark sphere.It can be shown that the sides and the side

83

IEEEStd 149-1979

H O R I Z O N T A LL I N E A R

Fig 49Polarization Box and Its Relation to the Poincare Sphere

POLARIZATION

diagonals are given by and

Sides Side Diagonals

YL = cos 2a XL = sin 2 (x

YD = cos 20 XD = sin 2 p

YC = cos 27 XC = sin 2 y

As an example of the use of the polarizationbox in converting between polarizationparameters, let PL and 6~ be known frommeasurements, and let it be required to findPC, 6~) y, and the tilt angle 7. The solution isas follows. From the polarization box it canbe seen that

6, = cos-’ [$I = cos-* [=JThe ambiguity in defining 6~ from its cosinealone is removed by inspection of the polariza-tion box. Finally, once PC and 6~ are known,the axial ratio r and the tilt angle 7 can beobtained by use of

so that

YC = X, sinhL

For some applications it is advantageous toexpress the polarization of a field in the formof a unitary vector. A wave W, for example,can be expressed in terms of two orthogonalelliptical polarizations as

cos 2-y = sin 2a sin 6,

From the definition of PL one can compute CY:

01 = tan-’ pL

Thus y can be determined, from which PC canbe calculated by the use of

PC = tan y

where EA , EB, and 5‘ are defined in Fig 43, and8~ is the relative phase between z~ and Eg.A and B can become (H, V), [45], [ 1351, or(L, R), in which case [ and 6~ become (a, SL),(p, 6D), or (y, 6c), respectively.

84

PC + 1r =---PC - 1 ’

7 = 6,/2

MEASUREMENTSIEEE

Std 149-1979

To illustrate the use of polarization vectors,let the incident wave W and the antennareceiving polarization A, be expressed in termsof the circular polarization components, that is,

cm YwW =

i 1sin ywe j(6c)W

A, =j(sCL

The normalized voltage response v of the an-tenna to the wave is proportional to the innerproduct of A, and W which can be expressed as

c = (A,, W) = A,tW

where the superscript t denotes the complexconjugate of the transpose. (Note that V is aphasor.) The matrix operation yields

v = cos yr cos yw + sin Tr sin TWeiA

where

n = (6C)w -&), = 2(~ -TV)

The polarization efficiency is then given by

p = TV* = 1712

The determination of the polar iza t ionefficiency by the vector method is a formal-ization of the process of resolving the powerdensity of the incoming wave and the effectiveaperture of the antenna each into two compo-nents corresponding to two orthogonal polariza-tions such that each component of the wave ispolarization matched to the correspondingcomponents of the antenna’s effective aperture.It should be noted that partial responses dueto the pairs of polarization-matched compo-nents do not, in general, add to give a maxi-

mum value of v. The condition for a polariza-tion match is that yr = yW and ~5 = 0, in whichcase

p = p/2 = (cos2 y + sin’ r)2 = 1

11.2 Polarization Measurements11.2.1 General. The radiation pattern of an

antenna designed for a specific polarization isusually described in terms of the field compo-nents for that polarization. It is only a partialdescription, since a cross-polarized componentmay be present. A complete description ofthe radiation pattern requires the measure-ment of polarization as a function of direction.In particular, away from the direction of thepeak value of the main beam, the polarizationmay be quite different from the design valueand, even over the main beam, its variationmay be of importance.

The various techniques used to measure thepolarization of antennas may be broadlyclassified into three categories:

(1) those that yield partial information aboutthe antenna’s polarization properties

(2) those that yield complete polarizationinformation, but require comparison with apolarization standard

(3) those that yield complete polarizationinformation, but require no standard or a prioriknowledge of the polarizations of the antennasused in the measurement

The methods of category (2) are termed trans-fer or comparison methods, and those of cate-gory (3) are referred to as absolute methods.The method one selects depends upon the typeof test antenna, the required accuracy, theamount of polarization data required, the timeavailable for the measurements, and the permis-sible cost.

The following methods may be employed tomeasure polarization [ 1, pp lO.l-10.381 :

(1) polarization-pattern method(2) rotating-source method(3) multiple-amplitude-component method(4) phase-amplitude method

To completely characterize the polarization


P O L A R I Z A T I O NP A T T E R N

\

POLA‘RIZATIONELLIPSE

Fig 50Polarization Pattern of a Wave

of an antenna, it is necessary to determine thepolarization ellipse (axial ratio and tilt angle)and the sense of rotation of the electric fieldvector of the wave radiated by the antenna.The polarization state of the wave can berepresented as a unique point on the Poincarksphere (see 11.1). Some methods of polariza-tion measurement yield insufficient informa-tion to completely characterize the state of thewave, hence a unique point on the Poincarksphere is not determined. For example, whenthe axial ratio and tilt angle are measured,but the sense of rotation is not determined,there will be an ambiguity between conjugatepoints in the upper and lower hemispheres ofthe Poincarb sphere. This information may beadequate when the sense of polarization isotherwise known or when the polarization isnearly linear so that the two conjugate pointslie close to the equator.

It should be noted that some of the methodsdiscussed hereafter require sequential measure-ments; one or more antennas are rotated for asecond measurement. These methods pose aproblem with regard to system stability, sinceany change in frequency or gain in the systemwill introduce errors in the measurement. Thisis particularly severe when it is necessary tocalibrate an antenna over a large field of view,

86

as is the case in the near-field-probing method(see 7.3). For these methods an extremelystable signal source is required. Frequencysynthesizers and phase-lock techniques may beemployed to achieve the required frequencystability. Those methods of polarization mea-surement where the measurements are madesimultaneously do not require such a highdegree of stability.

112.2 Measurement o f t h e PolarizarionPattern. The polarization-pattern method maybe employed to determine the tilt angle andthe magnitude of the axial ratio, but it doesnot determine the sense of polarization. For thismeasurement the antenna under test can beused in either the receive or the transmitmode. If it is used in the transmit mode, themethod consists of the measurement of therelative voltage response 17 1 of a dipole, orother linearly polarized probe antenna, as it isrotated in a plane normal to the direction ofthe incident field. The magnitude 1 VJ will beunity whenever the system is polarizationmatched.

The magnitude 1 iTI, when plotted as a func-tion of the tilt angle 7, of the receiving polari-zation of the linearly polarized probe antenna,is called a polarization pattern (see Fig 50).The polarization pattern is tangent with the

MEASUREMENTSIEEE

Std 149-1979

LHC

RHC

Fig 51Poincare Sphere Representation of the

Polarization-Pattern Method. LJ= (6~ )w - (SC),

polarization ellipse of the field at the ends ofthe major and minor axes. Hence the magnitudeof the axial ratio and the tilt angle of the inci-dent wave are determined.

The method can be illustrated by the use ofthe Poincark sphere. For this example assumethat the polarization of the incident wave W islocated on the sphere, as shown in Fig 51. Thereceiving polarization A, of the rotatinglinearly polarized probe antenna is always lo-cated along the equator of the sphere. Thesquare root of the relative response Iv 1 isequal to the cosine of one half of the angularseparation of W and A,, as shown in Fig 51. Asthe tilt angle of A, is rotated, its position onthe sphere moves along the equator so that themagnitude of 1” 1 varies. When it is plotted asa function of the tilt angle 7, the polarizationpattern is obtained. It should be noted that thesame polarization pattern would be obtainedif W were located in the conjugate point W’ onthe lower hemisphere. The sense of polarizationshall be measured in order to resolve theambiguity.

The polarization pattern can also be obtainedfrom the polarization vector formulation(see 11.1) since

V = cosy, cosy, + siny, s in y,ej*

where

A = ~~c)w - (SC), = 2 (7w - 7r)

This is illustrated by the phasor diagram inFig 52. As rr varies, the magnitude of Iv1 variesaccordingly, and when plotted as a function of7, the polarization pattern results.

Fig 52Generation of the Polarization Pattern

Using the Polarization Matrix Result.A = @c)w -Wr

87

_: _- - - -.. .~


I”.. 0 90 180

A N G L E ( D E GR E ES )

Fig 53Continuously Scanned Polarization Pattern as a Function of Angle 8

The polarization efficiency (see 11.1) for thesystem is equal to 1 V ( * so that it can be ob-tained directly from the polarization pattern.This is a useful result if the test antenna is tobe used in a system with a linearly polarizedantenna whose spatial orientation is known. Adisadvantage of the polarization-pattern methodis that the probe antenna shall be rotated 360’while the antenna under test is fixed so that itis not convenient to obtain polarization infor-mation as a function of direction.

11.2.3 Rotating-Source Method. The axialratio (and not the sense of polarization or thetilt angle) can be determined as a function ofdirection by using the rotating-source method.The method consists of continuously rotatinga linearly polarized source antenna as thedirection of observation of the test antenna ischanged. This method is of greatest value fortesting nearly circularly polarized antennas.The rotating source antenna causes the tiltangle 7w of the incident field to rotate at thesame rate. The rate of rotation i, shall be verymuch greater than that of 4 or 6 as 0 or Q cutsare made and recorded. Care shall be taken toensure that the time response of the recordingsystem can adequately follow the excursionsin 7W. An extension of this technique to aspiral cut pattern (see 3.2), where 7w, 8 and@ are all varying and iW > > 4 > > ,$, permits

recording on one pattern of the axial ratio foressentially all directions from the antenna [ 761.

A pattern of an eliptically polarized antennaobtained by the rotating source method isshown in Fig 53. If the amplitude variationsare plotted in decibels, the axial ratio, also indecibels, for any direction in space that isrecorded on the pattern is the width of theenvelope of the excursions. This particularantenna is essentially circularly polarized onaxis (19 = 0’) and elliptically polarized in thedirection of the maxima of the minor lobes.This sense of polarization is not available fromthis pattern.

11.2.4 Multiple - Amplitude - ComponenrMethod. The multiple-amplitude-componentmethod can be used to determine the polari-zation completely without the measurement ofphase. It has been shown that the polarizationof a wave can be determined from the magni-tudes of the responses of four antennas havingdifferent, but known, polarizations [70, pp540-5441, [77], [78]. The most convenientchoices of polarization for the sampling antennaare horizontal or vertical linear, 45” or 135”linear, and right-hand or left-hand circular, andin addition, any fourth component from thisset of six components. These sampling antennasshall have known gains, and the instrumentation

MEASUREMENTSIEEE

Std 149-1979

Fig 54Multiple-Amplitude-Component Method of Polarization Measurement

shall be suitably calibrated to compensate forgain differences. From these data the polariza-tion of the wave, and hence that of the testantenna, can be completely determined.Graphical construction or a solution of a sys-tem of linear equations may be used to find theStokes parameters.

Usually it is more convenient to measure themagnitudes of the polarization ratios. Hence allsix components are used [77]. This method isillustrated by use of the Poincark sphere asshown in Fig 54. The linear, diagonal linear,and circular polarization ratios (PL, PD, andPC? respectively) are measured. From thesedata 2 CX, 2p, and 2 y, the angles that define theloci of all possible polarizations on thePoincare sphere and correspond to the polari-zation ratios PL, pD, and PC, respectively, aredetermined. The common intersection of thethree loci determines the polarization of thewave.

The axial ratio r and the sense of polarizationare determined from pc since

PC + 1r=-

PC - 1If r is positive the sense is right handed, and if ris negative the sense is left handed. The phaseangle of the complex circular polarization ratiocan be computed by the use of

hf.2 = t a nYD- 1 _-

YL

where from the polarization box (see 11.1)

and

Y, =Cl- Ph

(I+ Ph II

Y, =(1 -Pi,

u+P;)

Since the tilt angle is one half of Fc, the polari-zation is completely determined.

A modified version of the multiple-amplitude-component method, which involves measure-

89

POLARIZATION

D U A L P O L A R I Z E DSAMPLING ANTENNA

M I X E R

IEEEStd 149-1979

A N T E N N AUNDER TEST W kEfl

P= = a D A M P L I T U D E 6

C0 R E C E I V E R

kEL

I

M-IXER

Fig 55Instrumentation of the Phase-Amplitude Method of Polarization Measurement

ments with a single linearly polarized antenna,may be employed to determine the axial ratioand tilt angle over the entire radiation patternof the test antenna, provided that the sense isnot required. The pattern of the test antenna ismeasured with the source antenna oriented at0” (horiztonal), 45’, 90” (vertical), and 135”.From these data the polarization ratios PL and&) are determined. With these results 6~ canbe computed. The value obtained can be usedto determine the tilt angle. The axial ratio isgiven by

r = -cot h

where

x = $ cos-l (Y; + yp2

This last result can be obtained from the polari-zation box (see 11.1). For highly accurateresults the linearly polarized antenna should bea polarization standard.

112.5 Phase-Amplitude Method. In thephase-amplitude method all the data requiredfor complete polarization determination can bemeasured simultaneously, permitting the com-plete polarization and radiation patterns of anantenna to be measured with a single runthrough a complete set of pattern cuts. Theinstrumentation required is illustrated in Fig55. A dual-polarized receiving antenna is usedto sample the field of the antenna under testwhich is, in this case, used as a transmitting

antenna. The outputs of the receiver will bethe magnitudes of the responses for each of thepolarizations of the sampling antenna and theirrelative phases. If the two polarizations areorthogonal, the complex polarization ratio canbe obtained. The polarizations of the samplingantenna shall be known, and the gains of thetwo antenna-receiver channels shall be madeidentical.

In general it is not economically feasible todesign a dual-polarized sampling antenna withknown pure polarizations. For antenna rangesthat employ automated instrument&ion with acomputer, it is not necessary to use antennasthe polarizations of which are precisely linearor circular, as the case may be, since the mea-sured data can be adjusted by computation,provided that the actual polarizations of thesampling antenna are known. For example,suppose that the measurement requires asampling antenna with orthogonal circularpolarizations, but the sampling antenna haspolarizations which are almost, but not quite,circular. The polarizations of the samplingantenna can be measured by use of the modifiedmultiple-amplitude-component method, pro-vided the senses of polarization are known.These data can be entered into the computer.The computer software can be designed to auto-matically compensate for the characteristicsof the sampling antenna.

If the antenna range is not automated and nocomputer is available, an alternate approachwould be the use of a polarization adjustmentnetwork, external to the sampling antenna, to

90

‘(_’Iii

MEASUREMENTS

LHC SAMPLING ANTENNA

IEEEStd 149-1979

T A N +,

RHC SAMPLING ANTENNA

Fig 56Phase-Amplitude Measurements with Circularly Polarized Antennas

provide the desired polarizations [ 1, pp 10.1%10.231. A typical sampling antenna mightconsist of two orthogonal linear antennas; theymay be feeds for a single reflector, for example.The antennas are connected to the inputs of aphase-amplitude receiver (see 5.4). The polari-zation adjustment network, in its simplestform, consists of an attenuator and a phaseshifter inserted in each channel. These networkscan be designed to operate at radio frequencyand can be inserted between the sampling an-tenna and the receiver’s mixer. Alternately, thenetwork can be designed to be operated atan intermediate frequency. Thus the same net-work can be used for any frequency covered bythe receiver. If the measurement system is cap-able of recording digital data, and a computeris available, a digital polarization adjustmentnetwork may be employed.

If it is desired that right- and left-handedcircular polarizations be used, adjustment ofthe network can be made by employing thepolarization-pattern method. A nominallylinear reference antenna can be used. Usuallythe accuracy of adjustment to circular polari-zation is limited by range reflections, misalign- y = tan-’ PC

91

ments, and so on. When precise measurementsare required, a polarization standard is neededfor the adjustment of the gains of the twochannels. If a linearly polarized standard isused, the output signals from the two channelsare adjusted to be equal. The phase angle 6~ isset equal to zero for horizontal polarization.The system can be checked by rotating thestandard through 360’. The output levelsshould remain constant, whereas the phaseangle 6~ should always be twice the value ofthe rotational angle of the standard, measuredwith respect to the horizontal.

The measurement of the polarization of thetest antenna using circularly polarized samplingantennas is illustrated by use of the Poincaresphere in Fig 56. The complex circular polari-zation ratio

A

PC = pcej% = f!$_ ,&[I

and


T A N a

Fig 57Phase-Amplitude Measurements with Linearly Polarized Sampling Antennas

are obtained from the measurement, and hencethe polarization of the wave radiated by thetest antenna is uniquely determined.

When orthogonal linear polarizations areused, a polarization standard, usually linear, isrequired to adjust the polarizations of thesampling antennas and the gains of the twochannels. If a linearly polarized standard isemployed, the polarizations of the two samplingantennas are first adjusted for optimum nullsfor the cases where the orientation of thestandard should be cross polarized for eachsampling antenna. The outputs of the twochannels are then adjusted to be equal withthe standard oriented to 45” or 135” withrespect to the horizontal. The phase 6~ shouldbe set to zero with the polarization standardrotated to 45” and should remain at zero or180’ as the standard is rotated through 360’.The sense of polarization can be checked usingany elliptically polarized antenna with a knownsense of polarization, such as a helical antenna.The phase angle 6, should be positive and lessthan 180” for left-hand elliptical polarization. Itshould be negative and with a magnitude lessthan 180’ for right-hand elliptical polarization.

When the field of the test antenna is sampled,

the measured result is

A

PL= pLeh

This can be converted to the complex circularpolarization ratio by use of the relation

& = pee 9cl+j;L

=

l-j&

from which the axial ratio and tilt angle maybe found (see 11.1). The method is illustratedby use of the Poincare sphere in Fig 57.

For many applications adequate accuracy canbe obtained if care is taken in the fabricationof a linearly polarized antenna which is to beused as a standard without proof of its polari-zation. Where precision is required, the stan-dard shall be calibrated by an absolute method.The three-antenna method can be employed forthis purpose [ 791, [80]. The only restrictionsplaced on the antennas for this method are thatat least two of them not be circularly polarizedand that the polarization be such that sufficientsignal levels are maintained. The method is

92

MEASUREMENTSIEEE

Std 149-1979

-<I3 - P-e- v31,0°

1731,90°

Fig 58Three-Antenna Absolute Method of Polarization Measurement

?“12,0 -- Relative Response at Antenna 2 with Its Orientation at 0” ;

h2,90 - Relative Response with Its Orientation at 90” ; etc

illustrated in Fig 58. The response is measuredtwice for each combination of two antennas.The receiving antenna in each case is rotatedthrough 90’ for the second measurement. Theratio of the two responses (the antenna at 0’and at 90”) is formed. The result yields threecomplex equations:

where

p^cI, SC2 and &s are the complex circular polari-zation ratios corresponding to the antennas’polarizations. These equations can be solvedsimultaneously, and the complex polarizationratios for each antenna can be determined.Thus all three antennas can be calibrated withthis procedure; for a modified approach see[153].

93

IEEEStd 149-1979

12. Measurement of Power Gainand Directivity

12.1 General. The power gain of an antenna,in a specified direction, is 4n times the ratioof the power radiated per unit solid angle inthat direction to the net power accepted bythe antenna from its generator. This quantityis an inherent property of the antenna and doesnot involve system losses arising from a mis-match of impedance or polarization. To deter-mine the power transfer in a complete system,the antenna input impedance (see 16.1) andthe antenna polarization (see 11.2) shall bemeasured and taken into account.

The directivity of an antenna in a specifieddirection is 47r times the ratio of the powerradiated per unit solid angle in that direction tothe total power radiated by the antenna. Thisterm differs from power gain because it doesnot include antenna dissipation losses.

The ratio of power gain to directivity in thesame direction is the radiation efficiency of theantenna (see Section 13).

The power gain (or directivity) of an antennais usually measured in the direction of its maxi-mum value, and that value is called the peakpower gain (peak directivity). Knowledge ofthe radiation pattern (see Section 3) permitsthe determination of the gain in any otherdirection. Of particular importance, when con-sidering pencil-beam antennas, is the determina-tion of side lobe levels. Most often the side lobelevels are referenced to the peak gain of theantenna. Another way to express the side lobelevels is relative to the gain of a lossless isotropicradiator. In both cases the levels are expressedin decibels. Since it is not uncommon for themagnitudes of these two results to be approxi-mately the same for a given side lobe, confusioncan arise if the gain reference is not properlyspecified.

The techniques that are employed for thedetermination of the power gain of an antennaare dependent upon its frequency of operation.Above 1 GHz, for example, free-space antennaranges (see Section 4) are commonly available

MEASUREMENT OF POWER GAIN

for gain measurements. For these frequenciesmicrowave techniques can be employed sincewaveguide devices, including electromagnetichorns, are readily available.

For frequencies between 0.1 and 1 GHz,ground-reflection ranges are usually requiredbecause free-space conditions are difficult tosimulate. Because of the longer wavelengths,microwave techniques become less practical atthese frequencies. Antennas operating in thisfrequency range are often mounted on struc-tures, such as aircraft, which affect theircharacteristics. For these cases, scale-modelingtechniques may be employed (see 7.1). Sinceit is impractical to scale the finite conduc-tivities and loss factors of the materials ofwhich the antenna and aircraft are constructed,power-gain measurements cannot be performedusing the scale model. However, the directivitycan be measured. If the efficiency of the full-scale antenna can be established by othermeans, then the power gain can be determinedsince the directivity of a properly scaled modelantenna and of the corresponding full-scaleantenna are the same. It is good practice toverify the results by requiring the aircraft, withthe full-scale antenna mounted, to fly a pre-scribed path relative to a suitable groundstation. The system performance, using thefull-scale test antenna, can be measured andcompared to that predicted from the scale-model measurements.

As the frequency is decreased below 0.1 GHz,the effect of the ground upon the antenna’scharacteristics becomes increasingly pro-nounced, making power-gain measurementsvery difficult. Directive antennas at these lowerfrequencies are physically large and must bemeasured in situ (see 12.3.3). Often it is satis-factory to calculate the antenna gain andestimate the effect of the ground. Again, scalemodels can be employed. However, because ofthe strong effect of the ground on the char-acteristics of the antenna, the electrical proper-ties of the ground shall also be scaled.

For frequencies below about 1 MHz theantenna power gain is not usually measured,but rather it is the field strength of the groundwave radiated by the antenna which is mea-sured (see Section 17).

94

AND DIRECTIVITY

For those frequencies for which power-gainmeasurements are practical, there are twogeneral categories into which the variousmethods can be placed: absolute-gain measure-ments and gain-transfer measurements.

For an absolute-gain measurement no a prioriknowledge of the gains of any of the antennasused in the measurement is requried. Thismethod is usually employed for the calibrationof gain-standard antennas and is rarely used inany laboratory except one that specializes inthe calibration of standards.

The gain-transfer method, which is alsoreferred to as the gain-comparison method, isthe most commonly employed method forpower-gain measurements. This method requiresthe use of a gain standard to which the gain ofthe test antenna is compared.

The directivity of a test antenna is obtainedby integrating the measured far-field radiationpatterns of the antenna over a closed sphericalsurface. If the antenna losses can be determinedby other means, then the power gain of theantenna can be determined from the directivitymeasurement [ 811.

12.2 Gain Standards12.2.1 Types of Gain Standards. Antennas

which are to be used for gain standards shouldhave the following characteristics:

(1) The gain of the antenna shall be ac-curately known.

(2) The antenna shall have a high degree ofdimensional stability.

(3) The antenna’s polarization should belinear or, for some applications, circular. Ifcircular polarization is required, two antennasare needed, one right-hand circular and theother left-hand circular. Whether the antenna islinearly or circularly polarized, a high degree ofpolarization purity is necessary.

Although any antenna meeting these criteriamay be used as a gain standard, the two uni-versally accepted antenna types are the dipoleand the pyramidal horn. The gains of theseantennas can be fairly accurately calculated,and because of their mechanical simplicity theycan be manufactured with a high degree ofreproducibility.

A thin dipole antenna, properly driven, has

IEEEStd 149-1979

a gain of approximately 2.15 dB when itslength is adjusted to half-wavelength resonance.(The calibration consists of the adjustment ofthe dipole length to obtain resonance.) Theseantennas are commercially available coveringfrequencies from 77 MHz to 1 GHz.

Pyramidal-horn antennas are commerciallyavailable for frequencies from 0.35 to 90 GHzwith nominal power gains from about 14 to25 dB. The calibration curves accompanyingthem are usually based upon data published bythe United States Naval Research Laboratory[82] . These calibration curves have a smoothcharacteristic, whereas the gain when measuredas a function of frequency exhibits an un-dulating characteristic that is believed to bedue to multiple reflections between the hornaperture and the transition between the wave-guide and the horn [ 131, [83]. The computed

,calibration is considered accurate to + 0.25 dBfor frequencies above 2.6 GHz and to + 0.50 dBbelow 2.6 GHz. If a higher degree of accuracyis required, the antenna shall be calibrated byan appropriate standards laboratory.

Both the dipole and the pyramidal-hornantennas are nominally linearly polarized. Thedipole antenna by itself has a very high degreeof polarization purity. However, because of itsbroad pattern its characteristics can be greatlyaffected by its environment and especially bythe transmission line used to feed it. For thisreason it is difficult to place limits of uncer-ta in ty on its polarization. Pyramidal-hornantennas, having directional patterns, are lessaffected by their environment when used forgain measurements. However, they are likelyto be slightly elliptically polarized on axis,with axial ratios between 40 dB and approach-ing infinity.

It is sometimes necessary to design a gainstandard with special properties. For example,the dipole antenna has an omnidirectionalpattern in its principal H plane, and thereforeits pattern can be greatly altered by the measure-ment environment. Because of this it is some-times necessary to design a directional antennasuch as a dipole array with a reflector, acomer reflector antenna or a log-periodicantenna which is then calibrated for use as a

95

IEEEStd 149-1979 MEASUREMENT OF POWER GAIN

P P0 r

I

- - - _ _ _CL

R

Fig 59Two-Antenna System Illustrating the Friis Transmission Formula

gain standard. The corrugated conical hornantenna is often used as a gain standard, partic-ularly for the measurement of the gain of an-tennas operating at millimeter wavelengths[151].

Absolute-gain measurements are usuallyrecommended fo r t he calibratioi? of gainstandards. It should be emphasized that if ahigh degree of accuracy is required, the an-tenna should be calibrated by a standardslaboratory specializing in the calibration ofgain standards.

12.2.2 Calibration of Gain Standards on aFree-Space Antenna Range. Absolute-gain mea-surements are based upon the Friis transmissionformula, which states that for a two-antennasystem as shown in Fig 59, the power receivedat a matched load connected to the receivingantenna is given by

If the two antennas are identical, it followsthat their gains are equal so that

The procedure in determining the power gainof the antennas is to measure R, X, and 10 logP, /Pr and then compute (GA)~B. Since twoidentical antennas are required, this method isreferred to as the two-antenna methbd. I fantennas A and B are not identical, a thirdantenna is required to determine the gains.

For the three-antenna method three sets ofmeasurements are performed using all combina-tions of three antennas. The result iS a set ofthree simultaneous equations of the form

P, = P,GAGB

where P, is the power received, PO is the poweraccepted by the transmitting antenna, GA isthe power gain of the transmitting antenna,and GB is the power gain of the receiving an-t e n n a [l, pp. 8.2-8.81. This form of thetransmission formula implicitly assumes thatthe antennas are polarization matched for theirprescribed orientations and that the separationbetween the antennas is such that far-fieldconditions prevail.

The Friis transmission formula can be writtenin logarithmic form, from which the sum of thegains, in decibels, of the two antennas can bewritten as

(GA),, +(GB)dB = 20 1%(y) -lOlog

tGA )dB + tGB)dB

tGA )dB + tGC )dB

tGB)dB + tGC)dB = 20 ‘og(y-)- 10 q$ BC

From these equations all three gains can bedetermined.

The block diagram of Fig 60 is typical of theinstrumentation required for the measurementof gain using the two-antenna or three-antennamethods. The instrumentation shall be highlystable with the source producing a singlesinusoidal frequency. With reference to Fig 60the procedure is to first calibrate the coupling

96

AND DIRECTIVITY

C A L I B R A T E D

IEEEStd 149-1979

COUPLING NETWORK

r - - - - - - - - - - - r - - - - - --.-- 1

A T T E N U A T O R If------_J L---_------J(TEST ASSEMBLY TEST ASSEMBLY

A B

TRANSMIT TEST POINT RECEIVE TEST POINT

C O U N T E R

Fig 60Typical Instrumentation for Two-Antenna and Three-Antenna Methods

of Power-Gain Measurement

network between the source and the trans-mitting antenna so that the power measured atthe transmit test point can be accurately relatedto the power into antenna A. Then all com-ponents of the system are impedance matchedusing tuners. The two antennas are separatedso that far-field conditions prevail and bore-sighted so that they are properly aligned andoriented.

The attenuator of the coupling network isadjusted so that the power level at the transmittest point is the same as that at the receive testpoint. From the calibration of the couplingnetwork the relative power level PO/P, can bedetermined.

If the gains of broad-band antennas are to bemeasured, it may be necessary to use the swept-frequency technique (see 7.4). Both the two-antenna and the three-antenna methods can beemployed [ 1, pp 8.20-8.241. A block diagramof a typical instrumentation setup is shown inFig 61. It should be noted that it is not possibleto match all the components over a band offrequencies. Therefore the impedances orreflection coefficients of all components shallbe measured. The swept-frequency techniqueshould be used for those measurements.

12.2.3 Calibration of Gain Standards on aGround-Reflection Antenna Range. For fre-quencies below about 1 GHz, antennas whichmight be chosen for use as gain standards will

necessarily have moderately broad beams.For these antennas, ground-reflection rangesare often used if accurate gain measurements arerequired. With some restrictions and modifica-tions, the two- or three-antenna method of gainmeasurement can be used on a ground-reflectionrange [84].

The method, as described hereafter, is limitedto linearly polarized antennas that couple onlyto the electric field. To be used with loopantennas, the equations would have to bemodified. Since the reflective properties of theearth are different for vertical and horizontalpolarizations, elliptically or circularly polarizedantennas are excluded. It is recommended thatthe antennas be oriented for horizontal polari-zation because of the rapid variation of theearth’s reflection coefficient as a function ofthe angle of incidence when vertical polariza-tion is employed. Such a phenomenon is notpresent for horizontal polarization.

The criteria for ground-reflection ranges,outlined in 4.4, should be satisfied. Thegeometry of the ground-reflection range forgain measurements is shown in Fig 62. It isdesirable that the range length R, be sucht h a t R, >> 2h,, where h, is the height of thereceiving antenna. When the height of thetransmitting antenna is adjusted so that thefield at the receiving antenna is at the firstmaximum closest to the ground, then the gain

97


r - - - - - - - - -1

Fig 61Typical Instrumentation for Swept-Frequency Two-Antenna and

Three-Antenna Methods of Power-Gain Measurement

RECEIVEANTENNA

TRANSMITANTENNA

-fh

t

I/

//

I /I / \ RANGEI /

/ SURFACEI /

ANTENNA

-R 0

Thr

I

Fig 62Ground-Reflection-Range Geometry

98

AND DIRECTIVITY

sum equation for the two-antenna or the three-antenna method can be modified to read

(GA)~B +(%)d~ =2olog (y#)- 10 log(gfj

‘-Rd+ F

R 1where Oil and DB are the directivities alongR. relative to the peak directivities of antennasA and B, respectively. The factor r is to bedetermined. It is a function of the electricaland geometrical properties of the antennarange, the radiation patterns of the antennas,and the frequency of operation. D, and Daare obtained from amplitude patterns of thetwo antennas which should be measuredprior to performing the gain measurement. Th‘equantities RD, RR, A, and the ratio P, /P, arequantities that are measured directly. Once ris determined, the gain sum can be evaluated.

To obtain r the preceding measurement isrepeated, but this time with the height of thetransmitting antenna adjusted to a positionsuch that the field at the receiving antenna isa minimum. To distinguish the quantities mea-sured with this geometry from those of theprevious one, let all the quantities associatedwith the latter geometry be represented by thesame letters, except with primes. With thisnotation the equation for r can be written as

where RR, RD, and P, are obtained with thetransmitting antenna adjusted for a maximumsignal at the receiving antenna, and Rh, Rb , andPi are obtained with the antenna adjusted fora minimum signal at the receiving antenna. Therelative directivities DA, DA, Dg, and Db areobtained from the amplitude patterns of thetwo antennas.

IEEEStd 149-1979

The instrumentation for this measurement isessentially the same as that for the free-spacerange measurement (see Fig 60). Accuraciesof _+ 0.3 dB are attainable with this method.

12.2.4 Calibration of Gain Standards on anExtrapolation Antenna Range. Multipath andproximity effects are always present for mea-surements made on antenna ranges (see 4.2 and12.5). The extrapolation technique [79] in-cludes provisions for rigorously evaluating andcorrecting for errors due to these effects. Thereceived signal is measured as a function of thespacing between the transmitting and thereceiving antennas. The multipath effect ismanifested by a cyclic characteristic on thecurve obtained by plotting the received signalas a function of spacing.

The cyclic variation in the received signal dueto the effect of multipath can be averaged outmathematically or by adjusting the time con-stant of the instrumentation so that it cannottrack the cyclic variations but rather allows theaverage value to be recorded. By fitting a curveto the averaged data, the signal level can beextrapolated to that which would be measuredin the far field. The method of curve fitting isdescribed in detail in [79] . In this mannerboth the proximity and the multipath inter-ference effects are removed. From this resultthe power gain can be computed.

The extrapolation technique, when com-bined with the generalized three-antennameasurement technique [32], [33], [79], iscapable of yielding not only the power gainsbut also the polarizations of the three antennas(see 11.2). There is the restriction that none ofthe three antennas be nominally circularlypolarized. If one of the antennas is circularlypolarized, then only that antenna’s characteris-tics may be completely determined. If two ormore of the antennas are nominally circularlypolarized, the method fails.

’

The extrapolation range shall have a precisionmovable tower which allows boresight betweenthe transmitting and the receiving antennas tobe maintained as it is moved over the length ofthe range. Measurements may be conducted atdistances less than 20 2/h, where D is the maxi-mum dimension of the antennas under test.

i

99

IEEEStd 149-1979

The tower heights should be at least 15 percentof the maximum separation between antennas.This places a practical limitation on the gainsof antennas that may be tested since the re-quired maximum spacing between towersincreases as the gains of the test antennas areincreased. Since the tower heights must alsobe proportionally increased, it becomes increas-ingly difficult and expensive to construct amovable tower that maintains boresight overthe required length of the range. Antennas withgains of about 40 dB represent a practicalupper limit [ 85] .

This technique is capable of yielding measuredgains with uncertainties as low as +0.05 dB andwith more routine measurements to +0.08 dB[152].

12.3 Gain-Transfer Measurements12.3.1 Measurement of Linearly Polarized

Antennas. The gain-transfer method is onein which the unknown power gain of a testantenna is measured by comparing it to that ofa gain-standard antenna [l, pp 8.15-8.181.The measurements can be performed on eithera free-space or a ground-reflection range. (Itcan also be performed with a test antenna thepower gain of which has to be measured i nsitu as discussed in 12.3.3 and 12.4.)

Ideally the test antenna is illuminated by aplane wave which is polarization matched toit, and the received power is measured into amatched load. The test antenna is replaced bya gain standard, leaving all other conditionsthe same. The received power into its matchedload is again measured. From the Friis trans-mission formula it can be shown that thepower gain (GT )dB Of the test antenna, indecibels, is given by

(GT)dB = (Gs)dB + ~o~ogpT/~S

where (Gs)dB is the power gain of the gain-standard antenna, PT is the power receivedwith the test antenna, and Ps is that powerreceived with the gain- standard antenna.

One method of achieving this exchange be-tween test and gain-standard antennas is tomount the two antennas back to back on either


side of the axis of an azimuth positioner.With this configuration the antennas can beswitched by a 180“ rotation of the positioner.Care shall be taken to position the antennas sothat they will be in the same location whenswitched. Usually absorbing material is requiredimmediately behind the gain standard to reducereflections in its vicinity which might perturbthe illuminating field.

Swept-frequency gain-transfer measurementscan be performed for testing broad-band an-tennas [l, pp 8.23-8.241. The procedure isessentially the same as that for the swept-frequency absolute-gain measurement (see12.2.2), except that the measurement is re-peated with the test antenna and the gainstandard. The reflection coefficients of all thecomponents shall be measured as a function offrequency so that corrections can be made tothe measured power gain.

12.3.2 Measurement of Circularly and Ellip-tically Polarized Antennas. For the special caseof circularly polarized test antennas it is possibleto design and calibrate orthogonal circularlypolarized gain-standard antennas. This is parti-cularly useful when production runs of power-gain measurements are required. In general,however, the power gains of circularly andelliptically polarized test antennas are mea-sured with the use of linearly polarized gainstandards. This is valid because the total powerof the wave radiated by an antenna can beseparated into two orthogonal linearly polarizedcomponents (see 11.1). Thus partial power-gainmeasurements can be performed using twoorthogonal linearly polarized gain standardsfrom which the total power gain of the testantenna can be determined. A single linearlypolarized gain standard can be employed androtated 90” to achieve the two orthogonalpolarizations.

A power-gain-transfer measurement as out-lined in 12.3.1 is performed with the gainstandard oriented for vertical polarization.The measurement is then repeated with thegain-standard and source antennas orientedfor horizontal polarization. From these partialpower-gain measurements the total power gain(GT)dB, of the test antenna, in decibels, can

100

AND DIRECTIVITY

be computed by use of the equation

(GT)dB = 10 lOk3 (GTv + GTH)

where GTV and GTH are the partial powergains with respect to vertical linear polarizationand horizontal linear polarization, respectively.These measurements can be performed on eithera free-space or a ground-reflection range.

12.3.3 Measurement in the High-FrequencyRange (3-30 MHz). At frequencies betweenabout 3 and 30 MHz medium- and long-distancepropagation of electromagnetic waves dependschiefly upon sky waves reflected from theionosphere. For paths of 2000 to 4000 km theantenna pattern and power gain at elevationangles between 5 and 30” above the horizonare of primary importance. Antenna-patternand power-gain measurements shall be madeat full scale and in situ as discussed in Sec-tion 9 [59].

The measurements require the use of anaircraft with a suitable transmitting atennamounted on it, which flies at fixed altitudesalong circular paths centered upon the testantenna. To measure the pattern, the ampli-tude of the received signal along with theazimuthal angles and elevation angle to theaircraft are recorded. The power gain is mea-sured by comparing the power received by thetest antenna to that received by a gain-standardantenna located near the test antenna.

A horizontally polarized dipole antenna maybe used as a gain standard, At these frequenciesthe pattern and power gain of a dipole antennaare greatly affected by the ground. For the gainmeasurement the dipole height is adjusted sothat its main lobe is pointing in the same direc-tion as that of the test antenna (between 5 and30’ from the horizon). The power gain (Gs )dBof the dipole antenna, in decibels, as a functionof height and elevation angle can be computedfrom the following equation:

(G,),, = 2.15 + 10 logRI,

,1 + Re(RH (90”)&)

le jkH sin q+ RH (*) e--jkH sin *

IEEEStd 149-1979

where Re means the real part of, R,, is the realpart of the self-impedance of the dipole in freespace, 2, is the mutual impedance betweenthe dipole and its image in a perfect conductor,k is the free-space wave number, His the heightof the dipole above ground, XI-’ is the elevationangle of the first lobe of the vertical radiationpattern of the dipole, and RH (!I!) and RH(90’) are the complex reflection coefficients ofthe imperfect ground for horizontal polariza-tion at angles \k and 90°, respectively, mea-sured from the horizontal (the complement ofthe angle of incidence). With the use of thisequation, contour plots of the, power gain ofthe dipole as a function of height above groundand elevation angle can be constructed [86] .

It is not practical to employ verticallypolarized gain-standard antennas for this pur-pose [87], since the power gain of such anten-nas at low elevation angles is low, and variesdrastically with the moisture content of theground. Thus if the test antenna is verticallypolarized, it is necessary to compare its cross-polarized partial gain (partial gain with respectto horizontal polarization) to a horizontallypolarized gain standard for the gain-transfermeasurement. From this result the partial gainwith respect to vertical polarization, which isthe quantity desired, can be determined.

This can be accomplished by rotating thetransmitting antenna on the aircraft in such away that the transmitted signal is alternatelyvertically and horizontally polarized. It isusually necessary to calibrate the rotatingantenna since its gain will vary as it is rotated.The signal is received by the test antenna andthe gain standard. A comparison between themaximum power received by the gain standardand the minimum signal received by the testantenna yields the partial gain of the testantenna with respect to horizontal polarization.A comparison of the maximum and minimumpowers received by the test antenna thenyields the partial gain with respect to verticalpolarization which is the. gain sought.

The accuracy of this technique depends uponhow well the characteristics of the ground are

101


10-l I IO 100FREQUENCY (GHz 1

Fig 63Flux-Density Spectra of Several Radio Stars (from [92] )

known and upon the terrain of the antenna site.If the site is flat and unobstructed and theground constants are accurately known, theuncertainty in the measured gain can bet 0.5 dB or better. In general, however,uncertainties of f 1 dB can be conservativelyexpected.

12.4 Measurement of the Power Gain ofElectrically Large Antennas

1 2 . 4 . 1 G e n e r a l . The measurement of thepower gain of electrically large antennas isgenerally impractical and in some cases evenimposs ib le to perform on an antenna tes t

range [88], [89]. T h i s i s m a i n l y d u e t o t h efact that the antenna-range length, required tosatisfy the far-field criterion, may be severaltens of kilometers. In addition, large steerableantennas usually experience gravity-inducedstructural deviations as the antenna is moved.This necessitates the measurement of gain as afunction of the antenna’s elevation pointingangle. These problems can be largely overcomeby the use of extraterrestrial radio sources orsatellite-borne beacons [ 58, pp 10-151, [ 851,[88] -[90].

The locations and radiation flux densities ofcertain extraterrestrial radio sources are ac-

102

AND DIRECTIVITYIEEE

Std 149-1979

Table 1Information about Several Radio Sources [ 921

Position Size VisibilityRight Right North South

Radio Ascension Declination Ascension Declination Latitude Latitude SpectralStar Type hours degrees hours degrees degrees degrees Index

Cas A SR 23.4 58.6 3.43 X 4 9 0 11 - 0.787Tau A SR 5.5 22.0 x 4 9 0 4 8 - 0,263Orion A EN 5.5 - 5.4 3.5 x 3.5 65 7 5 0Cyg A RG 20.0 40.6 1.6 x 1 90 2 9 - 1.205Virgo A RG 12.5 12.7 1 X 1.8 73 5 7 - 0.853DR 21 EN 20.6 42.2 < 0.3 9 0 2 8 1.75, - 0.13

*SR = supernova remnant; EN = emission nebula; RG = radio galaxy.

curately known so that they may be used forpower-gain measurements. Furthermore thesesources cover a broad frequency spectrum.Their flux densities vary with frequency, butthey have been accurately measured over the100 to 10 000 MHz range [91].

Man-made beacons aboard satellites have theadvantage of being coherent sources, and asufficient signal-to-noise ratio can usually beobtained to test moderate gain antennas. Thedisadvantages are that the beacon is usuallylimited to a single frequency, and for accuratemeasurements a geostationary satellite shallbe used. Thus the measurement is limited to asingle pointing direction. If the incident fluxdensities from satellites are not accuratelyknown, the gain-transfer method of gain mea-surement shall be employed for these sources.

12.4.2 Use of Extraterrestrial Radio Sourcesfor Power-Gain Measurements. The flux-densityspectra of several radio stars are given in Fig 63.The location, size, spectral index cy, and typeof phenomenon believed responsible for theradio star are given in Table 1. Also given isthe visibility of the source. Visibility refersto that interval of terrestrial latitudes fromwhich the source’s apparent daily path throughthe sky rises to at least 20’ above the horizon.This 20” minimum assumes that the sourcecan be seen by an antenna for at least one hourand is sufficiently far above the horizon to bediscernible above the atmospheric and groundnoise [92].

Cassiopeia A, Cygnus A, and Taurus A arethe strongest resources of relatively smallangular size. Their flux densities have beenaccurately measured over a frequency range of30 MHz to 16 GHz [93]. These are the mostuseful sources for antenna-gain measurements.It, should be noted, however, that Cassiopeia Ahas an annual decrease in flux density of0.9 + 0.1 percent, Cygnus A has a curvature inits spectrum, and Taurus A shows a significantdegree of linear polarizatio-n [ 891. These effectshave been thoroughly studied and documented.Because of their small angular size, they maybe treated as discrete or point sources, pro-vided the test antenna’s beamwidth is greaterthan 10 minutes.

The tabulated flux densities for celestialradio sources are usually given at discretefrequencies, which may be different from thefrequency at which the antenna’s gain is to bemeasured. The relation between power fluxdensity S and radio frequency f can be expressedas Sa f”, where (Y is the spectral index. Thusone can use the known flux density at thefrequency that is nearest to the test frequencyto compute the required flux density. If fk isthe frequency at which the flux density for agiven source is known, and f the frequency atwhich the gain is to be measured, then

/g\”S(f) = i S(fk)

u

103

MEASUREMENT OF POWER GAINIEEEStd 149-1979

The noise power radiated by a radio star is ingeneral randomly polarized. Thus the polariza-tion efficiency of the wave and antenna systemis %, and the power received when the an-tenna is pointing at the radio star is given by

SA, _ SX2GTp, =-_2 8n

W/Hz1

= kT,

where A, is the effective area, GT is the an-tenna gain, X is the wavelength, k is Boltzmann’sconstant, S is the power flux density, and T, isthe effective noise temperature of the radiostar referred to the antenna terminals. (It isthe measured antenna temperature due to thesource above that due to sky background.)From this the gain of the antenna can bewritten as

8rkT,GT = Shy

The equation is valid for a point-source radiostar radiating through a lossless atmosphere.Since this condition is not normally satisfied,correction factors shall be inserted in theequation, that is, the gain can be written as

8nkT,G, = sx2 K,K,

where K , corrects for atmospheric attenua-tion and K2 for the effect of the angular extentof the source.

The factor K, can be obtained from the ap-proximate expression

K, (0,) = (aOP, +PawPw) sec0, [dB]

where OZ is the zenith angle at which the an-tenna is pointing, PO (4 km) and V W (2 km) arecharacteristic heights for the atmosphericoxygen and water vapor, respectively, and a,and ~,w are the frequency-dependent attenua-

tion coefficients for molecular oxygen andwater vapor, respectively. The values of a,and Paw are available at some frequencies [ 941.However, it may be more accurate to measurethis effect than to calculate it [95] .

The correction factor K, is approximatelyunity when the angular diameter of the sourceis less than about one fifth the antenna half-power beamwidth. If the source is larger thanone fifth the antenna half-power beamwidth,K2 can be computed from the equation

where B (Sz) is the brightness distribution ofthe radio source, F, (52) is the normalizedpower pattern of the antenna, and s1, is thesolid angle subtended by the source. Curves ofthe correction factor K2 versus the ratio ofsource diameter to antenna half-power beam-width have been developed for several repre-sentative antenna power patterns [ 901 .

An examination of the equation for antennagain reveals that the gain measurement with theuse of celestial sources is basically a noise-poweror noise-temperature measurement. This typeof measurement requires the use of a radio-meter. The accuracy of the measurement isstrongly dependent upon the accuracy of theinstrumentation used.

A simplified block diagram of the instru-mentation is shown in Fig 64. The measure-ment of the antenna noise temperature due tothe source is accomplished by comparing thechange in output APs of the radiometer,which results from pointing tine antenna at thesource from a region in the vicinity of it, tothe change in output APK due to a knowntemperature change of a load connected inplace of the antenna. If ATK is the knownchange in temperature of the load, then theantenna temperature T, is given by

T, = ATK DpsAPK

104

AND DIRECTIVITYIEEE

Std 149-1979

HOT LOAD PRECISIONRF

ATTENUATOR POWERMETER

>LNA

COLD LOAD

PRECISIONRF

COLD LOAD

LNA = LOW N’OtSE AMPLiF E?

Fig 64Typical Instrumentation for the Measurement of Antenna Power Gain

and Temperature Using a Radio-Star Method

This result assumes that the radiometer islinear with respect to power.

To move the antenna from the sky to thesource, it is convenient to use the diurnalrotation of the earth. The measurement ismade by locking the antenna in a positionahead of the source and letting the earth’sdiurnal rotation move the source through theantenna’s electrical axis [ 58, pp 137-1561. Theadvantage of this procedure is that the back-ground emission received by the antenna’sside and back lobes remains unchanged duringthe measurement.

The change in load temperature ATK isobtained by switching from the calibrated coldload to the calibrated hot load. A& and A&can be measured with the use of the precisionradio-frequency attenuator shown in Fig 64.For example, APK can be determined by firstconnecting the cold load and adjusting theoutput to a convenient reference level. The

hot load is then substituted for the coldload. The radio-frequency attenuator is adjustedto restore the output to the original referencelevel. The change in attenuator setting givesAPK in decibels.

Once the antenna noise temperature dueto the source is measured, the gain can becomputed.

In addition to the correction factors K, ,which is due to the atmospheric absorption,and KZ, which is due to the relative angularsize of the source with respect to the antennabearnwidth, there are additional factors whichshould be considered if highly accurate mea-surements are required [96] . The most signi-ficant ones are:

(1) K,, the antenna pointing factor. If themain beam of the antenna does not passthrough the center of the source, the measure-ment will be in error. This factor includes theerror in alignment of the antenna’s axes.

(2) K4, the polarization factor (polarization

105

MEASUREMENT OF POWER GAINIEEEStd 149-1979

efficiency). If the source has a linearly polar-ized component, then the assumption that thepolarization efficiency is one half is in error.To determine this factor the polarization of thesource and the antenna shall be measured.

(3) K, , the system response factor. This isdue to the instabilities of the instrumentationand to the radiometer time constant whichmay result in a time delay of the recordedpower levels and therefore reduced ampli-tudes [ 58, pp 144-1461.

NOTE : See [ 961 for additional atmospheric effects.

If these factors are included, the gain of thetest antenna can be written as

8rkT,GT = sx2

___ K, K,K,K,Kj

12.4.3 Measurement of Absolute AntennaNoise Temperature and Figure of Merit G/T,In order to determine the absolute antennanoise temperature it is first necessary to mea-sure the effective noise temperature of thereceiver. This is accomplished by the Y-factortechnique which requires the use of calibratedhot and cold reference loads (see Fig. 64).The effective noise temperature of the receiverTR is given by [97].

TR=TH - Y, Tc

y -11

where TH and Tc are the temperatures of thehot and cold loads, respectively, and Y, isthe ratio of the noise power received with thehot load connected (TH + TR) to the n o i s epower received with the cold load connectediTc + TR).

Once Y, is determined, the hot load is re-placed by the test antenna. Its absolute noisetemperature T is given by

T = Y2Tc + TR (Yz - 1 )

where YZ is the ratio of the noise powerreceived with the antenna connected to thatwith the calibrated cold load connected.

The measurement of Y, and Y2 can be

accomplished by means of the precision radio-frequency attenuator shown in Fig 64 in thesame manner as outlined for the determinationof APK and APs (see 12.4.2).

Generally it is the figure of merit of theantenna-receiver system that is sought. Thesystem figure of merit is defined as the antennagain G divided by the system noise temperatureT referred to the antenna output terminals.The system noise temperature T is the sum ofthe antenna noise temperature and the effec-tive input noise temperature of the receiver[92]. The system G/T can be measured with aradio-star method [ 971, that is

G/T =8xk(Y-1) K K

SX? 1 2

where Y is the ratio of the noise power avail-able when the antenna is directed toward theradio star to that available when the antennais pointed toward the background sky at thesame elevation angle, K, is the correc t ionfactor for atmospheric attenuation, and K, isthe correction factor for the relative angularsize of the source with respect to the antenna’sbeamwidth. For high accuracy the additionalcorrection factors discussed in 12.4.2 applyhere also.

A typical procedure for the measurement isas follows. The antenna is pointed toward theradio star (either automatic or manual trackingis required). A peak response establishes areference level. The antenna tracking is stoppedand, as the radio star drifts away, the antennapoints toward the background sky. The radio-frequency attenuator is adjusted to restore theoriginal reference level. The change in attenu-ator setting is taken as the Y factor in decibels.

12.4.4 Measuremerrt of the Power Gain ofElectrically Large Antennas Using the Gain-Transfer Method. As discussed in 12.2 there isa practical limitation on the maximum gain ofa gain-standard antenna imposed by the avail-able techniques for accurate calibration. Atmicrowave frequencies the maximum gain isabout 30 dB. This gain is marginal for use witheven the strongest celestial sources. Therefore

106

AND DIRECTIVITY

the gain-transfer method in general is notrecommended when celestial sources are to beemployed. It is adequate if man-made beaconson satellites are employed. Highly accuratemeasurements can be performed for this case[851.

The procedure is basically the same as theone used on an antenna test range. Usuallythe gain standard is mounted on the largeantenna and oriented so that the beam axesof the two antennas are parallel. In some casesthe gain standard is permanentally installed fora continued system check.

The polarization mismatch between the waveradiated from the satellite and the test andgain-standard antennas can be a significantsource of error. To correct for this effect, thepolarizations of the wave and the antennasshall be measured. From these results thepolarization efficiencies for the wave-testantenna and the wave-standard gain antennasystems can be computed (see 11.1). Othersources of error include errors in tracking thesatellite, instabilities in the transmitting andreceiving systems, and receiver nonlinearityLB51 *

12.5 Errors in Power-Gain Measurements12.5.1 Gerreral. This section deals with the

estimation of the uncertainties that are en-countered when power-gain measurements ofantennas are performed. The term “uncer-tainty” is used here to mean the overall errorof the test parameters, that is, the sum of thesystematic errors and the random errors, asdistinguished from the various componenterrors.

Many of the sources of error associated withspecific techniques are discussed in the sectionsdescribing those techniques. There are sourcesof error, however, that most methods of power-gain measurement have in common. These willbe discussed in this section.

12.5.2 Sources of Error. Ideally the condi-tions required for power-gain measurementscan be catagorized as follows [83] :

(1) ilntenna range. Free-space conditions,uniform plane-wave field at the receivingantenna

IEEEStd 149-1979

(2) Antennas. Reciprocal, impedance match-ed, aligned and properly boresighted, in addi-tion the antennas shall be suitably polarizationmatched (see 11.1)

(3) Equipment operation. Components, idealand impedance matched, single sinusoidalfrequency, single waveguide mode, stablegenerator and receiver, adequate sensitivity anddynamic range

It is the deviations from these ideal condi-tions that produce uncertainties in the mea-surement of power gain.

In general, deviations from the conditionslisted in catagory (1) are the ones for which itis the most difficult to establish corrections.Those conditions imply an infinite separationbetween the transmitting and the receivingantennas, and an absence of multipath inter-ference, extraneous scattering of energy intothe receiving antenna, and reradiative mutualcoupling (see 4.2). Multipath interference,extraneous scattering, and reradiative mutualcoupling are always present. Their effects andthe methods of suppressing them are discussedin detail in 4.2.

The free-space conditions listed imply thatthe transmitting medium is reciprocal, lossless,linear, and isotropic. These conditions usuallycan be considered to prevail for antennaranges. (This is not true for the case of mea-surements employing radio stars or satellitebeacons as sources or for measurements onantenna ranges at millimeter wavelengths.)

A finite spacing between transmitting andreceiving antennas cannot be avoided. Theerror produced by finite spacing is of a sys-tematic type which results in a measured valueof gain lower than the actual far-field gain.This negative-valued systematic error decreasesasymptotically with increasing separation be-tween antennas. In order to estimate themagnitude of the correction factor, it is con-venient to think of the error as being caused bytwo effects. One is due to the fact that thegain of the test antenna does indeed vary withdistance due to the near-field components stillpresent, even though the spacing may begreater than 2D’/X, where D is the maximumdimension of the test antenna and X is the

107

-

IEEEStd 149-1979

wavelength. At 2D*/X, for a typical antenna,this correction factor is about 0.05 dB, and itwill decrease with increased spacing. Thesecond effect is that of a nonuniform illumina-tion of the test antenna due to the relativesizes of the transmitting and receiving antennasand their separation. A technique for estimat-ing the effect of nonuniform illumination ispresented in 4.2. For typical transmitting andreceiving antennas, for which the amplitudetaper across the aperture is about 0.25 dB ata spacing of 2D*/X, an error of about 0.1 dBcan be expected. This results in a total errorof 0.15 dB for a 2D*/h spacing. It should benoted that a reduction in amplitude taper ofthe illuminating field can only be achievedeither by increasing the spacing between thetwo antennas or by choosing a transmittingantenna with a broader beam. In either casethere is a danger of increasing multipathinterference. The effects of finite spacingbetween antennas has been studied extensivelyfor the special case of electromagnetic horns1131, 1821, [98141011~

Impedance and polarization mismatches canbe important sources of error in gain mea-surements. Corrections for these errors can bemade if measurements of the appropriateimpedances and polarizations of the antennascan be performed.

Impedance mismatches involve not just theantennas but also the generator and receiver(load). To illustrate this point, recall that thepower P, which appears in the Friis trans-mission formula is the power accepted by thetransmitting antenna. Usually one measuresthe available power from the generator PA byconjugately matching a power meter to thegenerator through a length of transmissionline. Alternately one can simply match thepower meter to the transmission line andmeasure what one might refer to as the line-matched power Pr,.l. The values PA and PMare equal only for the case where the genera-tor is matched to the transmission line. If thegenerator has a reflection coefficient rG look-ing into its output from an arbitrary referenceplane on the transmission line and the transmit-ting antenna has a reflection coefficient T‘T mea-


sured at the same reference plane, then thepower accepted by the antenna is given b yeither

P, = PAil/l,

or

p, = GM;

where

M, =(l-;rG I*) (I--jrT 1’)

?-rGrT /*

lu; =l-,rT I*

/brGrT j*

The form that is used is dependent upon whichpower was measured. If there is any significantloss between the reference plane and the an-tenna, that loss shall be taken into account.Likewise, if the test antenna has an inputreflection coefficient of rR and its load has areflection coefficient of I’,: then the powerinto the load is further reduced by a factorM2, where

lkf* =cl-,rR ;*I (1-IrL12)

1 &rRrL ] *

M 1 and M, can be evaluated by measuring rG,rT, rR, and rL. Often the load and generatorsare matched using tuners. In this case M,and M, reduce to

M, = (l+i*)

M, = (l-/rRj*)

These corrections can be made in the absolutegain method (see 12.2) and in the gain-transfermethod.

The equation for the measured gain of thetest antenna (C+)dB, using the gain-transfermethod, can be written to include theseeffects as

108

AND DIRECTIVITY

(GT)dB = (Gs)dB

where (Gs)dB is the gain of the standard indecibels , PT/Ps is the ratio of the powerreceived with the test antenna and the standardantenna, (MZ)T is the correction factor M,with the test antenna connected to the load,and (AJ2)s is the correction factor Al2 with thestandard connected to the load.

For absolute-gain measurements the correc-tions can be made in the same manner. How-ever, by treating the measurement as aninsertion-loss measurement, the correctionfactors for impedance mismatches are includedautomatically [83] . The procedure, in thiscase, is to connect the generator directly to theload and measure the absorbed power PL. Theantennas are then inserted at the proper spacing.By using a precision attenuator, the power levelat the load is adjusted to restore the originallevel. The ratio of the power Pe delivered withthe antennas inserted to the power Pt deliveredwhen the load is connected directly to thegenerator is obtained by noting the change inthe settings of the precision attenuator. Theproduct of the gains of the receiving andtransmitting antennas is given by

where

If the transmitting and receiving antennas arenot polarization matched, then the Friis trans-mission formula, as given in 12.2, shall bemultiplied by the polarization efficiency p(see 11.1). Thus there is an apparent error inthe measured gain unless the factor p is knownand the measured gain is corrected. The effectof polarization mismatch can be illustratedwith the following examples. First consider

IEEEStd 149-1979

Table 2Errors in the Measured Gain of a Purely Circularly

Polarized Antenna Due to a Finite Axial Ratioof the Transmitting Antenna

TransmittingAntenna

Axial Ratio(dB)

20253035‘404550

Measurement Error [ dB ]

Same Sense Opposite Sense

+0.828 -0.915+0.475 -0 .503+0.270 -0 .279+0.153 -0 .156+0.086 -0 .109+0.049 -0 .049+0.027 -0 .028

the case where the gain of a circularly polarizedtest antenna is measured using the method ofpartial gains (see 12.3.2). Suppose that the testantenna is purely circularly polarized, and thatthe gain standard is purely linearly polarized,but that the transmitting antenna has a finiteaxial ratio. Then Table 2 indicates the errorsone might expect due to the finite axial ratioof the transmitting antenna.

Next consider the case of the measurementof a linearly polarized antenna with an axialratio of 25 dB using an ideal gain standard(linearly polarized) and a transmitting antennahaving a finite axial ratio. The errors expectedfor several different axial ratios are given inTable 3.

TabIe 3Errors in the Measured Gain of a Linearly

Polarized Antenna Due to a Finite Axial Ratioof the Transmitting Antenna*

TransmittingAntenna Measurement Error [ dB]

Axial Ratio(dB) Same Sense Opposite Sense

20 -0 .035 +0.06325 -0 .014 +0.04130 -0 .002 +0.00335 +0.005 +0.02240 +0.009 +0.01945 +O.Oll +0.01650 +0.012 +0.015

*The test antenna has an axial ratio of 25 dB. The gainstandard is purely linearly polarized.

109

IEEEStd 139-19’i9

For the preceding examples the errors dueto polarization mismatch can be significant,especially when a linearly polarized gainstandard is used to measure circularly polarizedantennas by means of the partial-gain method.The effect is less severe when a nominallylinearly polarized antenna is being measured.As a point of reference, a typical electromag-netic-horn gain standard is likely to have anaxial ratio between 40 dB and approachinginfinity. If corrections for these effects are tobe made, the polarizations of all antennas usedshall be measured. In an automated systemwith a minicomputer, the computations couldbe made a part of the computer software.

12.5.3 Estimation of Uncertainty in GainMeasurements. The determination of the gainof an antenna involves the measurement of anumber of quantities which are used to computethe gain. For example, consider the insertion-loss method of absolute gain measurementdiscussed 12.5.1 The gain-product equationcan be written as

where f is the frequency in hertz, c is the free-space velocity of light, cy is the measured powerratio, and M is the correction factor for imped-ance mismatch. To determine GTGR onehas to measure the length of the range R, thefrequency f, the power ratio cr, and the imped-ances, so that the correction factor M may becalculated. There will be random and sys-tematic errors associated with each of theseindividual measurements. In addition to thesequantities, which appear explicitly in the gainequation, there are other factors that affectthe accuracy of the gain such as polarizationmismatch, the effect of the amplitude taper ofthe incident field over the aperture of thereceiving antenna, multipath effects due torange reflections, and system errors such asalignment of the antennas, equipment in-stabilities, and so on. One can account forthese effects by multiplying the gain equationby correction factors corresponding to eachof the preceding as follows:


CYMK, K, K, K,

For the case of nearly identical antennas(where GT = GR), the gain uncertainty of theantennas can be shown to be

-=_

This result assumes that the individual uncer-tainties associated with each term are indepen-dent. Unfortunately the signs of the individualuncertainties are not known; however, theprobability that all signs are the same is quitelow. Thus the root-sum-square method ofcomputation is usually used to determine theuncertainty, that is

‘c’=[(~)2+(!x)2+(c!5)2+($!)

+(q2 +(g$2

+($)‘+ (-,;I 1’2Often both the arithmetic sum and the root-sum-square values of uncertainty are reported.

The limits of the 99 percent confidenceinterval are often taken to be the limits of theuncertainity which, for a single measurement,corresponds to + 30, where u is the standarddeviation. The standard deviation for the gainmeasurement can be obtained from the standarddeviations of the individual measurements,which make up the gain measurements, by theuse of the root-sum-square computation[991.12.6plete

Directivity Measurements. When the com-radiation pattern of an antenna is known

110

AND DIRECTIVITY

or is measurable, it may be used to determinethe directivity of the antenna, The particularquality of the pattern that is employed is theradiation intensity cP(fl, 4), which is the powerradiated from the antenna per unit solid anglein a given direction. The peak directivity of anantenna is the maximum radiation intensity@, (0 ‘, C#J ‘), divided by the average radiationintensity. The latter quantity multiplied by 47ris the total power radiated. To compute thedirectivity D, , the following relation maybe employed :

~rn(cJo = Qrn(d ‘7 44 =p ,4n

t

‘a,(6 ‘7 4’)sin

where Pt is the total power radiated by the testantenna, and the angles 19’ and 4’ give the direc-tion for which the radiation indensity is amaximum.

Since the radiation intensity cP(0, @) can becontained in any two orthogonal polarizations,it is convenient to write the expression fordirectivity as

Drn VUJ’) = ‘km1 UC 9’)-~(1/4n)(Pt1 + Pt2)

cp+ m2 (0’~ 4’)(1/4n)(Pt1 + 42)

= D, (6’, @‘) + Dz (6’, q5’)

where the subscripts indicate the two orthog-onal polarizations. D, (0’, 4’) and D, (e’, o’)are called partial directivities. Note that thetotal power radiated in both polarizations isrequired in order to determine either D, (0’,+‘) or D2 (e’, 4’). Usually the orthogonalpolarizations employed are 0 and o linear orright and left circular.

When the antenna under test has been de-

IEEEStd 149-1979

signed to be essentially linearly polarized, thentl and 4 linear polarizations are most oftenused. For this case it is desirable that the testantenna be oriented such that the designpolarization corresponds with either the (3 orthe @ polarization. If, for example, the antennais designed to have 0 polarization, the 4 polari-zation is referred to as the cross polarization.De (0’, c#J’-) and Do (0; G’) are the partial direc-tivities with respect to 0 and @ polarization,respectively. Often it is the partial directivitywith respect to the design polarization that issought. Note that if right and left circularpolarizations were employed, the directivitycould be obtained, but the partial directivitywith respect to the design poIarization wouldnot be obtained.

For test antennas that have been designed tobe essentially circularly polarized, it may bedesirable to use right and left circular polariza-tions. In this case, if the maximum coupling isobtained with the use of, say, right circularpolarization, then left circular is the crosspolarization. Note that if 0 and o linear polari-zations were employed, the partial directivitywith respect to the design polarization wouldnot be obtained.

For either of the preceding cases the mannerin which the test antenna is used operationallywill usually dictate which orthogonal polariza-tions should be used. For example, if a circularpolarized test antenna is to be used opera-tionally to receive a known linearly polarizedsignal, then appropriate orthogonal linearpolarizations may be desirable. Additionally,practical considerations may dictate whichapproach will be taken. For example, it isusually more difficult to produce a circularlypolarized field than a linearly polarized onewith sufficient purity to make accurate mea-surements.

For conciseness in the following discussion,the equation for directivity will be treated asthough all the radiated power were containedin a single polarization and that polarization isused for the measurement, that is, the crosspolarization is taken to be zero, the discussionapplies equally well to both partial directivities.

111

IEEEStd 149-1979 DETERMINATION OF

The radiation intensity may be measured bysampling the field over a sphere centered onthe test antenna (see Section 3). This can beaccomplished by making conical cuts, succes-sive C$ cuts through the pattern at incrementsof 8, or by great-circle cuts, successive 0 cutsat increments of 4. For either case, the numberof increments required is determined by thecomplexity of the antenna’s pattern structure.In general the number of increments increasesas the pattern becomes less uniform. Also, anantenna with a narrow major lobe is likely tohave a large number of narrow minor lobeswhich must be included in the measurementand subsequent integration. Generally it ispractical to determine directivity accuratelyonly for those antennas having antenna patternsthat are not highly directive.

A normalized value of the radiation intensityis usually recorded instead of the absolutevalue. Typically the normalization is withrespect to the maximum value, and the peakdirectivity may be expressed as

sinode

where

If o cuts are employed, then the 8 intervalfrom 0 to 7r rad is divided into M equal spher-ical sectors, and the expression for directivitybecomes

o,(e’,o’) = M4M

S Cl4 sin eii=l 1

where Bi = ir/M rad.If the normalized radiation intensity is

recorded in rectangular coordinates, then eachof the integrals within the brackets is equal tothe area under the curve, and a planimeter, forexample, can be used to evaluate them. Whenpolar coordinates are used, it is necessary torecord voltage rather than power because the

area of a differential wedge in polar coordinates

is given by + V* de, where V is the voltage

recorded at the angle 4. Since the radiationintensity is proportional to V* , the area in thepolar curve will be proportional to the integralwithin the brackets.

If 0 cuts are employed, the o interval from 0to 2n is divided into N equal increments, andthe expression for peak directivity can bewritten as

Qn (K 6) = N2 N

Cj=l

i,?* (0, @ji, sin 0 de

Note that s(e, @j) shall be multiplied by sin 19during the process of integration. If it is neces-sary to use graphical integration, then a specialsine graph paper can be used.

The directivity can be evaluated numericallyby use of a digital computer. The patterns canbe recorded digitally and can be entered intoa digital computer at a later time. If an on-linecomputer is available, real-time computationcan be made. As discussed in 5.7, an on-linecomputer offers the possibility of automatedmeasurements.

13. Determination of Radiation Efficiency

The radiation efficiency of an antenna isthe ratio of the total power radiated by theantenna to the net power accepted by theantenna at its terminals during the radiationprocess. The difference between these twopowers is the power that is dissipated withinthe antenna. Radiation efficiency is an inherentproperty of an antenna, and is not dependenton system factors such as those due to imped-ance or polarization mismatch.

A fundamental method for determiningradiation efficiency relies on the measurementsdescribed in 12.2 and 12.3. As noted in 12.1,radiation efficiency is equal to the ratio of the

112

RADIATION EFFICIENCY

power gain in any specified direction to thedirectivity in that same direction. It is us-ually convenient to take the direction ofmaximum radiation for this determination ofradiation efficiency. Thus

peak gainradiation efficiency = ~

peak direc tivity

In measuring peak gain and directivity, all theprecautions mentioned in 12.2, and 12.6 shallbe carefully observed. Even when this is done,the results are not very accurate for low-losshighly-directive antennas because of the dif-ficulty in calculating directivity from themeasured patterns with sufficient accuracy.

Another method may be available when theantenna is electrically small and simple. In thiscase an equivalent series circuit can frequentlybe found in which the real part of the inputimpedance, that is, the antenna resistance, isequal to the sum of the radiation resistance anda loss resistance [ 1021 .

The radiation resistance accounts for allradiated power, and the loss resistance accountsfor all dissipation within the antenna. Forantennas such as dipoles and loops, where thetheoretical pattern can be integrated, theradiation resistance is best found by calcula-tion from the dimensions [73, pp 136, 143-148,166-1691, [ 103 section 111. The antennaresistance is obtained from measurements ofinput impedance (see 16.1). The radiation ef-ficiency is then

radiation efficiency =radiation resistance

antenna resistance

This method is valid only if the antenna can beaccurately represented as a series circuit. Whenthe dissipation cannot be represented by aresistance in series with the radiation resistance,as in the case of an antenna coated with lossydielectric or an antenna over a lossy ground,the method should not be used. Furthermore,the calculated radiation resistance and themeasured antenna resistance shall be referredto the same set of antenna terminals. It should

IEEEStd 149-1979

also be noted that the input impedance of thistype of antenna may present a large mismatchto the connecting transmission line, and amatching network having appreciable dissipa-tion might be used. Such a loss is not usuallyincluded within the meaning of radiationefficiency, although it is clear that the losswould be important to the system as a whole.

14. Special Measurements forAngle-Tracking Antennas

14.1 General. In directive antennas the direc-tion of the beam or the tracking axis often hasto be determined precisely on the basis of anelectrical indication from the antenna system.Such a direction is called the electrical bore-sight. This electrical boresight is determinedwith respect to a reference direction, calledthe reference boresight. The latter is either aspecified stationary direction or a directionderived from a measurable parameter, such asan optical axis or a mechanical axis of sym-metry, or a prior electrical indication. Theangular deviation of the electrical boresight ofan antenna from its reference boresight iscalled the boresight error and is a measuredquantity. On the basis of such measurement,the electrical boresight of the antenna system isaligned with the axis of symmetry or the axrabout which the antenna rotates, and the bore-sight error is minimized.

The beam direction of an antenna with asingle major lobe is usually determined bynoting the direction of maximum response orthe direction halfway between equal responseseither side of the peak. The precision of suchdetermination is usually on the order of one+znth the half-power beamwidth.

Greater precision of direction is demandedof tracking antennas that employ two or moreoverlapping beams or lobes as indicated inFig 65(a). These lobes are compared to indi-cate a direction through equality of ampli-tude or phase. When the comparison is donesequentially, it is termed sequential lobing,and typically amplitudes are compared. Twoexamples of this class are conical scanning

113

IEEEStd 149-1979 SPECIAL MEASUREMENTS FOR

LOBE 2, OR LOBE IAT A LATER TIME

CROSSOVER

0 ANGLE

(0)

LOBE I-LOBE 2(DIFFERENCE)

DIFFERENCE-PATTERNM I N I M U MOR NULL 0 ANGLE

(b)

Fig 65Signals Received by Tracking Antenna Versus Angle of Target.

(a) Lobe Patterns, Monopulse, or Conical Scanning.(b) &Ionopulse Sum and Difference Patterns

[104] and lobe switching. When the com-parison is done simultaneously, it is calledsimultaneous lobing or monopulse. In thiscase either amplitude or phase may be com-pared [105]. The electrical boresight withsuch antennas can be determined with a preci-sion on the order of one hundredth of thebeamwidth, depending upon the availablesignal-to-noise ratio.14.2 Conical Scanning Angle Tracking. Conicalscanning antennas use a single beam, the peaktraversing a nearly circular path, so that thebeam axis describes a cone. Fig. 65(a) representsa cross section through the pattern showingtwo extreme positions of the beam on op-posite sides of its circular path.

During the scan cycle the radio-frequencyvoltage received or transmitted by the antennais amplitude modulated because the signalvaries with the beam position when the targetis not on the crossover axis, which is deter-mined by equal voltages for the two lobes asin Fig 65(a). The amplitude modulationenvelope is a complex signal in which thelowest observed frequency, or fundamental,is the scan frequency [106]. The percentmodulation of the radio-frequency voltage bythe fundamental is of importance in tracking

systems. Harmonics of the fundamental scanrate arise from the inherent nonlinearity ofthe scanning process. At crossover the har-monics ideally go to zero as does. the funda-mental. In practice, however, antenna patternimperfections, such as unequal E - and H-planebeamwidths, leave a residual harmonic signal.

The fundamental and higher harmonic modu-lations are often measured as a function oftarget error, which is the target displacementfrom the boresight direction. The null of thefundamental modulation (or error signal) de-fines the electrical boresight. The slope andlinearity of the error signal with respect totarget error, as well as the angular range overwhich the error signd maintains unique polarityare important to the tracking system design.The uniqueness of polarity determines theextent of the angular region over which theantenna system can acquire a target. Theharmonic content of the modulation envelopeis important to the demodulator design.

These modulation properties may be mea-sured by a system which employs one-way

/

transmission as in a normal pattern measure-ment, but which utilizes the square-law char-acteristic of a bolometer to simulate the two-way transmission of normal radar angle track-

114

ANGLE-TRACKING ANTENNAS

ing. The equipment required consists of amicrowave transmitter, a bolometer detectorand mount, an amplifier, and a wave analyzer.The transmitter, located in the far field andoriented to simulate the target position, maybe a square-wave-modulated klystron or a pulsedmagnetron which is amplitude modulated ata suitable repetition rate. The antenna undertest is terminated with the bolometer mountfrom which the detected signal is fed to theamplifier. For the measurement of the funda-mental modulation, the bolometer-amplifiercombination shall have an essentially flatresponse for the frequencies that lie betweenthe repetition rate plus or minus the conicalscan frequency. The fundamental of the ampli-fied signal and the first-order sidebands, whichare at the repetition rate plus or minus thescan frequency, are compared using a waveanalyzer to determine the percent modulationwhich is defined as

lower sideband amplitude

% modulation =+ upper sideband amplitude

amplitude at repetition rate

x 100%

The upper and lower sidebands are theoreticallyequal. In practice, however, they may beslightly different due to the measurementerrors and are averaged in the preceding equa-tion.

Before the system measurements are at-tempted, a check for linearity shall be made.This can be done by inserting a radio-frequencyattenuator ahead of the bolometer, and notingthe relative output voltage displayed on thewave analyzer for a given change in attenua-tion. Over a range of at least 20 dB of inputlevel, the output voltage variation in decibelsshall be double the change in the radio-fre-quency attenuation.

I Because the angle sensitivity of the two-way

I

fundamental modulation is not easily mea-sured, it may be determined, to a good approxi-mation at least in the region of crossover (wherethe harmonic content of the modulation waveform is not large), by the following relation:

IEEEStd 119-1979

70 modulation = 2 log, 10 X v?%- x 100%

degree 5 P

1.6 fi= x 100%P

where D is the crossover depth in decibels, pis the 3 dB beamwidth in degrees, and %modulation is two-way percent fundamentalmodulation.

The angle sensitivity can also be expressedin terms of a dB difference curve which is thealgebraic difference between two azimuthpatterns at the extreme scan positions as shownin Fig 65(a), with the data being measured indecibels.

70 modulationdB difference 2 (20 log,, e) X

100%

% modulation= 8.686 x

100%

14.3 Monopulse Angle Tracking. Measurementsof monopulse tracking antennas are made atthe sum and difference ports. The difference-signal pattern minimum defines the electricalboresight (see Fig 65(b)). There are additionalerrors contributed by the interaction betweenthe antenna and the associated circuits in thetracking loop. These can be determined fromadditional measurements and from a graphicalcomputation [ 1071. The angle-sensitivity infor-mation usually desired is the slope of the lobe-signal patterns at crossover or the asymptoticslope of the difference-signal patterns. Duringmeasurement these slopes should be referred toa reference level of signal voltage, such as asignal from an isotropic radiator or some otherappropriate reference [log], to permit an eval-uation of the antenna design and the systemsignal-to-noise ratio.14.4 Electrical Boresight Measurements. Thetest equipment for determining the electricalboresight of an antenna comprises the antennaunder test, the associated circuits for perform-ing the appropriate signal processing to obtainan error signal, a distant source, a precisemeans for orienting the antenna direction

115

IEEEStd 149-1979

(or moving the source), and a highly accurateoptical or mechanical indication of antennadirection (or source location). The antenna(or source) is oriented for minimum outputat the appropriate error-signal port of thecircuit. This antenna direction (or sourcelocation) is then noted under the particularcondition of frequency, environment, or othertest variable. The direction may be comparedwith a reference direction to determine a com-ponent of boresight error. If the antenna sys-tem output cannot be noted continuously as afunction of source angle, the minimum pointmay be interpolated. Since the absolute direc-tion of the electrical boresight during a test isoften of less interest than its variation with thesystem parameters, the chief requirement of theantenna direction-indicating mechanism maybe one of high precision over a very smallangular range. A telescope rigidly fixed to theantenna mount may be used to sight a cali-brated optical target at the distant source, ora dial indicator on a long-radius arm may beused. A camera or laser may be similarly em-ployed to measure dynamic errors betweenelectrical and mechanical axes when tracking amoving target,

15. Measurement of the ElectricalProperties of Radomes

15.1 General. For many applications opera-tional antennas have to be protected from theeffects of the environment. These protectingstructures, called radomes, are generally de-signed to be nearly transparent to the electro-magnetic radiation and occasionally to affectradiation in some desired way. Most radomesconsist of dielectric materials, although somecontain significant amounts of conductingmaterials. Metallic space frames have beenutilized to increase the strength of large ra-domes, and perforated metallic layers havebeen developed for both flat and curved ra-domes. Radomes may take the shape of hollowshells or flat sheets. The hollow shells, in theform of blunt curved radomes, are frequentlyused with shipborne antennas and occasionally

MEASUREMENT OF THE ELECTRICAL

on aircraft. Pointed shapes such as ogives areused on the nose of aircraft and missiles. Flatsheets cover the radiating elements of electroni-cally scanned arrays and also contribute tothe antenna impedance match.

This section describes the measurement ofthe effect or perturbations introduced by theradomes on the performance of the antennasthey enclose. For realistic tests to be per-formed, ancillary parts such as pitot tubes andlightning rods found on some radomes shouldbe included. Multiple scattering betweenmultiple antennas in one radome can beappreciable, thus all antennas shall be presentand appropriately terminated during the tests.Because the effects of the radome on the an-tenna occur in the near field, the evaluation ofthe radome shall be performed in the presenceof the antenna’s auxiliary structures as well asof the actual antenna.

Radome testing is an extension of antennatesting because the tests are performed onantennas with and without the radome. Theapparatus and tests are more extensive, espe-cially when a mechanically or an electronicallyscanned antenna is involved.

In the following discussion the quantitiesthat define the significant electrical parametersof the radome are delineated, and the apparatusand the procedures to be used are described.

The coordinate systems used for radometesting are usually the same as those describedin 3.1. Two sets of coordinates are used, one isfixed to the antenna to describe the pattern,and a second set is fixed to the radome todescribe antenna orientation. Special coordi-nates are sometimes used for airborne guidancesystems.

15.2 Significant Antenna-Radome Parameters.The performance of a radome is characterizedby the power transmittance, the power reflec-tance, the squint angle, the boresight error, andthe radiation patterns in the presence of theradomes.

The power transmittance IS the ratio of thepower density in a given direction emergingfrom the radome with an internal radiatingsource to the power density radiated by thesource without the radome. Power trans-

116

PROPERTIES OF RADOMES

mittance is usually measured in the direction ofthe peak of the main lobe. Alternately it maybe specified in terms of the crossover level fora conical-scan antenna, or the null direction fora monopulse system. The positions and orienta-tions of both antenna and radome shall bespecified in measurements of power trans-mittance.

The power reflectance is the ratio of thepower density that is internally reflected fromthe radome to that incident on the radomefrom the internal radiating source. It indicatesthe contribution of the radome to the antennamismatch.

The squint angle (beam shift) is the angulardisplacement of the main lobe of the sumchannel patterns from a reference beam direc-tion. The squint angle often is treated as avector quantity with two components that areusually resolved into the directions of thecoordinates specifying the antenna-radomeorientation.

Boresight error describes the angular dis-placement from the zero error direction in atracking system which uses an antenna systemthat generates a conical scanning, lobing, ormonopulse difference pattern. The boresighterror can be considered to be a vector withtwo components resolved into the directions ofthe coordinate system that specifies antenna-radome orientations.

The antenna patterns in the presence of theradome change most significantly in the side-lobe regions. The side lobe levels may be ex-pressed in terms of peak or mean levels, depend-ing on the application. For all the quantitiesdiscussed, polarization shall be specified andtaken into account in the measurements.

In performing radome measurements, all therelevant quantities shall be measured for asufficiently wide range of coverage to allowthe significant effects of the radome to beobtained. For example, antenna patterns shouldbe measured over the full hemisphere to locatewide-angle side lobes caused by reflectionsfrom the radome. Similarly, patterns should berecorded for wide ranges of antenna position.However, the scope of all tests, including data

IEEEStd 149-1979

reduction and evaluation, shall be carefullyplanned because of the costs involved. Thus toreduce the total number of variables, antennapatterns are measured with the radome andenclosed antenna in a fixed relative orientation,while both are rotated as a unit. Measurementsare then repeated for several relative orienta-tions of the antenna radome. Specializedautomated equipment and computerization areavailable that gather and reduce the data ef-ficiently.

15.3 Apparatus. The apparatus to be used andthe procedures to be employed are outlined forthe quantities enumerated in the precedingparagraphs. In many instances the degrees offreedom of the antenna with respect to the ra-dome and exterior region are complex; fixturesshall be adequate to duplicate the angular reia-tionships in the actual system. The fixtures aredevices that hold and position the antenna tosimulate actual system operation. The fixtureshall also duplicate the sequence of axis rota-tions to be found in the actual system [log].The fixtures themselves shall not add extrane-ous reflections. A check can be made by takingsuccessive measurements with the antenna(without the radome) displaced in steps of frac-tions of a wavelength so that reflections addand subtract. Thus the magnitude of the stand-ing wave will show up in the several quantitiesto be measured. Side-lobe levels are especiallysensitive to the antenna motion. Particularattention shall be paid to the overall imple-mentation of the measurements to minimizespurious range reflections and to illuminate thetest antenna and radome with a plane wave.The far field is defined in terms of the maxi-mum dimension of the overall structure, in-cluding the radome and the nearby portions ofthe vehicle to which it may be attached. Tocheck for reflections within the radome, mea-surements can be repeated with successivequarter-wave shifts of the antenna relative tothe radome in the direction of the antennaaxis. These displacements reveal reflectionsthat can affect the measured values of trans-

117

MEASUREMENTIEEEStd 149-1979

Fig 66Orthogonal Arrays for Beam-Shift Measurement

(D-detector, other components omitted)

mittance, reflectance, and side-lobe levels. Pre-cautions shall also be employed in the designof the transmitting and receiving systems toassure sufficient amplitude and frequencystability, particularly in measurements thatinvolve phase.

Power transmittance and beam-shift, or ra-dome boresight error, measurements can beperformed simultaneously. For these tests theradome is pivoted about the enclosed testantenna, which remains stationary except pos-sibly for small displacements that permit mea-surement of the radome boresight error.

Two techniques are used for automaticmeasurements of radome boresight errors. Inthe first technique an external antenna (calleda null seeker) is automatically positioned totrack the direction of the null as the radome ispivoted. In the second technique, called theclosed-loop test, the antenna being testedis moved to maintain a null as the radome ispivoted.

For boresight measurements with antennasthat have a null tracking system, the usualpractice is to operate the test antenna in recep-tion, because many systems operate only inthe receiving mode. The boresight error can beobtained by rotating the test antenna topreserve the antenna null that exists withoutthe radome. Alternately the external antennacan be laterally displaced to maintain the nullof the test antenna. Precision in angular mea-

surements is greater for the latter case becausethe displacement is the product of the rangebetween transmit and receive and the boresighterror. For large antennas that require longranges, the necessary displacement may be solarge that moving the external antenna is im-practical. With either technique, the angulardeviations are obtained from the error voltagesgenerated by the servo mechanism of the sys-tem. Different measurement setups are re-quired with the two techniques.

For boresight measurements of antennaswithout tracking nulls, the test antenna is usedin transmission. Reception is obtained byusing two orthogonally oriented pairs of an-tennas as shown in Fig. 66. The plane contain-ing each pair is perpendicular to the electricalboresight. Each pair is interconnected so thatthe amplitude or the phase of the received sig-nals may be compared. A balanced output fromeach pair is initially established without the ra-dome. In some cases the test antenna is fixed,and the receiving antennas are laterally displacedto preserve the balanced output while the ra-dome is pivoted. In other cases the receivingantennas are fixed and error voltages are derivedfrom the test antenna. Alternately, the test an-tenna can be displaced to preserve the balancedcondition with the fixed receiving antennas.

Power transmittance measurements are madewith additional apparatus. When the beam shiftis measured, a fifth antenna is used. It is located

118

OF IMPEDANCES

at the intersection of the two base lines thatconnect the beam-tracking antennas (see Fig66). Received power is measured while theradome is pivoted and the beam is tracked.The received power is compared with the valuethat is received when the radome is removed.When the boresight error is measured with theradome-enclosed test antenna in the receivingmode, the power is measured at a port of theenclosed antenna. For monopulse antennas,power is extracted from the sum port. Forconical-scan antennas, a directional couplersamples the power ahead of the detector thatis used to define the boresight error.

Reflectance is measured when the enclosedantenna is in the transmitting mode. The ra-dome is pivoted and the enclosed antenna isstationary.

Very large radomes and the antennas theyenclose can only be measured when they arelocated in their operational environments.Preliminary tests may be performed using partsof the radome in the form of flat sheets. Thesame quantities are measured and then extrapo-lated to the complete structure. The measure-ments for large radomes involve the extensionof techniques used for the in situ measurementsdiscussed in Section 9. The environment inwhich the antenna and radome must operateshall be taken into account. These environ-mental conditions include rain, snow, wind etc.The effects of rain are discussed separately inthe next subsection.15.4 Testing of Wet Radomes. A weatheredradome, especially one covered with dirt orone that absorbs water, will show appreciablechanges in use. Rain or sea spray can cover aradome with a thin layer of water that canproduce significant effects. The water reducestransmittance and causes reflections that canincrease side-lobe levels. In communicationlinks the noise temperature can also beincreased.

The effects of water can be tested by spray-ing the radome during the previously describedtests of transmittance, patterns, and so on.However, testing with water present is compli-cated because accurate measurements of waterthickness and distribution are difficult. Capaci-tance measurements can be utilized to deter-

IEEEStd 149-1979

mine the thickness in controlled measurements,but the sensors can perturb the propagation. Inaddition, because of surface roughness, watercan accumulate locally so that the contour ofthe radome should be checked. The significanceof test results should be assessed in terms ofthe amount of. water than can accumulatewhen the radome is being used in its operationalenvironment, for example, rainfall rates shouldbe measured. If salt water is of interest, thesalinity of the water used in the tests shall bedetermined.

16. Measurement of Impedances

16.1 Input-impedance Measurements. The in-put impedance of an antenna at the specifiedterminal pair (or port) affects the interactionbetween the antenna and its associated circuits.Antenna impedance can be an important factorin the consideration of power transfer, noise,and stability of active circuit components.Frequently it is the antenna impedance thatlimits the useful bandwidth of the antenna.

The optimum impedance relationship betweenan antenna and its associated circuits is deter-mined by the application. In some receivingsituations, in the interest of a minimum noisefigure, the antenna impedance should be lowerthan the matched impedance. In some trans-mitting situations, in order to attain maximumpower efficiency, a mismatch in the oppositedirection may be required. However, in manyapplications a matched condition is the ideal.By this is meant a conjugate match betweenthe antenna and the circuits; maximum powertransfer is attained in this case. When a conju-gate match does not exist, some of the avail-able power is lost, as follows:

where Zant is the input impedance of the an-tenna, Z&t is the input impedance of thecircuits at the antenna terminals, and Z*,,t isthe complex conjugate of the circuit imped-

119

IEEEStd 149-1979

ante. This expression may be useful duringthe measurement of power gain with a mis-matched system, as mentioned in 12.5.

Most antennas are connected to the electronicnetworks via a transmission line, and the de-sired degree of match or mismatch could beadjusted at either end of the line. In practice,however, it is usually advantageous to performthe matching as near to the antenna terminalsas possible. This minimizes the line losses andthe voltage peaks on the line, and generallymaximizes the useful bandwidth of the system.Imperfect matching of the antenna to the trans-mission line creates a reflected wave in the line;the reflected power relative to the incidentpower is:

Prefl zant -2, 2---=Pint I IZant + 20

where 2, is the characteristic impedance ofthe line. This ratio is related to the voltagereflection coefficient r and to the voltage-standing-wave ratio (VSWR) by the standardtransmission-line relationships

If the transmission line has a characteristicimpedance that is purely real, and if the elec-tronic networks are perfectly matched to thetransmission line, this ratio yields the loss ofavailable power caused by the reflection at theantenna terminals; if not, this loss shall be de-termined using methods given in ANSI/IEEEStd 148-1959 (Reaff 1971), Measurement ofWaveguides and Components. (See also 12.5.)

Measurement of the input impedance ismade at a single port of the antenna. For theusual problems and procedures related to thismeasurement, reference should be made toANSI/IEEE Std 148-1959 (Reaff 1971). Thereis, however, a particular problem inherent inradiating structures, since the input impedanceis modified by the environment of the antenna.

MEASUREMENT

For this reason the antenna shall be placed in asimulation of its operating environment beforethe measurement is made. Usually this require-ment is easily approximated for narrow-beamantennas that can be pointed away from re-flecting obstacles, but it may be more difficultfor omnidirectional antennas where much ofthe surrounding structure affects the inputimpedance.

The measurement of impedance, or equiva-lently, of the complex reflection coefficent,can be performed using impedance bridges orslotted lines at frequencies where these techni-ques are applicable [110], [ill], [154]. Theincreasing availability of broad-band, swept-frequency network analyzer systems thatmeasure impedance or the entire networkscattering matrix has made measurements ofthis type the choice for many applications.These instruments are commercially availableat frequencies up to and through the microwavespectrum and are suited for computer-controlleddata collection, storage, and display, as well asfor analog display using X-Y recorders andoptical display units. At microwave frequenciesthe units feature automat‘ic swept-frequencydisplay of complex reflection and transmissioncoefficients with various overlay charts to relatethe reflection coefficient locus to impedancecoordinates. Two such coordinate systems arein common use. One involves the resistive andreactive components of the impedance [112],while the other involves the magnitude andphase of the impedance [ 1131. The impedanceso determined is in a form which is normalizedto that of the reference transmission line. Thisis usually the convenient form at microwavefrequencies where transmission lines arecommon.

16.2 Mutual-Impedance Measurement. In anarray antenna, there is usually an interactionbetween the array elements which significantlyaffects the behavior of the antenna, and it isoften necessary to determine the effect of thisinteraction. An important measure of the inter-action is the mutual impedance between theelements of the array antenna.

The mutual impedance between any two ele-

120

OF IMPEDANCES

ments is defined by the equation

z vm= -mn

In

where In in the current at the reference pointof the driven element 12, and V, is the voltageproduced at the reference point in element M,when all the elements except the driven ele-ment are open-circuited at their referencepoints.

The mutual coupling between two elementscan also be described in terms of the incident,reflected, and coupled waves measured in thetransmission lines connected to each element.This description of coupling utilizes the cross-coupling coefficient of a scattering matrix:

Sbm=-

mnan

where a , is the incident wave at a referencepoint in the transmission line of the drivenelement n, and b, is the received wave in thetransmission line of element m, when all otherarray elements are terminated in matchedloads.

These measures of mutual impedance orcoupling depend upon the type of elements,their size and spacing in terms of a wavelength,and the geometrical arrangement of the ele-ments and their environment. In general themagnitude and the phase angle of the mutualimpedance or coupling decreases with increasedspacing.

The reference terminals (or ports) at whichthe currents and voltages are determined maybe selected for greatest convenience, but in anycase they shall be specified. In tower radiators,which are the elements of a broadcast array an-tenna, the reference terminals are customarilytaken at the base of the towers, even thoughthe towers may be of unequal heights, and theeffect of the base insulators is often subtractedfrom the measured values. In dipole arrays,such as those used at high frequency and veryhigh frequency, the reference terminals aregenerally taken to be at the current maximumof the radiating portion of the elements. In suchcases the current distributions themselves are

IEEEStd 149-1979

sometimes measured using loop probes [114]mounted within slits in the dipole or along thedipole surface. In microwave arrays the refer-ence points are often taken to be at the portswhere the waveguides connect to the elementsor at the array face for waveguide elements. Forcases in which the element feed transmissionlines are of uncertain or unequal lengths, it isconvenient to use symmetric probe techniquesto assure a consistent measurement of thereference plane [114], [115].

An impedance or scattering matrix can bemade up of a set of mutual, impedance valuesZmn or a set of scattering coefficients S,, fora given antenna array. These matrices includethe self-impedances Z,, or the complex reflec-tion coefficients S,, . These two types ofmatrices are interrelated and can be convertedfrom one to the other by means of a matrixtransformation [ 1161 . The self-impedance orreflection coefficient terms are properties ofeach element individually excited in its arrayenvironment for specific loading conditions,and differ from the active impedance or reflec-tion coefficient that exists when the entirearray is excited. The quantities measured atthe terminals under the latter condition areinfluenced by mutually coupled contributionsthroughout the array [117]. The active impe-dance, or active reflection coefficient, is de-fined as the input impedance or reflectioncoefficient of an element with all other ele-ments excited as in aclual array operation.

In general the active input impedances orreflection coefficients will change if the excita-tion is changed, because the changed relativephases or amplitudes of excitation producedifferent resultant terminal voltages. Suchchanges occur, for example, when the patternof an array of broadcast towers is being ad-justed, or when the major lobe of a large planararray is being scanned. If either the impedanceor the scattering matrix is known, the activeimpedance or reflection coefficient of all ele-ments in an array can be evaluated under allscan conditions [ 1181, [ 1191 . For microwavearrays the scattering matrix is more convenientto use for measurement and analysis. A par-

121

IEEEStd 149-1979

tially filled scattering matrix consisting of onlythe outward coupling coefficients about a givenexcited element provides an efficient means todetermine the active array performance from aspecified set of array incident-wave excitationsby indexing it over the full array [ 1201. Thisprocedure eliminates the need to invert verylarge impedance matrices as well as the needto determine the complete scattering matrix.

The mutual impedance between two elementsof an array is sometimes determined by a calcu-lation [ 1171, [ 1211, [ 1551. Such a calculationis usually practical only when all the elementsof the array have some simple, idealized config-uration. This situation occurs if the form of thecurrent distribution or aperture distributionis not significantly affected by the environmentin which the element is located.

The direct determination of the mutualimpedance from voltage and current measure-ments is most commonly made on antennaarrays in the medium-frequency range. A com-mon example is a broadcast antenna arrayconsisting of several electrically-small towerelements. A knowledge of the mutual impe-dance in such arrays is necessary to determinethe branching and coupling circuit require-ments which will provide the proper currentamplitude and phase relations in t.he arrayelements to produce the desired directionalcharacteristics. Frequently such antenna arraysemploy tower elements of different heights inirregular geometric arrangements with unequalcurrent amplitudes and irregular phase relation-ships in order to produce a radiation patternthat best fits the desired coverage and projecteddirections. In an antenna array containing onlya few radiating elements, the nonrepetitiverelative position of each element in the arraymay result in a substantially different mutualimpedance between different pairs of elements.

At much lower frequencies, antenna arraysare not generally used. When they are, however,the currents and voltages which establish themutual impedances are directly measurable asat medium frequencies. In the higher radio-frequency region, where currents and voltagesare not readily measurable, the mutual impe-dances between elements of an array may be

GROUND-WAVE

derived from measurements of the input im-pedances of the elements at their referencepoints under appropriate conditions. Two ormore terminating conditions for the coupledelements are required, such as short and opencircuits at the terminals, the latter permitsmeasurement of the self-impedances of the ele-ments. Such measurement procedures are basedon an exact analogy between the current andvoltage relations between the terminals in asystem of radiating elements and the corres-ponding current and voltage relations in acircuit network. An impedance bridge, admit-tance meter or vector voltmeter is often usedfor such measurements, and a specific methodis described in the literature [ 1221.

In UHF and microwave antenna arrays con-sisting of a large number of identical regularly-spaced radiating elements, the mutual impe-dance between all pairs of elements having thesame relative position tends to approximatea constant value, except near the edges of thearray [1X3], [123]. In such cases measure-ments are sometimes made on a limited-sizearray having an element design and arrangementlike that of the full-size array. By measuringthe mutual impedance or coupling betweenelements near the center of the limited array, apartial impedance or scattering matrix can bedetermined which has values that approximatethe corresponding ones in the full-size array.However, care shall be taken in interpreting theresults in terms of various related aspects ofarray performance such as the active impedanceas a function of electronic beam scanning, orthe gain and shape of the element pattern. Theomission of the effect of the missing elementson the limited array may correspond to signifi-cant errors in the aformentioned quantitieswhen they are used for the full-size array.

Alternative techniques that overcome thislimitation have been described [124], [ 1251for the measurement of the active impedanceof an element in a large array of identical,regularly spaced elements. The effect ofelements in an infinite array is simulated by theuse of waveguide techniques. These techniquesare based upon image theory and yield resultsfor a l imi ted number o f d i sc re te beampositions.

122

t

MEASUREMENTS

1’7. Ground-Wave Measurements

The measurements described in Sections 3, 7,and 9-12 presumed that the performance ofthe antenna is desired in the far-field region.However, some antenna systems deliver theiruseful power to regions were the simplifyingfar-field relationships do not apply. An impor-tant class of such antennas are ground-basedvertically polarized antennas which operate atfrequencies that are low, which rely on theground as an essential means for wave propaga-tion, and which radiate their useful power toreceiving antennas located on or near theground. In this case the ground can be con-sidered a significant part of the antenna itself,and cannot be in the far-field region.

The useful component of radiation for suchantennas is usually represented by the groundwave (see IEEE Std 211-1977, Standard Defini-tions of Terms for Radio Wave Propagation).This is a wave associated with currents thatflow in the ground, which is an imperfect con-ductor. Power is absorbed by the ground, andthe ground wave ordinarily decays at a morerapid rate than would a wave in free space.The concepts of power gain, directivity, radia-tion resistance, and radiation efficiency are notdirectly applicable in this situation; althougharbitrary assumptions regarding the extent ofthe antenna system are sometimes made. Thispermits a limited use of these terms. Further-more the electrical properties of the imperfectlyconducting ground and the type of terrain mayvary over the useful coverage area surroundingthe antenna, so that irregularities in the ground-wave field strength are likely to exist.

A complete measurement of the ground-waveradiation by an antenna involves the measure-ment of the field strength at every significantpoint on the ground within the useful coveragearea (see IEEE Std 291-1969, Standards Re-port on Measuring Field Strength in RadioWave Propagation). The power supplied to theantenna during these measurements shall alsobe determined. A pattern may then be plottedshowing the field strength throughout thecoverage area for the specified value of antenna

IEEEStd l-19-1979

power. Such a pattern is often plotted in termsof contours of equal field strength [103, Sectionlo], [ 1261.

When certain conditions exist, a relativelysmall number of measurements can be used, inconjunction with well-known propagationformulas and published curves [ 127]-[ 1301,to estimate the field strengths at various otherpoints throughout the coverage area. Forantennas operating below approximately 5 MHzthe useful component of the ground wave isusually the surface wave because the space-wave component cancels out near the ground[ 103, Section lo] , [ 1311 . If the electrical pro-perties of the ground are reasonably constantalong lines radially outward from the antenna,the surface-wave field strength can approximatesimple functions of distance on each line. Thesesimple functions usually apply for distancesfrom the antenna greater than 1 wavelengthand greater than five times the vertical heightof the antenna, and for distances less than8 X lo6 (f)-l/3 meters (where f is the fre-quency in hertz), beyond which the curvatureof the earth begins to have a significant effect.

Four theoretical curves of relative fieldstrength versus relative distance are shown inFig 67 for this ideal situation. Each curverepresents a different set of electrical propertiesof the ground. In addition, the relative distancep is a function of the ground properties, so thateach curve requires a translation in the scaleof relative distance. By making about 10 or20 measurements of field strength for evenlyspaced intervals along a radial line within therange of distances mentioned, and by plottingthese on the same logarithmic scales as thoseof Fig 67, a curve may be chosen and locatedhorizontally and vertically so that it representsthe best fit to the measured data. This processis repeated for each radial line. By interpola-tion and extrapolation the field strength maynow be estimated at any point within thedescribed range [132]. If more detailed theo-retical curves are desired, they may be found inboth graphical [133] and tabular form in theliterature [128] (see also IEEE Std 291-1969).The electrical properties of the ground may be

123

IEEEStd 149-1979 POWER-HANDLING

0.0001 I :

QM PI I IO IM

PRELATIVE DISTANCE

Fig 67Decay of Surface-Wave Component of the Ground Wave for a Plane Earth

determined from the quantities b and p inFig 67 and from relationships presented in theliterature [ 1271, [ 1281 . This is necessary inorder to determine that the best fit to the mea-sured data as established in the preceding,represents a reasonable value.

It is often desired to estimate the fieldstrength at distances beyond the range where itapproximates a simple function of distance,and a helpful factor for this purpose is theunattenuated surface-wave field-strength pat-tern at a standard distance from the antenna.This is a hypothetical pattern that may beobtained from the measurements and the curve-fitting procedure described. For each radial setthe line marked “unattenuated field strength”

in Fig 67 is extended until it intersects thestandard distance (such as 1 kilometer or 1mile), and the value of field strength is readoff. The resulting pattern of unattenuated fieldstrength versus azimuth angle for the standarddistance may be used, in conjunction withwell-known propagation formulas and pub-lished curves [127] -[130] , to estimate thefield strengths at greater distances, whereeffects such as the curvature of the earthbecome important.

For antennas operating above about 5 MHzthe space-wave component of the ground wavecannot usually be neglected, and the contourof the ground is a significant factor. However,a hypothetical unattenuated pattern at a

124

MEASUREMENTSIEEE

Std 149-1979

standard distance is still a helpful concept inestimating field strengths over a coveragearea. In this case the unattenuated field strengthmay be determined by a comparison methodin which the antenna being tested is comparedwith a standard antenna. The comparisonmethod requires that the antenna under testbe removable. This is generally possible inthe frequency range above 5 MHz where an-tennas are usually not extremely large. Thestandard antenna may consist of either a verti-cal loop or a vertical dipole for which thevertical electric field Es expected over a per-fectly conducting ground plane can be calcu-lated when the antenna current is specified.This value is just twice the free-space value ofthe vertical field strength. The method is validonly if the vertical-plane patterns of the antennaunder test and of the standard antenna are suchthat the relative effect of the ground-reflectedcomponent of the space wave is the same ineither case. First the signal voltage Vx receivedin a vertically polarized pickup antenna placedat some standard distance from the antennaunder test is measured. This distance shouldbe greater than 1 wavelength and greater thanten times the largest dimension of the testantenna. Then the antenna under test is removedand the standard antenna is placed in the sameeffective location with due consideration givento the probable difference in current distribu-tions of the two antennas. The signal voltageVs received in the pickup is now measured fora current in the standard antenna equal to thatassumed in the field-strength calculation. Theunattenuated field strength Ex of the testantenna at the standard distance is then deter-mined from the following relation:

VWE, = --A Es

VS

To determine the complete unat tenuatedground-wave field-strength pattern in thehorizontal plane, this procedure is repeated ina number of directions from the antenna undertest.

In addition to its use in estimating fieldstrengths, the unattenuated ground-wave field-strength pattern is helpful as an indication of

the quality of the antenna design. With theunattenuated pattern some of the variableeffects of the ground are removed, particularlythose at distances far from the antenna. Anexample of this is the determination of equiva-lent radiated power for ground-wave trans-mission (see IEEE Std 291-1969).

It is sometimes desired to determine thevertical-plane pattern of a ground-based verti-cally polarized antenna. This pattern is ofinterest, for example, when the “skywave”radiation characteristics are important [ 103,Section lo] (see also IEEE Std 211-1977). Al-though the pattern in the far-field region of theantenna-ground system may be desired for theprediction of skywave characteristics, practicallimitations may rule out the far-field measure-ments described in Section 9. In particular, atlow elevation angles and low frequencies, theremay be a special problem in the measurementsbecause of the presence of ground-wave compo-nents in the total radiation field of the antenna.In this case the shape of the vertical-plane pat-tern can depend very markedly on the distanceatwhich the pattern is measured [127], [134].However, it is sometimes possible to estimatethe far-field vertical-plane pattern from a seriesof near-field measurements by subtracting out,with appropriate relative phase, the surface-wave and the space-wave components of theground wave.

18. Power-Handling Measurements

Antennas are often tested to determine theirability to handle the power generated by theirassociated transmitters. These tests may beconcerned primarily with limitations imposedby metallic or dielectric heating at high averagepower levels, or with limitations imposed bythe arcing, voltage breakdown, or corona dis-charge associated with high electric fields athigh peak-power levels.

In cases wnere the antenna is subject to lowatmospheric pressures it is necessary to simu-late the high-altitude conditions by means ofa bell jar or some other suitable chamber in

125

IEEEStd 119-1979

order to achieve satisfactory peak-power-handling measurements [ 1351, [ 1361, sinceextrapolation from tests conducted at sea levelis not reliable. At very low pressues the break-down power level decreases with increasingpressure, reaches a minimum at a pressure thatdepends on the wavelength (the glow dischargeregion), and then increases with pressure athigh pressures. The use of a radioactive materialto create a continuous supply of free electronsin the atmosphere in the near vicinity of theantenna under test is a useful means of pro-viding reliability and consistency in the break-down tests, and does not lower the absolutebreakdown threshold. At satellite altitudes,multipacting [137] may cause problems andshall be considered as a possible breakdownmechanism in any antenna power-handlingtests. Multipacting is an antenna resonancephenomenon occurring under high vacuumconditions. When the travel time of electronsfrom one antenna electrode to another iscomparable to the inverse of the operatingfrequency, a buildup in electron density maytake place. This is due to surface ionizationcaused by electron bombardment.

Because of the great variety of antenna con-figurations and environmental conditions, noattempt will be made to provide a specificprocedure for conducting power-handling mea-surements. For accurate results it is impor-tant to ensure that the source of power for themeasurements has the same characteristics asthe actual transmitter with which the antennais to be used, that is, the modulation, pulseshape. pulse width, pulse rate, and so on shouldbe the same and should not change with theapplied power level during the test. Tempera-ture rise may hc measured by means of thermo-couples or temperacure-sensitive paints appliedto critical surfaces. Care should be taken thatthese added components are not themselvesheated by the radio-frequency power. Depend-ing son the type, breakdown may be detectedby visual observation, audible indication,change in signal picked up by a transmission-monitor antenna, or change in a reflected-wave-monitor signal as the applied power levelis increased. It is sometimes important, as in

ELECTROMAGNETIC

the case of glow discharge, to determine thepower level at which breakdown ceases once ithas been initiated. Sharp comers, dirt, metalparticles, and corrosion are often major of-fenders in lowering the breakdown powerlevel.

In the case of antennas carried on reentryvehicles, it is necessary to determine the power-handling capability in the presence of theionized air enveloping the reentry vehicle[138]. As the power transmitted from theantenna is increased, a level is reached wherethe power absorbed by the ionized gas is suf-ficient to heat the gas and increase the electrondensity. The critical plasma density is the mostimportant parameter in determining break-down in such cases. It is defined as that electrondensity where the natural oscillation frequencyof the ionized gas is equal to the frequency atwhich the antenna is being driven. When thenatural oscillation frequency of the ionizedgas is much lower than the frequency appliedto the antenna, the plasma is termed under-critical or underdense. If the initial plasmadensity is greatly undercritical, the effect ofthe unperturbed gas upon the antenna fieldmay be neglected. Breakdown is then usuallydefined as that input power at which the per-turbed ionized gas has become overcritical(overdense) and a precipitous drop in the trans-mitted power beyond the plasma takes place. Ifthe initial plasma density is close to or overcriti-cal, the definition of breakdown is more com-plex. The unperturbed ionized gas may have astrong effect on the antenna field and, in general,may not be neglected. One commondefinitionofbreakdown for this case is that input powerlevel above which an increase in power nolonger results in a significant increase in thepower transmitted beyond the plasma.

In order to measure the power-handling ca-pability of a missile antenna in the laboratoryit is necessary to model the plasma conditions.The breakdown power level may dependcritically upon a number of plasma parametersalong with the magnitude and spatial variationof electron density. A number of techniquesare available to produce plasmas in the labora-tory suitable for antenna-breakdown measure-

126

RADIATION HAZARDS

ments [ 1391, [ 1401. When testing an antenna,it is desirable to measure not only the break-down power level, but also the changes in inputimpedance, radiation pattern, power gain, anddistortion of the transmitted signal for powerlevels below, at, and above the breakdownlevel.

Analytical and numerical techniques [ 1411,have been developed to predict the power-handling capabilities of missile antennas. Thebreakdown power levels predicted by thesetechniques are in good agreement with flight-test results. If it should be impossible to modelparticular reentry conditions in the laboratory,these prediction techniques may prove suf-ficient in providing data for antenna designs.

Whenever measurements are performed inwhich high average power levels exist, it isessential that proper safety devices and proce-dures be employed for the protection of thepersonnel in the vicinity (see Section 19).

19. Electromagnetic Radiation Hazards

19.1 General. Individuals involved in the gen-eration of electromagnetic signals and in thetesting of antennas shall consider carefully thepotential hazard of exposure of humans toexcessive electromagnetic radiation. The spec-trum of nonionizing radiation covers the fre-quency range up to the ultraviolet-light region.The nonionizing radiation is the radiant energythat interacts with body tissue causing heatingbecause the body is a lossy dielectric.

NOTE: The biological effects of nuclear radiation(ionizing radiation) can be largely attributed to ioniza-tion and electronic excitation which causes destructionof various molecules.

This section presents brief comments perti-nent to health and safety precautions, whichapply primarily to radiant energy in the radio-frequency region. The purpose is to alert thoseengaged in testing antennas, microwave compo-nents, and so on, to potential radio-frequencyradiation hazards. In addition to radiationdamage, bums may be incurred when contactis established by arcing between the body andexposed components of a system operating at

IEEEStd 149-1979

high RF power levels.The interaction of radio waves and the human

body involves very complex processes that arenot completely understood. Considerable re-search has been conducted, and indeed is beingconducted, by researchers in many differentcountries. (For example, there are severalhundred references listed in four publications[ 142]-[ 1451.) Safe exposure limits are set as aresult of an analysis of all the data generatedby such research. As a consequence, one canexpect that standards for exposure limits willpossibly change as new knowledge is acquired.

In the United States the organizations whichare the principal proponents and sponsors ofwork leading to the setting of standards are theBureau of Radiological Health of the US PublicHealth Service, the National Science Founda-tion, the American National Standards Institute(ANSI Committee C-95), and the IEEE Com-mittee on Man and Radiation (COMAR). Simi-lar organizations exist in other countries.

The injurious effects on human beings arebelieved to be principally due to the heating oftissues at various depths of penetration. Suchheating can be especially detrimental to sensi-tive organs, such as the eyes, testicles, the kid-neys, and the liver. Excessive exposure can alsocause skin bums, but at RF the energy usuallypenetrates below the surface of the skin causingeither deep-skin bums or damage to internalorgans. It should be noted that the body orparts of the body exhibit resonances. In fact,the whole body is resonant in the VHF rangeof frequencies. The exact frequency dependsupon the size and shape of the body.

The principal factors which affect the amountof RF energy absorbed by the human body arethe frequency, polarization and power fluxdensity of the incident wave, the exposure time,and the electrical properties of the body.

The ability of the body to tolerate heat stressdue to absorbed RF energy without deleteriouseffects depends upon a number of factors.These include the environmental temperatureand humidity, the amount of heat already gen-erated in or absorbed by the body, and the stateof health of the person being exposed. Becauseof the variability of conditions under whichpersonnel must work, the safety limits must be

I

127

IEEEStd 149-1979 ENVIRONMENTAL

Table 4Electromagnetic Radiation Safety Limits

CountryRegulating

Body Safety Limit Exposure Time

us AmericanNationalStandardsInst i tute t

10 mW/cm’(averaged overany 6 minutes)

6 minutes

u s Air Force [ 1A61ttR F > lO>IHz 50 mW/cm’ over 6 minutes

300 mW-min/cmZ under 6 minutes

RF < 10 MHz 10 mW/cm2 over 6 minutes60 mW-min/cm’ under 6 minutes

USSR State Committee 0.01 mW/cmz over 24 hours**on Standards* 0.1 mW/cm? up to 2 hours

1 .O mW/cm* 20 minutes

?A11 exposures limited to Mean Squared Electric Field Strength of 40 000 V‘/m’, Mean Squared MagneticField Strength of 0.25 A’/mZ and an Energy Density of 1 mW-h/cm’

??A11 exposures limited to a Peak Electric Field Strength of 100 kV/m -- -7*For frequencies between 0.01-0.03 GHz the RMS Electric Field Strenphis limited to 20 V/m: for frequencies’

between 0.03-0.05 GHz, 10 V/m, 0.3 A/m; for frequenciesI ~~~,

duration of one working day**Continuous exposure.

between 0.05-0.3 GHz, 5 V/m all for an exposure

considered to be upper limit guide numbers fora normal or moderate environment. If, forexample, the temperature and humidity arehigh in the area where personnel are exposed toRF. radiation, it is appropriate to include asafety factor, thereby reducing the guide num-bers recommended in the applicable standard.Also, it should be noted that individuals withcirculatory difficulties are especially vulnerableto adverse affects, due to heat stress.

19.2 Safe Radiation Limits. Uniform radia-tion limits have not been established world-wide. For instance, there is a wide discrepancybetween the United States and the USSR in theestablishment of maximum allowable limits forthe radiation exposure of human beings. ANSIc95.1-1974, American National StandardSafety Level of Electromagnetic RadiationRespect to Personnel, includes limitations onthe E and H fields as well as the power density.It recommends a radiation protection guide of10 mW/cm*, averaged over any possible 0.1hour period for frequencies from 10 MHz to100 GHz. In the USSR’s Occupational SafetyStandard, GOST 12.1.006-76, the safety limits

pergiven in step-function levels of milliwattssquare centimeter for different exposure

times and frequencies between 0.3 GHz and300 GHz. For frequencies between 0.01 GHzand 0.3 GHz, the limits are on the electric andmagnetic field strengths [ 1561. Table 4 sum-marizes the safety limits of the two countries.Other countries have similar standards whichshall be adhered to by engineers and scientists[ 1571 practicing within their boundaries.

The safety limits presented in Table 4 arebased upon plane wave or locally plane wavepropagation. This means that they apply un-ambiguously only for the far field of an antenna[147]. In the absence of reflecting obstacles,the far field power flux densities can often becalculated. In the near field, however, suchcalculations are too inaccurate and measure-ments are required [ 1481, [ 1491. For the nearfield case where reactive fields are significant,the concept of power flux density is invalid[ 1471. One can, however, measure the magni-tude of either the electric or magnetic fields orthe squares of these quantities, which are pro-portional to the energy density of the electricor magnetic fields respectively [150], Theseresults are sometimes used to obtain the power

128

FACTORS

flux density of an equ,valent plane wave inorder to reconcile the results with a standard.

19.3 Measurement and Instrumentation. In theUnited States the appropriate standard on mea-surements is ANSI C95.3-1973, Techniquesand Instrumentation for the Measurement ofPotentially Hazardous Electromagnetic Radia-tion at Microwave Frequencies (also see [ 1501and [158] ). Most often in antenna measure-ments it is the near field of an antenna that isof concern. For this case an ‘isotropic’ probe isrequired to sample the field. These probesusually consist of three mutually perpendicular,and very short, dipoles. The induced currentson the dipoles are detected by diodes, bolo-meters or thermocouples; the three detectedsignals are combined and amplified. The outputof the instrument is calibrated so that the mag-nitude of the electric field or its square can bepresented on a meter or other type of indicator.Some instruments are calibrated in terms of anequivalent power flux density as discussed in19.2. Such instruments are commercially avail-able.

The measurement must be performed withcare since the probe and its associated ‘instru-mentation can distort the field being measuredto the degree that erroneous results are obtain-ed. In addition the presence of the person per-forming the measurement can also disturb thefield being measured. Measurements made re-motely are to be preferred.

It is good laboratory practice to experiment-ally survey any region where potentially hazard-ous fields may exist and appropriately postthose areas where the measured field strengthsexceed prescribed limits. This is to alert allpersonnel to the possible hazard. Also, therange operations manual [see Section 81 shouldcontain warnings to personnel of any potentialhazardous situations that may commonlyoccur.

20. Environmental Factors

An antenna can be considered adequately .tested only when the tests have included theenvironmental conditions under which the

IEEEStd 149-1979

antenna will operate. Many of these environ-mental factors are of a specialized nature, andit is impractical to include all the appropriateeffects in this test procedure. Instead a fewcases will be briefly mentioned as pertinentexamples.

One category of environmental factors maybe defined as those directly affecting thematerial properties or structure of an antennaand, thus, indirectly affecting the electricalcharacteristics. Mechanical loading of theantenna structure by wind and ice is a com-mon, but nevertheless important, effect to beconsidered in the testing of many antennas.Vibration and shock tests are often made toassure that an antenna that is subject to severeaccelerations will maintain its structuralintegrity, and to determine whether dynamicdeformations are within allowable electricallimits. Antennas in exposed locations Jreoften provided with lightning protection andanti-icing devices; the effect of such devices onthe electrical characteristics must be evaluated.

Various natural or man-made environmentsmay impose special requirements. For example,shipborne antennas may have to withstandwater-wave impact and salt-water corrosion.Antennas on hypersonic vehicles shall with-stand very high temperatures and pressures,and antennas designed for satellite or space-probe application shall withstand intenseionizing radiation, hard vacuum, and extremetemperatures. Ground-based antennas that areintended to operate in the vicinity of a nuclearblast should maintain their essential propertiesin the wake of seismic waves, atmosphericshock waves, thermal radiation, ionizing radia-tion, blast-product erosion, and radioactivedebris.

Certain antenna applications necessitate un-usual attention to tests that involve quiteordinary aspects of the physical environment.For instance, rapid-scan antennas employingferrite components are especially sensitive tochanges in ambient temperature. Antennas inwhich intermodulation distortion has to beminimized may have difficulty with nonlinearcorrosion films. In the case of precision track-ing antennas, a boresight error can be causedmerely by deflections resulting from nonuni-

129

IEEEStd 149-1979 BIBLIOGRAPHY

form heating of the antenna by the sun. Large from the physical environment. However, in soantennas rotated in the elevation plane may doing it introduces an electrical environmentalso undergo significant deflections because the effects of which should not be overlooked.of the change in the effect of the gravitational Radome testing is described in Section 15.force on the antenna.

The other category of environmental factorsmay be defined as those directly affecting theelectrical properties of the antenna. In somecases this electrical environment is permanentand is included in the design as an integralpart of the antenna. Familiar examples of thissituation are the presence of the ground forground-based antennas, the superstructure forshipbome antennas, and the vehicle for air-borne or space antennas. Techniques for testingin these circumstances have been developedand are widely used. They are described in 7.1and Section 9. Another example is the noiseenvironment of an antenna, including galacticradiation, thermal atmospheric radiation, at-mospheric and man-made static, and thermalearth radiation. A less familiar but moreextreme example of permanent electrical en-vironment occurs with low-frequency antennasdesigned to operate underground or beneaththe surface of the sea.

In many cases the electrical environment ofan antenna is transient, and causes problems ofa more elusive nature. Moisture, when com-bined with impurities, can create conductingpaths that are troublesome, especially inregions of high electric field. Even pure water,if it exists in a continuous film, may affect theperformance of a microwave antenna becauseof its high dielectric constant. Ice or snow arealso factors that can directly affect the char-acteristics of an antenna. In the case of antennasfor missile application, ionized gases mayexist in the exhaust or may be generated byhigh vehicle velocity through the atmosphere;these gases are also a part of the electricalenvironment. Finally, certain antennas may besusceptible to precipitation static, which is aseries of noise pulses created when chargedparticles such as raindrops discharge on theantenna.

There is a particular type of antenna environ-ment that deserves special mention, namely,the radome. A radome is a structure that isoften used to enclose an antenna to protect it

130

[II

[21

[31

141

[Sl

[61

171

[81

21. Bibliography

HOLLIS, J. S., LYON, T. J., andCLAYTON, L., Jr, Eds. MicrowaveAntenna Measurements. Atlanta, GA.Scientific-Atlanta, 1970.Standard Coordinate System and DataFormats for Antenna Patterns, Elec-tronic Trajectory Measurements Work-ing Group, Inter-Range InstrumentationGroup, Range Commanders Council,United States National Ranges, IRIGDocument AD 637 189, May 1966.CUMMING, W. A. Radiation Measure-ments at Radio Frequencies: A Surveyof Current Techniques, Proceedings ofthe IRE, vol 47, May 1959, pp 705-735.CUTLER, C. C., KISG, A. P., and .KOCK, W. E. Microwave AntennaMeasurements, Proceedings of the IRE.vo135, Dee 1947, pp 1462-1471.PIPPARD, A. B., BURRELL, 0. J., andCROMI, E. E. The Influence ,of Re-Radiation on Measurements of thePower Gain of an Aerial, Journal ofthe Institution of Electrical Engineers.vol 93, pt III A, 1946, pp 720-722.KING, H. E., SHIMABUKURO, F. I..and WONG, J. L. Characteristics of aTapered Anechoic Chamber, I E E ETransactions on Antennas and Propa-gation (Communications), vol AP-15.iMay 1967, pp 488-490.BACHMAN, C. G., KING, H. E., andHANSEN, R. C. Techniques for Mea-surement of Reduced Radar CrossSections, pt 1, Microwave Journal.~016, Feb 1963, pp 61-67.MOELLER, A. W. The Effect ofGround Reflections on Antenna TestRange Measurements, Microwave Jour-nal, vol 9, Mar 1966, pp 47-54.

BIBLIOGRAPHY

[91

[lOI

[Ill

[I21

[I31

[I41

[I51

1161

[I71

[I81

[ 191

ARNOLD, P. W. The “Slant” AntennaRange, IEEE Transactions on Antennasand Fropaga tion (Communications), volAP-14, Sept 1966, pp 658-659.COHEN, A., and MALTESE, A. W.The Lincoln Laboratory Antenna TestRange, Microwave Journa l , voI 4,,4pr 1961, pp 57-65.JORDAN, E. C., and BALMAIN, K. G.Electromagnetic Waves and RadiatingSystems, Englewood Cl i f fs , NJ ,Prentice-Hall, 1968, 2nd ed.REED, H. R., and RUSSELL, C. M.Ultra High Frequency Propagation,Cambridge, MA, Boston Technical Pub-lishers, 1966, 2nd ed.JULL, E. V., and DELOLI, E. P. AnAccurate Absolute Gain Calibration ofan Antenna for Radio Astronomy,IEEE Transactions on Antennas andPropagation, vol AP-12, July 1964,pp 439-447.Minimum Standards for Land-MobileCommunication Antennas. Part 1 -Baseor Fixed Station Antennas, Engineer-ing Department, Electronic IndustriesAssociation, Washington, DC, Dee 1966.JOHNSON, R. C., ECKER, H. A., andMOORE, R. A. Compact Range Tech-niques and Measurements, IEEE Trans-actions on Antennas and Propagation,vol AP-17, Sept 1969, pp 563-576.JOHNSON, R. C., ECKER, H. A., andHOLLIS, J. S. Determination of Far-Field Antenna Patterns from Near-FieldMeasurements, Proceedings of the IEEE,~0161, Dee 1973, pp 1668-1694.EMERSON, W’. H. ElectromagneticWave Absorbers and Anechoic Cham-bers Through the Years, IEEE Trans-actions on Antennas and Propagation,vol AP-21, July 1973, pp 484-490.

POUND, R. V. Electronic FrequencyStabilization of Microwave Oscillators,Review of Scientific Instruments, vol17, Nov 1966, pp 490-505.DYSON, J. D. The Measurement ofPhase at UHF and Microwave Fre-quencies, IEEE Transactions on Micro-

[201

c211

I221

~31

[241

v51

[261

1271

[=I

[291

[301

[311

IEEEStd 149-1979

wave Theory and Techniques, v o lMTT-14, Sept 1966, pp 410-413.LOMBARDI, A. J., and POLGAR, M.S.,Jr. Wide Band Antenna Facility,Electrical Communication, ~0149, no 1,1974, pp 94-98.APPEL-HANSEN, J. Reflectivity Levelof Radio Anechoic Chambers, I E E ETransactions on Antennas and Propaga-tion, vol AP-21, July 1973, pp 490-498.BUCKLEY, E. F. Outline of EvaluationProcedures for Microwave AnechoicChambers, Microwave Journal, vol 6,Aug 1963, pp 69-75.SINCLAIR, G., JORDAN, E. C., andVAUGHAN, E. W. Measurement ofAircraft-Antenna Patterns Using Models,proceedings of the IRE, vol 35, Dee1947, pp 1451-1462.STRATTON, J. A. ElectromagneticTheory, New York, McGraw-Hill, 1941,pp 488-490.KING, R. W. Electromagnetic Engineer-ing, vol 1, New York, McGraw-Hill,1945, pp 316-320.MARCUVITZ, N. Waveguide Hand-book, M.I.T. Radiation LaboratorySeries, vol 10, New York, McGraw-Hill,1951, pp 18-23.SINCLAIR, G. Theory of Models ofElectromagnetic Systems, Proceedingsof the IRE, vol 36, Nov 1 9 4 8 . p p1364-1370.FANO, R. M. Theoretical Limitationson the Broadband Matching of ArbitraryImpedances, Journal of the FranklinInstitute, vol249,1950, pp 57-154.BOLLJXHN, J. T., and REESE, R. F.Electrically Small Antennas and theLow-Frequency Aircraft Problem, IRETransactions on Antennas and Fropaga-tion, vol AP-1, Ott 1953, pp 46-54.

GREENE, F. M. NBS Field-StrengthStandards and Measurements (30 Hz to1000 MHz), Proceedings of the IEEE,vol 55, June 1967, pp 970-981.

CHENG, D. K., and MOSELEY, S. T.On-Axis Defocus Characteristics of the

131

IEEEStd 149-1979

~321

[331

[341

[351

[361

E371

[381

[391

1401

[411

Paraboloidal Reflector, IRE !lYansuc-t ions on Antennas and Propagation(Communications), vol AP-3,Oct 1955,p p 2 1 4 - 2 1 6 .KERNS, D. M., and DXYHOFF, E. S.Theory of Diffraction in MicrowaveInterferometry, Journal of Research ofthe National Bureau of Standards,vol64B, Jan/Mar 1960, pp l-13.KERNS, D. M. Correction of Near-FieldAntenna Measurements Made with anArbitrary but Known Measuring An-tenna, Electronics Letters, vol

AntennaMeasurements,

JOY, E. B., and PARIS, D. T. SpatialSampling and Filtering in Near-FieldMeasurements, IEEE Transactions onAntennas and Propagation, vol AP-20,May 1972, pp 253-261.NEWELL, A. C., and CRAWFORD,M. L. Planar Near-Field Measurementson High Performance Array Antennas,National Bureau of Standards, Boulder,CO, Rep NBSIR 74-380,1974.

Special Issue on Fast Fourier Trans-form, IEEE Transactions on Audio andElectroacoustics, vol AU-17, June1969, pp 65-172.LEACH, W. M., Jr, and PARIS, D. T.Probe Compensated Near-Field Measure-ments on a Cylinder, IEEE Transac-tions on Antennas and Propagation,vol XP-21, July 1973, pp 435-445.JENSEN, F. Electromagnetic Near-FieldFar-Field Correlations, Ph.D. disserta-tion. Technical University of Denmark,Lyngby, July 1970.LUDWIG, A. C. Near-Field Far-FieldTransformations Using Spherical-WaveExpansions, IEEE Transactions onAntennas and Propagation, vol AP-19, Mar 1971, pp 214-220.WXCKER, P. F., Non-Planar Near-FieldMeasurements: Spherical Scanning, ,va-

[421

[431

[441

[451

[461

I471

1481

[491

[501

1511

BIBLIOGRAPHY

tional Bureau of Standards, Boulder.CO, Rep NBSIR 75-809,1975.BANTIN, C. C., and BALMAIN, K. G.Study of Compressed Log-Periodic Di-pole Antennas, IEEE Transactions onAntennas and Propagation, vol AP-18,Mar 1970, pp 195-203.BALMAIN, K. G., BANTIN, C. C.,OAKES, C. R., and DAVID, L. Opti-mization of Log-Periodic Dipole An-tennas, IEEE Transactions on Antennasand Propagation (Communications), volAP-19, Mar 1971, pp 286-288.OAKES, C. R., and BALMAIN, K. G.Optimization of the Loop-Coupled Log-Periodic Antenna, IEEE Transactions onAntennas and Propagation, vol AP-21,Mar 1973, pp 148-153.GRUNER, R. W., and ENGLISH, W. J.Antenna Design Studies for a U.S.Domestic Satellite, COMSTAT Tech-nical Reuiew, vol 4, no 2, Fall 1974,pp 413-447.KREUTEL, R. W., DIFONZO, D. F.,ENGLISH, W. J., and GRUNER, R. W.Antenna Technology_ for FrequencyReuse Satellite Communications, Pro-ceedings of the IEEE, ~0165, Mar 1977,pp 370-378.FITZGERRELL, R. G. Swept-FrequencyAntenna Gain Measurements, I E E ETransactions on Antennas and Propaga-tion, vol AP-14, Mar 1966, pp 173-178.KENEFICK, J. F. Ultra-Precise Analyt-ics, Photogrammetric Engineering, vol37, Nov 1971, pp 1167-1187.FINDLAY, J. W., and VON HOEMER, S.A 65-Meter Telescope for hlillimeterWavelengths, National Radio Astron-omy Observatory, Charlottesville, VA,1972.PAYNE, J. M. An Optical DistanceMeasuring Instrument, Review of Scien-tific Instruments, vol 44, Mar 1973 ,pp 304-306.

LALONDE, M. The Upgraded AreciboObservatory, Science, vol 186, Ott1974, pp 213-218.

132

BIBLIOGRAPHY

[521

[531

[541

[551

[561

[571

[58 1

[591

[GOI

[611

PAYNE, J. M., HOLLIS, J. M., andFINLEY, J. W. New Method of Mea-suring the Shape of Precise AntennaReflectors, Review of Scientific Instru-ments; ~0147, Jan 1976, pp 50-55.FINDLAY, J. W., and PAYNE, J. M.An Instrument for Measuring Deforma-tions in Large Structures, IEEE Trans-a c t i o n s o n Instrumentation andMeasurement, vol IM-23, Sept 1974,pp 221-226.BALE, F. V., GOURLAY, J. A., andMEADOWS, R. W. Measuring theShape of Large Reflectors by a SimpleRadio Method, Electronics Letters,~012, July 1966, pp 252-253.CLARK, G. A., and SLATER, R. H.Experimental Laser System for Moni-toring Deformations in Large RadioReflectors, Proceedings of the In-stitution of Electrical Engineers, vol118, Nov 1971, pp 1562-1568.BRUECKMANN, H. Antenna PatternMeasurement by Satellite, IEEE Trans-actions on Antennas and Propagation,vol AP-11, Mar 1963, pp 143-147.KENNEDY, J. T., and ROSSON, J. W.The Use of Solar Radio Emission forthe Measurement of Radar Angle Errors,Bell System Technical Journal, vol 41,Nov 1962, pp 1799-1812.KUZ’MIN, A. D., and SALOMONO-VICH, A.E. Radio-Astronomical Meth-ods of Antenna Measurement, NewYork, Academic Press, 1966.FITZGERRELL, R. G. Gain Measure-ments of Vertically Polarized Antennasover Imperfect Ground, IEEE Trans-actions on Antennas and Propagation,vol AP-15, Mar 1967, pp 211-216.

DYSON, J. D. Determination of thePhase Center and Phase Patterns ofAntennas, in Radio Antennas forAircraft and Aerospace Vehicles,BLACKBAND, W. T., Ed, AGARDConference Proceedings, no 15, Slough,England, Technivision Services, 1967.

HU, Y. Y. A Method of DeterminingPhase Centers and Its Application to

IEEEStd 149-1979

Electromagnetic Horns, Journal of theFranklin Institute, vol 271, Jan 1961,pp 31-39.

[62] BAUER, K. The Phase Centers of Aper-ture Radiators, Archiv der Elek trichenfibertragung, vol 9,1955, p 541.

[63] CARTER, D. Phase Centers of Micro-wave Antennas, IRE Transactions onAntennas and Propagation, vol AP-4,Ott 1956, pp 597-600.

[641

[651

[661

[671

[68 1

[691

[701

NAGELBERG, E. R. Fresnel RegionPhase Centers of Circular ApertureAntennas, IEEE Transactions on An-tennas and Propagation (Communica-tions), vol AP-13, May 1965, pp479-480.DYSON, J. D. The Measurement ofNear Fields of Antennas and Scatterers,IEEE Transactions on Antennas andPropagation, vol AP-21, July 1973,pp 446-460.ELLERBRACH, D. A. UHF and Micro-wave Phase-Shift Measurements, Pro-ceedings of the IEEE, vol 55, J u n e1967, pp 960-969.

SCHAFER, G. E. Mismatch Errors mMicrowave Phase Shift Measurements,IRE Transactions on Microwave Theoryand Techniques, vol MTT-8, Nov1960, pp 617-622.

HUNTON, J. K. Analysis of MicrowaveMeasurement Techniques by Means ofSignal Flow Graphs, IRE Transactionson Microwave Theory and Techniques,vol MTT-8, Mar 1960, pp 206-212.

LEED, D. Use of Flow Graphs to Eva1uate Mistermination Errors in Loss andPhase Measurements, IRE Transactionson Microwave Theory and Techniques(Correspondence), vol MTT-9, Sept1961, p 454.

BOOKER, H. G., RUMSEY, V. H.,DESCHAMPS, G. A., KALES, M. I.,and BOHNERT, J. I. Techniques forHandling Elliptically Polarized Waveswith Special Reference to Antennas,Proceedings of the IRE, vol 39, May1951, pp 533-552.

133

IEEES t d 1 4 9 - 1 9 7 9 BIBLIOGRAPHY

[711

[72]

[731

[741

[751

[761

[771

[781

1791

[801

DESCHAMPS, G. A., and MAST, P. E.Poincare Sphere Representation of Par-tially Polarized Fields, IEEE Transac-tions on Antennas and Propagation,vol AP-21, July 1975, pp 474-478.COHEN, M. H. Radio AstronomyPolarization Measurements, Proceedingsof the IRE, vol 46, Jan 1958, pp 172-183.KRAUS, J. D. Antennas, New York,McGraw-Hill, 1950.BECKMANN, P. The Depolarization ofElectromagnetic Waves, Boulder, CO,Golem Press, 1968, pp 183-187.TAI, C. T. On the Definition of the Ef-fective Aperture of Antennas, I R ETransactions on Antennas and Propa-ga tion (Communications), vol AP-9,Mar 1961, pp 224-225.DYSON, J. D. An Approach to a FullyIsotropic Wide-Band Source of Radia-tion, Abstracts of the 20th AnnualUSAF Symposium on Antenna Researchand Development, Sponsored by theAF Avionics Labratory. Monticello,IL, Ott 1970.CLAYTON, L. Jr., and HOLLIS, J. S.Antenna Polarization Analysis by Am-plitude Measurement of Multiple Com-ponents, Microwave Journal, vol 8 ,Jan 1965, pp 35-41.KNITTEL, G. H. The PolarizationSphere as a Graphical Aid in Determin-ing the Polarization of an Antenna byAmplitude Measurements Only, IEEETransactions on Antennas and Propaga-tion, vol AP-15, Mar 1967, pp 217-221.

NEWELL, A. C., BAIRD, R. C., andWACKER, P. F. Accurate Measurementof Antenna Gain and Polarization atReduced Distances by an Extrapola-tion Technique, IEEE Transactions onAntennas and Propagation, vol AP-21, July 1973, pp 418-431.

JOY, E. G., and PARIS, D. T. A Prac-tical Method for Measuring the ComplexPolarization Ratio of Arbitrary An-tennas. IEEE Transactions on Antennas

[811

[=?I

[831

[841

[851

[861

[871

[881

[891

and Propagation, vol AP-21, July 1973,pp 432-435.LUDWIG, A. Gain Computations fromPattern Integration, IEEE Transactionson Antennas and Propagation (Com-munications), vol AP-15, Mar 1967,pp 309-311.SLAYTON, W. T. Design and Calibra-tion of Microwave Antenna GainStandards, Naval Research Laboratory,Washington, DC, Rept 4433, Nov1954.BOWMAN, R. R. Field Strength Above1 GHz: Measurement Procedures forStandard Antennas, Proceedings of theIEEE, vol 55, June 1967, pp 981-990.HEMMING, L. H., and HEATON, R. A.Antenna Gain Calibration on a GroundReflection Range, IEEE Transactionson Antennas and Propagation, vol AP-21, July 1973, pp 532-538.KANDA, M. Accuracy Considerationsin the Measurement of the Power Gainof a Large Microwave Antenna, IEEETransactions on Antennas and Propa-gation (Succinct Papers), vol. AP-23,May 1975, pp. 407-411.FITZGERRELL, R. G. The Gain of aHorizontal Half-Wave Dipole overGround, IEEE Transactions on Anten-nas and Propagation (Communications),vol AP-15, July 1967, pp 569-571.FITZGERRELL, R. G. Limitations onVertically Polarized Ground-Based An-tennas as Gain Standards, IEEE Trans-actions on Antennas and Propagation(Communications), vol AP-23, Mar1975, pp 284-286.

SMITH, P. G. Measurement of the Com-plete Far-Field Pattern of Large Anten-nas by Radio-Star Sources, I E E ETransactions on Antennas and Propa-.gation, vol AP-14, Jan 1966, pp 6-16.

BAARS, J. W. M. The Measurement ofLarge Antennas with Cosmic RadioSources, IEEE Transactions on Anten-nas and Propagation, vol AP-21,July 1973, pp 461-474.

134

BIBLIOGRAPHY

[goI

[911

1921

[931

[941

[951

1961

1971

[981

GUIDICE, D. _4., and CXSTELLI, J. P.The Use of Extraterrestrial RadioSources in the Measurement of AntennaParameters, IEEE Transactions on Aero-space and Electronic Systems, v o lAES-7, Mar 1971, pp 226-234.BXARS, J. W. M., MEZGER, P. G., andWENDKER, H. The Spectra of theStrongest Nonthermal Radio Sourcesin the Centimeter Wavelength Region,Astrophysical Journal, vol 142, 1965,pp 122-134.WAIT,D. F., DAYWITT, W.C., KANDA,M., and MILLER, C. K. S. A Study ofthe Measurement of G/T Using Cassi-opeia A, National Bureau of Standards,Boulder, CO, Rep NBSIR 74-382,June 1974.BAARS, J. W. M., andHARTSUIJKER,A. P. The Decrease of Flux Density ofCassiopeia A and the Absolute Spectraof Cassiopeia A, Cygnus A and TaurusA, Astronomy and Astrophysics, ~0117,1972, p 172.

Influence of the Non-Ionized Regionsof the Atmosphere on the Propagationof Waves, CCIR Plenary Assembly Rep234-1,1966.

DICKE, R. H., BERINGER, R., KYHL,R., and VANE, A. B. AtmosphericAbsorption Measurements with a Micro-wave Radiometer, Physical Review,vol 70, Sept 1 and 15, 1946, pp 340-348.

KANDA, M. Study of Errors in AbsoluteFlux Density Measurements of Cassi-opeia A, National Bureau of Standards,Boulder, C O , R e p SBIR 7 5 - 8 2 2 ,Ott 1975.

K R E U T E L , R . W . , J r , a n d PAC-HOLDER, A. 0. The Measurement ofGain and Noise Temperature of a Satel-lite Communications Earth Station,Microwave Journal, vol 12, Ott 1969,pp 61-66.JAKES, W. C., Jr, Gain of Electromag-netic Horns, Proceedings of the IRE.~0139, Feb 1951, pp 160-162.

‘I991

[lOOI

[loll

[lo21

11031

11041

[IO51

[lo61

[IO71

11081

IEEEStd 149-1979

WRIXON, G. T., and WELCH, W. J.Gain Measurements of Standard Elec-tromagnetic Horns in the K and K,Bands, IEEE Transactions on Antennasand Propagation, vol AP-20, Mar 1972,pp 136-142.CHU, T. S., and SEMPLAK, R. A.Gain of Electromagnetic Horns, Bel lSystem Technical Journal , vol 4 4 ,June/Aug 1965, pp 527-537.BEATTY, R. W. Discussion of Errorsin Gain Measurements of StandardElectromagnetic Horns, National Bureauof Standards, Boulder, CO, Tech Note351.NEWMAN, E. H., BOHLEY, P., andWALTER, C. H. Two Methods for theMeasurement of Antenna Efficiency,IEEE Transactions on Antennas andPropagation, vol AP-23, July 1975.pp 457-461.TERMAN, F. E. Radio Engineers’Handbook, New York, McGraw-Hill,1943.CADY, W. M., KARELITZ, M. G.,and TURNER, L. A. Radar Scannersand Radomes, M.I.T. Radiation Labora-tory ser., vol 26, New York, McGraw-Hill, 1948, pp 7,62,66.RHODES, D. R. Introduction to Mono-pulse, New York, McGraw-Hill, 1959.POVEJSIL, D. J., RAVEN, R. S., andWATERMAN,P. AirbomeRadar,Prince-ton, NJ, Van Nostrand, 1961, p 520.REDLIEN, H. W. The MonopulseDifference Chart, IEEE InternationalConvention Record, pt I, Antennas andPropagation, 1963, pp 129-131.KINSEY, R. R. Monopulse DifferenceSlope and Gain Standards, IRE Trans-actions on Antennas and Propagation(Communications), vol AP-10, May1962, pp 343-344.LYON, T. J. in Radome EngineeringHandbook, WALTON, J. D., Jr, Ed.,New York, Marcel Dekker, 1970,p 145.TERMAN, F. E., and PETTIT, J. M.

New York,Electronic Measurements,McGraw-Hill, 1952, chap 3

135

IEEEStd 149-1979

ill11

[I121

[I131

ill41

u151

I1161

[I171

[1181

Cl191

[I201

MONTGOMERY, C. G. Techniques ofMicrowave Measurements, M .I.T. Radia-tion Laboratory ser., vol 11, New York,McGraw-Hill, 1947, chap 8.SMITH, P. H. An Improved Transmis-sion Line Calculator, Electronics, vol17, Jan 1944, pp 130-138,318.CARTER, P. S. Charts for Transmission-Line Measurements and Computations,RCA Review, vol 3, Jan 1939, pp 355-368.KING, R. W. P., MACK, R. B., andSANDLER, S. S. Arrays of CylindricalDipoles, New York, Cambridge Uni-versity Press, 1968, chap 8.MAILLOUX, R. J., and LARUSSA, F.A Microwave Phase Bridge Techniquefor Measuring the Mutual Coupling ofIdentical Coupled Antennas, I E E ETransactions on Microwave Theory andTechniques (Correspondence), vol MTT-6, Feb 1968, pp 129-130.MONTGOMERY, C. G., DICKE, R. H.,and PURCELL, E. M. Principles ofMicrowave Circuits, M.I.T. RadiationLaboratory ser, vol 8, New York.McGraw-Hill, 1948, p 140.OLINER, A. A., and MALECH, R. G.Mutual Coupling in Infinite ScanningArrays, in Microwave Scanning Anten-nas, vol 2, HANSEN, R. C., Ed., Aca-demic Press, 1966, chaps 2, 3,4.

CARTER, P. S., Jr. Mutual ImpedanceEffects in Large Beam Scanning Arrays,IRE Transactions on Antennas andPropagation, vol AP-8, May 1960,pp 276-285.

KURTZ, L. A., ELLIOTT, R. S.,WEHN, S., and FLOCK, W. L. Mutual-Coupling Effects in Scanning DipoleArrays, IRE Transactions on Antennasand Propagation, vol AP-9, Sept 1961,pp 433-443.ALLEN, C. C. Mutual Coupling Effectsin Phased Array Antennas, Abstractsof the 13th Annual USAF Symposiumon Antenna Research and Development,University of Illinois, Urbana, Ott 15,1963, DDC ?#AD-421483.

[I211

[I221

1231

1241

[1251

11261

ii271

[I281

[I291

[I301

BIBLIOGRAPHY

CARTER, P. S. Circuit Relations inRadiating Systems and Applications toAntenna Problems, Proceedings of theIRE, ~0120, June 1932, pp 1004-1041.BROWN, G. H. Directional Antennas,Proceedings of the IRE, vol 25, J a n1937, pp 78-145.ALLEN, J. L. Gain and ImpedanceVariation in Scanned Dipole Arrays,IRE Transactions on Antennas andPropagation, vol AP-10, Sept 1962,pp 566-572.HANNAN, P. W., MEIER, P. J., andBALFOUR, M. A. Simulat ion ofPhased Array Antenna Impedance inWaveguide, IEEE Transactions on An-tennas and Propagation (Communica-tions), vol AP-11, Nov 1963, pp 715-716.GUSTINCIC, J. J. The Determinationof Active Array Impedance with Multi-element Waveguide Simulators, IEEETransactions on Antennas and Propaga-tion, vol AP-20, Sept 1972, pp. 589-595.BURROWS, C. R., HUNT, L. E., andDECINO, A. Ultra-Short-Wave Propaga-tion : Mobile Urban TransmissionCharacteristics, Bell System TechnicalJournal, vol 14, Apr 1935, pp 253-272.NORTON, K. A. The Propagation ofRadio Waves over the Surface of theEarth and in the Upper Atmosphere,Proceedings of the IRE, vol 24, Ott1936, pp 1367-1385; vol 25, Sept1937, pp 1203-1236.NORTON, K. A. The Calculation ofGround-Wave Propagation, Proceedingsof the IRE, ~0129, Jan 1941, pp. 16-24.Proceedings of the IRE, vol 29, Dee1941, pp 623-639.BURROWS, C. R. Radio Propagationover Plane Earth, Field Strength Curves,Bell System Technical Journal, vol 16,J a n 1 9 3 7 , p p 4 5 - 7 5 : Ott 1 9 3 7 ,pp 574-577.BURROWS, C. R., and GRAY, M. C.The Effect of the Earth’s Curvature onGround-Wave Propagation, Proceedingsof the IRE, ~0129, Jan 1941, pp 16-24.

136

BIBLIOGRAPHY

[I311

[1321

[I331

[I341

Cl351

[I361

11371

[I381

[I391

[I401

I1411

~421

JASIK, H. Antenna Engineeringbook, New York, McGraw-Hill,pp 33.1-33.24.

Hand- [143]1961,

CUNNINGHAM, J. E. The CompleteBroadcast Antenna Handbook-Design,Installation, Operation, and Mainte- [144]nance, Blue Ridge Summit, PA, TABBooks, 1977.Federal Communications CommissionRules and Regulations, vol III, Sept [145]1961, set 3.184; see also Dee 1963,set 73.184.WAIT, J. R., and CONDA, A. M.Pattern of an Antenna on a Curved [146]Lossy Surface, IRE Transactions onAntennas and Propagation, vol AP-6,Ott 1958, pp 348-359.MACDONALD, A. D. Microwave Break- [ 1471down in Gases, New York, Wiley, 1966.CHOWN, J. B., SCHARFMAN, W. E.,and MORITA, T. Voltage BreakdownCharacteristics of Microwave Antennas,Proceedings of the IRE, vol 47, Aug [148]1959, pp 1331-1337.WOO, R. Multipacting Discharges Be-tween Coaxial Electrodes, Journal ofApplied Physics, vol 39, Feb 1968,pp 528-533. [I491TAYLOR, W. C., SCHARFMAN, W. E.,and MORITA, T. Voltage Breakdownof Microwave Antennas, in Advancesin Microwaves, vo l 7 , New York , [150]Academic Press, 197 1.KUNKEL, W., Ed., Plasma Physics inTheory and Application, New York,McGraw-Hill, 1966, chap 10.VENUGOPALAN, M., Ed., Reactionsunder Plasma Conditions, New York, [I511Wiley-Interscience, 1971, chaps 5, G.MAYHAN, J . T . , F A N T E , R . L . ,O’KEEFE, R., ELKIN, R., KLUGER-MAN, J., and YOS, J. Comparison of [152]Various Microwave Breakdown Predic-tion M/:dels, Journal of Applied Physics,~0142, Dee 1971, pp 5362-5369.Special Issue on Biological Effects of [153]Microwaves, IEEE Transactions onMicrowave Theory and Techniques,vol M T T - 1 9 , F e b 1 9 7 1 , p p 128-253.

137

IEEEStd 149-1979

MICHAELSON, S. M. Human Exposureto Nonionizing Radiant Energy - Po-tential Hazards and Safety Standards,Proceedings of the IEEE, vol 60,Apr 1972, pp 389-421.TYLOR, P. E., Ed., Biologic Effectsof Nonionizing Radiation, Annals ofthe New York Academy of Sciences,~01247, Feb 1975.BARANSKI, S., and CZERSKI, P.Biological Effects of Microwaves,Stroudsburg, PA, Dowden, Hutchinsonand Ross, 1976.U.S. Air Force, Radiofrequency Radia-tion Health Hazards Control, AF Regu-lation 161-42, Washington, DC, Nov 7,1975.SCHWAN, H. P. Microwave Radiation:Biophysical Considerations and Stan-dards Criteria, IEEE Transactions onBiomedical Engineering, vol BME-19,July 1972, pp 304-312.Instrumentation for EnvironmentalMonitoring Radiation, EnvironmentalInstrument Group, Lawrence BerkeleyLab., University of California, May 1,1972.An Evaluation of Selected SatelliteCommunication Systems as Sources ofEnvironmental Microwave Radiation,EPA-52012. -74-008.WACKER, P. F., and BOWMAN, R.R.Quantifying Hazardous ElectromagneticFields: Scientific Basis and PracticalConsiderations, IEEE Transactions onMicrowave Theory and Techniques,vol MTT-19, Feb 1971, pp 178-187.CHU, T. S. and LEGG, W. E. Gain ofCorrugated Conical Horns, InternationalSymposium Digest, IEEE Antennas andPropagation Society, pp 427-430.KUMMER, W. H. and GILLESPIE, E. S.Antenna Measurements, 1978, Proceed-ings of the IEEE, vol 66, April 1978,pp 483-507.NEWELL, A. C., Improved PolarizationMeasurements Using a Modified ThreeAntenna Technique, IEEE AB-S Inter-na t iona l Symposium, pp 337-340,1975.

[ 1541 BEATTY, R. W., Microwave ImpedanceMeasurements and Standards, NationalBureau of Standards, Monograph 82,August 1965.

[155] HARRINGTON, R. F., Field Compu-tation by Moment Methods, New York,McGraw Hill, 1968.

[ 1561 Occupational Safety Standards, Electro-magnetic Fields of Radiofrequency,General Safety Requirements, GOST12.1.006-76, State Committee on Stan-dards of the Council of Ministers of theUSSR, Moscow, Jan 22,1976.

[157] STUCHLY, M. A., and REPACHOLI,M. H., Microwave and RadiofrequencyProtection Standards, Trmsac tions ofthe In terna tional Microwave PowerInstitute, vol 8, Microwave Bioeffectsand Radiation Safety, 1978, pp 95-101.

[158] BOWMAN, R. R., Quantifying Hazard-ous Microwave Fields, Transactions ofthe International Microwave PowerInstitute, vol 8, Microwave Bioeffectsand Radiation Safety, 1978, pp 113-128.

138

Appendixes

(These appendixes are not a part of IEEE Std 149-1979, Test Procedures for Antennas.)

Appendix AField Regions

Al. General

The distribution of field strength aroundan antenna is a function of the distancefrom the antenna. Three field regions aredistinguished to express antenna patternsand field distributions in zones surroundingthe antenna. In close proximity to the antennathe field strength may include, in addition tothe radiating field, a significant reactive (non-radiating) field. The strength of the reactive-field components decays rapidly with thedistance from the antenna (inversely as distanceraised to powers greater than unity). That regionof space immediately surrounding the antennain which the reactive components predominateis known as the reactive near-field region. Thesize of this region varies for different antennas.Formost antennas, however, the outer limit ison the order of a few wavelengths or less. Forthe particular case of an electrically smalldipole, as indicated in Fig Al(a), the reactivefield predominates out to a distance of approxi-mately X/27r, where the radiating and reactivefields are equal.

Beyond the reactive near-field region theradiating field predominates. The radiatingregion is divided into two subregions: theradiating near-field region and the far-fieldregion. The former region exists for mostelectrically large antennas, but not for elec-trically small antennas. The latter regionexists for all antennas.

In the radiating near-field region the relativeangular distribution of the field (the usualradiation pattern) is dependent on the distancefrom the antenna. There are two reasons forthis behavior.

(1) The relative phase relationship of fieldcontributions from different elements of the

antenna changes with distance(2) The relative amplitudes of these field

contributions also change with distance.For an antenna focused at infinity, as indi-cated in Fig Al(b), the radiating near-fieldregion is sometimes referred to as the Fresnelregion in analogy to optical terminology.

In the far-field region the relative angulardistribution of the field becomes essentiallyindependent of distance. Correspondingly, theamplitude of the field is given, in the limit, bythe reciprocal of the first power of distance.The reason for this behavior is that the relativephase and amplitude relationships betweenthe field contributions from different elementsof the antenna approach a fixed relationship.Although this situation is not attained preciselyuntil the observation point is an infinite distancefrom the antenna, the relative angular distribu-tion of the field at a comparatively shortdistance is often an adequate approximationof the field distribution at infinity. For anantenna focused at infinity, the far-field regionis sometimes referred to as the Fraunhoferregion in analogy to optical terminology.

For electrically large antennas of the broad-side-aperture type, such as that shown inFig Al(b), a commonly used criterion [Al],[A21 to define the distance in free space tothe boundary between the radiating near-fieldand the far-field regions is

20’R=--x

where D is the largest dimension of the aper-ture. The difference in path length betweenthe center and the edges of the aperture to thepoint at the region boundary is X/16. At thisboundary the antenna gain over most of the

139

IEEEStd 149-1979 APPENDIX A

(a)

RADIATINGNEAR-FIELDREGION

(b)

Fig AlField Regions for Two Antenna Types

(a) Electrically Small Dipole.(b) Electrically Large Reflector

major lobe of the radiation pattern differsfrom that at infinity by a very small factor,the exact value depending on the shape of theaperture and the illumination taper. However,for directions corresponding to minor lobesand pattern minima the gain at the regionboundary may differ from that at infinity bymany decibels (see 4.2).

For most applications it is the radiationpattern in the far-field region that is important.It is, therefore, customary to make measure-

ments m the far-field region. If the distancesinvolved are prohibitively great, measurementsin what would normally be the near-fieldregion can be used and the pattern in the far-field region inferred (see 7.2 and 7.3).

For electrically large antennas other thanconventional broadside-aperture types there isno recognized criterion defining the distanceto the far-field inner boundary. However, the2 D ’ /h criterion, w h e n D is taken as thelargest linear dimension, will usually give a

140

APPENDIX B

distance that is within the far-field region.Special precautions should be observed whenthe antenna’s environment plays a part in theformation of the radiation pattern. For such sit-uations the distance to the inner boundary ofthe far-field region is determined by the dimen-sion of the entire radiating structure. This struc-ture may involve a large metallic supportingsurface (for example, an aircraft fuselage) in afixed installation (see Section 9).

Certain measurement situations do not per-mit separation of the useful space into simplenear- and far-field regions. An important

IEEEStd 149-1979

example of this class is that of a verticallypolarized antenna operating over a ground offinite conductivity (see Section 17).

AZ. Bibliography

[Al] SILVER, S. Microwave Antenna Theoryand Design, M.I.T. Radiation Labora-tory ser, vol 12, New York, McGraw-Hill, 1949, set 6.9.

[A21 HANSEN, R. C. Microwave ScanningAntennas, vol 1, New York, AcademicPress, 1964, pp 30-46.

Appendix BReciprocity

Bl. General

The reciprocity principle [Bl] -[B3] is offundamental importance in the determinationof many antenna properties because, for areciprocal antenna, these properties may bedetermined from measurements with the an-tenna in either the transmitting or the receivingmode. An antenna is said to be reciprocal ifthe constitutive parameters of the transmittingmedia through the antenna can be characterizedby symmetric tensors. This generally meansthat there are no ferrite or plasma devices (withsteady magnetic fields applied) within theantenna or in the transmission medium.

For such antennas the reciprocity principlecan be used to produce relationships betweentheir transmitting and receiving properties.

The reciprocity principle as applied to anten-nas can be divided into three areas:

(1) antenna pattern characteristics and mea-surements

(2) gain measurements(3) polarization measurementsThe polarization characteristics of a recipro-

cal antenna are discussed in 11.1.If the antenna is not reciprocal, its trans-

mitting and receiving properties are not simplyrelated, and measurements shall be performedfor the mode or modes in which it is designedto be used.

A modified reciprocity principle applies tothe case of antennas containing magnetizedferrites, or to antennas in the presence ofmedia such as ionized gases in static magneticfields [B4]. The condition under which themodified reciprocity principle holds is that thestatic magnetic field shall be reversed when thetransmitting and receiving roles are inter-changed. When this condition is satisfied, allthe antenna properties mentioned are pre-served.

B2. Antenna Patterns

The pattern characteristics of an antennaunder test, called the test antenna, are obtainedby using the antenna in either of two ways.In the first way the test antenna is used in areceiving mode with the signal being trans-mitted by a source antenna. The signal isgenerated by a transmitter connected to theinput of the source antenna. Amplitude andphase patterns are recorded at the output of a

141

IEEEStd 149-1979 APPENDIX B

receiving system connected to the terminalsof the test antenna. The phase is referred to anarbitrary reference derived from the trans-mitting signal.

In the second way the transmitting source isconnected to the test antenna, and the receivingsystem is connected to the source antenna.The antennas are not disturbed in the inter-change of the transmitter and the receivingsystem. The amplitude and phase patternsmeasured by the receiving system will be thesame in both cases. The phase reference shallremain unchanged, otherwise the phase patternwill be shifted by some constant value ofphase.

NOTE: Strictly speaking, reciprocity applies to inter-changes between current sources and open-circuitvoltages or between voltage sources and short-circuitcurrents. These conditions can be satisfied by includingthe transmitter impedance as a part of the transmittingantenna and the load impedance as a part of the receiv-ing antenna. However, when a physical source and areceiver are interchanged, their impedances shall movewith them. Then unless these impedances are equal,the reciprocity condition is not satisfied. If theseimpedances are equal, reciprocity is satisfied for anyseparation between test and source antennas. However,if the antennas are separated sufficiently far so thatmultiple interactions are negligible, the amplitudepatterns will be proportional and the phase patternswill differ only by a constant phase when the trans-mitter and the receiver are interchanged.

B3. Gain and Effective Area

Another antenna property which followsfrom the reciprocity principle is that the powergain G and the effective area A, of a reciprocalantenna are related by the equation [B3]

47rA,G=--h2

(Eq 1)

This latter property is often interpreted as: Thepower gain of a reciprocal antenna is the samewhen used for transmitting as it is when usedfor receiving. Strictly speaking, power gain isonly defined for the transmit case. For thereceive case the effective area is the appropriatequantity to be used.

B4. Expanded Reciprocity Relations [ B5]

Applied to an antenna, the reciprocityprinciple implies that the far field radiated ina given direction and a given polarization, fora specified excitation, is proportional to theresponse of the antenna when a plane waveof opposite direction and of matched polariza-tion is incident upon it.

NOTE: The incident plane wave is said to be polari-zation matched to the antenna if, in the same plane ofpolarization, their polarization ellipses have the sameorientation, the same axial ratio, and the same senseof rotation. However, because of the usual conventionin specifying the orientation of the ellipses, the tiltangles will be different (see 11.1).

The factor of proportionality depends onhow the excitation and the response are ex-pressed. Both shall be measured at the samepoint, or port, in the transmission line con-nected to the antenna. Furthermore, a refer-ence point 0 shall be specified in the vicinity ofthe antenna to make-the phase of the incidentplane wave and that of the radiated far-fieldunambiguous. A direction shall be specified bya wavevector k acd an associated polarizationby a unit vector u pzrpendicular to k(i - k = 0,

1 u) = 1). In general u is a complex vector.If the excitation is by a source that is matched

to the transmission line and produces a travel-ing wave of unit amplitude (that is, a wave thatcarries unit power and has zero phase at theantenna port), then the far-field in direction kand polarization i, referred to the origin 0,will have the form

e-jkrE(r;k,u^)=l;*.E(iTr)=flo Afk,+-

(Eq 2)

where E is the unit vector in the direction of k,r is the distance from 0 to the observationpoint Err, and 2, is the intrinsic impedance ofspace. Conversely, let a plane wave of wave-

142

f1

APPENDIX B

vector -k and electric field c * 42, be incidentupon the antenna. Note that u^* incidentrepresents the polarization Fhich is matched tothe antenna’s polarization u (see 11.1) and be-cause of the factor 6, this wave has unitintensity. Its Poynting vector is -k. This inci-dent field produces, at the antenna port, awave of amplitude a traveling toward thematched load. Reciprocity is expressed by therelation

a (-k, i*) = -jXA (k, s) (Eq 3)

valid for any k and i vectors. Equation 3 holdswhether the transmission line is matched to theantenna or not.

If the antenna is excited by a unit currentsource of zero internal admittance, the far-fieldin direction k, polarization 4 at a distance ris expressed by

h (k ;)s-jkrr (Eq 4)

where h is the effective length for directionvector k and polarization vector u^. In general his a complex number. For an incident planewave of wave vector -k and electric field vectori* at point 0, the open-circuit voltage V (alsocomplex) is, according to reciprocity,

V (-k, t;*) = h (k, u^) 0% 5)

The electric far-field radiated in direction k‘ canbe expressed by the vector

E (zr) = jh ZO h (k)g (Eq 6)

The complex vector h (effect ive currentmoment per unit current) indicates the polari-zation, has the dimensions of a length, and canbe written as a product h U of a real effectivelength h and a unit magnitude vector U. Thiseffect lengt*h, which is not to be confusedwith h (k, u), is a function of vector k only. If

IEEEStd 149-1979

an incident plane wave in direction -k has anelectric field vector U* at reference point 0,the resulting open-circuit voltage is precisely hvolts. It therefore is of zero phase. It is alsolarger in magnitude than the voltage that anyother plane wave of unit intensity wouldproduce. Reciprocity can therefore be expressedby saying that the scalar r_eal effective length h(a function of direction k ) is the same whentransmitting and when receiving. For trans-mitting, h is such that the current moment ofmagnitude h 111 when the excitation currentis I and for receiving a plane wave of matchedpolarization the open-circuit voltage is ofmagnitude h 1 E 1 if E is the electric field. Ifthe phase is important, one shall carefullyspecify the reference port and the referencepoint 0 and use the precise Eqs 3 or 5.

[Bll

F321

LJ331

[B41

IJ351

B5. Bibliography

JORDAN, E. C., and BALMAIN, K. G.Electromagnetic Waves and RadiatingSystems, Englewood Cliffs, NJ, Prentice-Hall, 1968, chap 11.

COLLIN, R. E., and ZUCKER, F. J.Antenna Theory, pt 1, New York,McGraw-Hill, 1969, chap 4.

SESHADRI, S. R. Fundamentals ofTransmission Lines and Electromag-netic Fields, Reading, MA, Addison-Wesley, 1971, chap 10.

HARRINGTON, R. F., and VILLE-NEUVE, A. T. Reciproci ty Rela-tionships for Gyrotropic Media, I R ETransactions on Microwave Theory andTechniques, vol MTT-6, July 1958,pp 308-310.

DESCHAMPS, G. E. I -Principe deReciprocite en Electromagnetisme; II -Application du Principe de Reciprociteaux Antennes et aux Guides d’ondes,Cethedec no. 8,4 e t r imes t re 1966,pp 91-101.

143

Date post:	03-Sep-2018
Category:	Documents
Upload:	dotruc
View:	259 times
Download:	4 times

IEEE Standard Test Procedures for Antennas · IEEE Std 149-1979 (Revision of IEEE Std 149-1965)...

Documents