Antenna Measurements, Testing, and Diagnostics

Introduction and Context

Antenna theory predicts quantities such as radiation pattern, gain, impedance, polarization, efficiency, and bandwidth. Antenna measurement determines whether a manufactured or installed antenna actually achieves those predictions. It is therefore the bridge between electromagnetic design, production engineering, regulatory compliance, and field performance.

Measurement is not simply the act of connecting an antenna to a vector network analyzer (VNA). The measured result is produced by the complete measurement chain: the antenna under test (AUT), test environment, reference antenna, cables, instrumentation, positioning system, calibration, data-processing algorithms, and uncertainty. At postgraduate level, antenna testing must be treated as an inverse problem: the observable voltage or scattering parameter is interpreted to estimate physical antenna characteristics.

The central distinction is between verification, which checks conformance with a specification, and diagnosis, which identifies the physical cause of nonconformance. A failed gain result may arise from a damaged radiator, incorrect calibration, cable movement, chamber reflection, polarization mismatch, or an error in the AUT coordinate system. Reliable diagnosis consequently requires physical insight as well as measurement skill.

Learning Objectives

After completing this chapter, the learner should be able to:

1. Explain the purposes and limitations of antenna measurements.
2. Select appropriate far-field, near-field, compact-range, and over-the-air test methods.
3. Measure and interpret S-parameters, impedance, radiation patterns, gain, efficiency, and polarization.
4. Apply calibration, error correction, data reduction, and uncertainty concepts.
5. Diagnose measurement anomalies using physically justified tests.
6. Design a technically defensible antenna test plan for research or production.

1. Measurement Quantities and Their Relationships

1.1 Input impedance and reflection coefficient

The input impedance of an antenna is

where V_in and I_in are the terminal voltage and current, R_in is resistance, and X_in is reactance, all evaluated at a specified frequency and reference plane. The resistance includes radiation resistance and loss resistance, while reactance represents stored electric and magnetic energy.

For a transmission line with characteristic impedance Z₀, the reflection coefficient is

For a real reference impedance Z₀, commonly 50 Ω, the voltage standing-wave ratio is

These equations assume a single-port, linear, time-invariant antenna and a known reference plane. Return loss is

A small |Γ| indicates good matching, but matching alone does not establish good radiation performance. A lossy antenna can have excellent matching while radiating little power.

1.2 S-parameters and calibration planes

A VNA measures travelling-wave ratios. For a one-port antenna, S₁₁ is the input reflection coefficient at the calibrated reference plane. For a two-port device, S₂₁ represents forward transmission from port 1 to port 2. Calibration removes systematic errors such as directivity, source match, reflection tracking, and transmission tracking.

The reference plane may be moved electronically using port extension or physically by calibration at the connector. Cable flexure after calibration changes phase and may change amplitude; this is especially serious in pattern and near-field measurements.

1.3 Radiation pattern

The radiation intensity U(θ,φ) is power radiated per unit solid angle. In the far field,

where S_r is the radial time-average Poynting-vector density and r is distance. The normalized power pattern is

The half-power beamwidth (HPBW) is the angular separation between points at which the power is one-half of its maximum, or −3 dB. First-null beamwidth is measured between the nearest nulls surrounding the main beam. Sidelobe level is normally stated relative to the main-beam maximum in dB.

Patterns must specify coordinate convention, polarization, frequency, angular step, cut plane, and whether the result is field magnitude, power, or gain. A plot without these metadata is not reproducible.

1.4 Gain, directivity, and efficiency

Directivity is the ratio of radiation intensity in a specified direction to the average radiation intensity:

where P_rad is total radiated power. Maximum directivity is D₀. Gain includes radiation and loss effects:

where P_in is accepted input power under the stated reference conditions. Radiation efficiency is

and, when mismatch is separated from internal antenna losses,

with η_mismatch = 1 − |Γ|² for a single port and a real reference impedance. In this convention, realized gain is related to gain by

provided G is defined relative to accepted power. Terminology must be checked against the applicable standard because manufacturers sometimes use “gain” and “realized gain” inconsistently.

The Friis transmission equation is useful for gain comparison:

where P_t and P_r are transmitted and received powers, G_t and G_r are gains in the link direction, λ is wavelength, and R is separation. It assumes free-space far-field propagation, matched polarization, aligned boresights, and negligible multipath.

1.5 Polarization

Polarization describes the time-varying orientation and magnitude of the electric-field vector. Important measured quantities include co-polarized response, cross-polarized response, axial ratio, and polarization discrimination. For an elliptically polarized field with orthogonal components E_x and E_y and phase difference δ, the polarization ellipse can be determined from their complex amplitudes. For circular polarization, equal magnitudes and a ±90° phase difference are required.

Cross-polarization can be caused by the AUT itself, the probe, mechanical misalignment, chamber scattering, or coordinate transformation errors. Therefore, a low cross-polar level must not automatically be interpreted as intrinsic antenna purity.

2. Test Environments and Measurement Regions

2.1 Far-field testing

A conventional outdoor range or anechoic chamber measures the AUT in the radiating far field. A commonly used engineering criterion is

where D is the largest AUT dimension and λ is wavelength. This criterion ensures that phase variation across the aperture is sufficiently small for plane-wave interpretation. It is a practical criterion, not a universal boundary; electrically large or precision applications may require a stricter distance.

The chamber should suppress reflections from walls, floor, ceiling, supports, and feed cables. Quiet-zone evaluation normally includes amplitude ripple, phase ripple, and cross-polarization. Absorber performance is frequency-dependent, so a chamber satisfactory at 10 GHz may be inadequate at 1 GHz.

2.2 Near-field measurement

Near-field systems measure amplitude and phase over a surface close to the AUT, then mathematically transform the measured field into a far-field pattern. Common geometries are planar, cylindrical, and spherical.

A planar scanner is efficient for pencil-beam antennas with a principal boresight normal to the scan plane. A cylindrical scanner is appropriate when azimuthal coverage is important. A spherical scanner provides complete angular information but requires more samples and careful probe correction.

Near-field transformation assumes accurate probe positioning, adequate spatial sampling, stable phase, known probe response, and an AUT that does not move during scanning. Probe correction is essential because the probe samples a field through its own receiving pattern rather than measuring an ideal point field.

2.3 Compact antenna test range

A compact range uses a reflector system to convert a spherical wave from a feed into a quasi-plane wave in a quiet zone. It can provide far-field-like testing in a laboratory-sized facility. Its limitations include reflector-edge diffraction, feed blockage, quiet-zone nonuniformity, and frequency-dependent performance.

2.4 Over-the-air testing

Over-the-air (OTA) testing evaluates complete wireless devices, including antennas, enclosures, radios, and signal-processing functions. It may measure total radiated power, total isotropic sensitivity, throughput, beam-steering accuracy, and user-equipment performance. OTA results are system-level quantities and should not be confused with isolated antenna gain.

3. Measurement Systems and Calibration

A typical test system includes a signal source or VNA, receiver, AUT positioner, reference antenna or probe, cables, amplifiers, attenuators, timing and control software, and environmental monitoring. The system must be treated as a calibrated measurement chain.

3.1 One-port VNA calibration

SOLT calibration uses known standards: short, open, load, and through. It is convenient at coaxial connectors. TRL calibration uses a through, reflect, and line standard and is often preferable for fixtures, waveguides, and broadband structures where accurate standards are difficult to realize. Calibration should be performed at the measurement reference plane and verified using an independent check standard.

Calibration does not remove random noise, chamber reflections that vary with position, or errors introduced after calibration. A cable should be immobilized, and connectors should be clean, correctly torqued, and inspected.

3.2 Pattern measurement procedure

A defensible pattern measurement normally follows these steps:

1. Define frequency, polarization, coordinate system, angular range, and required dynamic range.
2. Verify the chamber quiet zone and position the AUT at the mechanical phase center as closely as practical.
3. Align the AUT and probe, documenting boresight and polarization axes.
4. Calibrate the VNA and characterize cable stability.
5. Measure a reference or probe response where required.
6. Scan with an angular step small enough to capture the narrowest expected feature.
7. Repeat selected angles or a complete cut to assess repeatability.
8. Apply probe, cable, mismatch, and range corrections with traceable software.
9. Report raw and corrected results, uncertainty, and configuration metadata.

For a sinusoidal spatial variation with maximum spatial frequency k_x,max, a sampling interval Δx should satisfy the Nyquist condition Δx ≤ π/k_x,max. In practice, margins are used because truncation, probe size, evanescent fields, and interpolation affect the transformation. Angular step selection should also resolve sidelobes and nulls, not merely the main beam.

4. Gain and Efficiency Measurements

4.1 Gain comparison method

In the gain comparison method, the AUT replaces a calibrated standard-gain antenna while transmitter, receiver, distance, polarization, and alignment remain unchanged. If received powers are P_AUT and P_ref, and the reference gain is G_ref, then under identical conditions:

This expression assumes the power ratio is corrected for mismatch, cable loss, and any change in polarization or alignment. If the power meter reads voltage rather than power, the conversion must be handled correctly; equal impedances imply power proportional to voltage squared, not voltage itself.

4.2 Three-antenna method

The three-antenna method determines absolute gain without requiring one antenna to be a calibrated standard. Three pairwise measurements are made at the same frequency and geometry. In logarithmic form, the three unknown gains can be solved from the three link equations. The method assumes reciprocal antennas, stable instrumentation, far-field conditions, and adequate isolation. It is useful for calibration laboratories but becomes sensitive to alignment and chamber errors when the three antennas have substantially different patterns.

4.3 Wheeler cap and reverberation approaches

A Wheeler cap estimates radiation efficiency by comparing accepted power in free space with accepted power when radiation is suppressed by a conductive enclosure. The method is convenient for electrically small antennas but requires a cap that suppresses radiation without significantly changing internal loss or stored-energy conditions.

Reverberation chambers produce a statistically rich multipath field and can estimate total radiated power and efficiency efficiently for wireless devices. They do not directly provide the directional pattern obtained from a traditional anechoic range. Chamber stirring, loading, calibration, and statistical confidence must be considered.

5. Measurement Uncertainty and Data Quality

A result should be reported as an estimate with an uncertainty, not as an infinitely precise number. An uncertainty budget identifies contributors such as:

• VNA amplitude and phase accuracy;
• reference-antenna calibration;
• positioning and angular encoder errors;
• cable drift and connector repeatability;
• chamber reflections and quiet-zone ripple;
• mismatch and polarization errors;
• temperature, humidity, and mechanical vibration;
• data interpolation, truncation, and transformation errors.

For independent standard uncertainties u_i with sensitivity coefficients c_i, the combined standard uncertainty is commonly estimated as

If the measurand is a direct sum or difference, c_i is often ±1; for a nonlinear model, sensitivity coefficients are obtained from partial derivatives. An expanded uncertainty is U = k u_c, where k is a coverage factor selected according to the desired coverage probability and effective degrees of freedom. Contributors that are correlated must not be combined as independent terms.

A useful practice is to separate repeatability—variation under nominally identical conditions—from reproducibility, which includes changes in operators, setups, days, or laboratories. Outlier rejection must be justified physically; deleting an inconvenient point is not uncertainty analysis.

6. Diagnostics: From Anomaly to Cause

Diagnostics begins by classifying the symptom. A broad, physically informed fault tree is more useful than repeatedly rerunning the same measurement.

A diagnostic measurement should change one factor at a time. For example, moving the AUT by a small known amount can reveal whether a pattern ripple is a chamber standing-wave effect. Measuring both raw and corrected data can identify an inappropriate correction algorithm. Near-field data may be transformed with and without probe correction to test whether unexpected sidelobes are probe-generated.

6.1 Time-domain gating

A VNA can transform frequency-domain data to time-domain information. Reflections associated with delayed paths can then be isolated or suppressed using a time gate. If a reflection arrives after delay τ, its path-length difference is approximately cτ, where c is wave velocity in the relevant medium. Gating can reduce multipath influence, but it cannot recover information lost through insufficient frequency span, poor windowing, or overlapping reflections. Gated results must therefore be reported as processed results, not as direct raw observations.

6.2 Near-field diagnostics and holographic insight

Near-field phase and amplitude contain spatial information about aperture illumination. A localized defect, failed element, or phase-shifter error often appears as a localized disturbance in the aperture field. This makes near-field measurement valuable not only for far-field prediction but also for manufacturing diagnosis and active-array maintenance.

7. Active Arrays and Advanced Measurements

Phased arrays require measurements beyond a single passive pattern. Important tests include element-level impedance, embedded element pattern, active reflection coefficient, scan blindness, beam-pointing accuracy, sidelobe control, polarization purity, calibration drift, and thermal behavior.

The active reflection coefficient of an element depends on the excitation of all other elements because mutual coupling changes the terminal environment. Consequently, a low isolated-element |S₁₁| does not guarantee a low active reflection coefficient in a scanned array.

Array calibration may estimate complex excitation errors, mutual coupling, phase-shifter offsets, receiver gain mismatch, and element failures. Calibration algorithms should be evaluated for identifiability: multiple error mechanisms may produce similar far-field symptoms. Combining near-field, embedded-element, and built-in-test data improves diagnostic confidence.

8. Worked Example: Far-Field Feasibility and Gain Comparison

An AUT has maximum dimension D = 0.60 m and is tested at 10 GHz. The wavelength is

Using the engineering far-field criterion,

Thus, a 24 m separation is the minimum indicated by this criterion; a range should normally provide additional margin and verify its quiet-zone performance.

Suppose a calibrated reference antenna has gain 12 dBi. With the reference installed, received power is −40 dBm; with the AUT installed, it is −43 dBm. The power ratio is 10^(−3/10) = 0.501. Therefore,

This estimate is valid only if distance, polarization, alignment, cable configuration, mismatch, and frequency are held constant or correctly corrected. If the AUT has a significant mismatch, the result should be identified as gain or realized gain according to the adopted measurement convention.

9. Illustrative Case: Diagnosing a Low-Gain Patch Array

Background

A four-element 5.8 GHz patch array is predicted to provide 14 dBi gain. A production test reports 10.8 dBi, while the measured main-beam direction is correct but the sidelobes are asymmetric.

Problem

The result could indicate element detuning, unequal feed amplitudes, phase error, connector damage, chamber reflection, or a faulty gain reference. Immediate redesign would be premature.

Analysis

First, the one-port measurement is repeated after a fresh SOLT calibration. One element shows a shifted resonance and higher loss. Inspection finds a partially lifted feed via. However, the 3.2 dB gain deficit is larger than expected from one failed element alone. A reference antenna measurement then reveals a 1.1 dB discrepancy caused by an incorrect stored calibration file.

The remaining asymmetry is investigated by repeating the pattern with the AUT rotated by 180°. The asymmetry rotates with the AUT, proving that it is not fixed in chamber coordinates. Near-field scanning shows a phase error in the feed network near the damaged via.

Approach and solution

The via is repaired, the reference calibration file is corrected, and the array is remeasured. The pattern becomes symmetric, and gain rises to 13.6 dBi. The residual difference from 14 dBi is attributed to production dielectric variation and connector repeatability, supported by measurements from several conforming units.

Learning points

1. A pattern symptom should be tested in both AUT and chamber coordinates.
2. Absolute gain requires independent verification of the reference chain.
3. S-parameters, near-field data, and far-field data provide complementary evidence.
4. A measurement process should distinguish product failure from test-system failure.

10. Standards, Governance, and Professional Practice

A test plan should identify the governing standard, frequency range, polarization convention, uncertainty method, calibration interval, acceptance limits, and data-retention policy. Commonly consulted families include IEEE antenna measurement standards, CISPR and regulatory emission requirements, IEC procedures, and national spectrum-authority rules. The exact edition and applicability must be stated because terminology and procedures evolve.

Traceability requires documented standards, calibration certificates, equipment serial numbers, software versions, environmental conditions, operator identity, and raw data. Automated test software should preserve an audit trail and prevent silent changes to correction files. AI may assist with anomaly detection, surrogate modelling, or classification of array faults, but it must not replace calibrated measurements. Training data, false-alarm rates, explainability, and human approval are important when AI influences compliance decisions.

Safety is also part of measurement quality. High-power transmitters create RF exposure, burns, arcing, and equipment hazards. Interlocks, shielding, grounding, lockout procedures, and controlled access are essential. Export controls, privacy considerations for device OTA data, and responsible reporting of nonconforming results may also apply.

11. Contemporary Developments and Best Practices

Modern laboratories increasingly combine robotic positioners, broadband VNAs, real-time receivers, digital twins, near-field scanning, reverberation chambers, and automated uncertainty budgets. Massive MIMO and millimetre-wave systems require fast electronic scanning, over-the-air characterization, and calibration under thermal and environmental stress.

Best practice is to:

• define the measurand before selecting equipment;
• validate the chamber with a known antenna or quiet-zone survey;
• calibrate at the actual reference plane;
• control cables, connectors, temperature, and mechanical alignment;
• measure a reference before and after a test batch;
• preserve raw data as well as corrected data;
• include uncertainty and repeatability in acceptance decisions;
• use independent cross-checks for surprising results;
• report coordinate systems, polarization, frequency, distance, and processing;
• separate measurement evidence from engineering interpretation.

Chapter Summary

Antenna measurement converts electromagnetic behaviour into traceable engineering evidence. Input impedance and S-parameters describe terminal interaction; patterns and polarization describe spatial radiation; gain, directivity, and efficiency describe power performance. Far-field, near-field, compact-range, and OTA methods serve different purposes and have different assumptions. Calibration establishes a trustworthy reference plane, while uncertainty analysis expresses the limits of that trust. Diagnostics requires controlled experiments that distinguish AUT faults from chamber, instrumentation, alignment, and processing errors. For arrays and modern wireless devices, embedded patterns, active reflection, near-field data, and system-level OTA metrics are increasingly important. Professional practice combines standards, safety, data governance, repeatability, and independent verification.

Key Terms

Antenna under test (AUT): The antenna being characterized.

Anechoic chamber: A shielded room designed to suppress electromagnetic reflections.

Axial ratio: Ratio of the major to minor axes of a polarization ellipse.

Calibration plane: The physical or mathematical reference location at which measured quantities are defined.

Compact range: A reflector-based facility that creates a quasi-plane-wave quiet zone.

Directivity: Ratio of radiation intensity in a direction to average radiation intensity.

Far-field region: Region in which the radiated field has approximately plane-wave angular behaviour.

Gain: Directional radiation performance including antenna loss, according to a stated convention.

HPBW: Angular width between the −3 dB points of a main beam.

Near-field measurement: Measurement close to an antenna followed by field transformation.

Probe correction: Compensation for the receiving probe’s nonideal response.

Radiation efficiency: Radiated power divided by accepted input power.

Realized gain: Gain including mismatch loss.

Reflection coefficient: Complex ratio of reflected to incident travelling-wave amplitude.

S-parameter: Scattering parameter describing travelling-wave relationships at ports.

Sidelobe level: Relative level of a sidelobe, usually expressed in dB.

Quiet zone: Region of a test range meeting specified field uniformity requirements.

Uncertainty budget: Structured evaluation of contributions to measurement uncertainty.

VSWR: Ratio of maximum to minimum standing-wave voltage.

Review Questions

Short Answer Questions

1. Why does good impedance matching not necessarily imply high antenna efficiency?
2. State the practical far-field distance criterion and define its variables.
3. What information does S₁₁ provide for a one-port antenna?
4. Distinguish repeatability from reproducibility.
5. Why is probe correction needed in near-field measurements?

Long Answer Questions

1. Explain the operating principles, advantages, and limitations of planar, cylindrical, and spherical near-field measurements.
2. Discuss the calibration and verification steps required for a reliable VNA-based antenna test.
3. Derive the gain-comparison relationship from the Friis transmission equation and state its assumptions.
4. Explain how measurement uncertainty should be constructed for an antenna gain measurement.
5. Discuss the additional measurements required for phased-array characterization compared with a passive single antenna.

Analytical / Application Questions

1. An antenna has D = 0.40 m and operates at 6 GHz. Calculate the minimum far-field distance using R_ff ≥ 2D²/λ, and discuss whether a 10 m chamber should be accepted without a quiet-zone survey.
2. A pattern has the correct beamwidth but a 4 dB lower absolute gain than expected. Develop a diagnostic sequence that separates AUT loss, cable/reference error, mismatch, and chamber effects.
3. An array’s cross-polarization is 8 dB worse than predicted only when scanned to 45°. Propose measurements that distinguish mutual coupling, scan blindness, polarization-coordinate error, and chamber multipath.

Multiple Choice Questions with Answers

Question 1

For a one-port antenna referenced to a real impedance Z₀, which expression gives the reflection coefficient?

A. Γ = (Z_in + Z₀)/(Z_in − Z₀)

B. Γ = (Z_in − Z₀)/(Z_in + Z₀)

C. Γ = Z_in Z₀

D. Γ = Z₀/Z_in

Correct Answer: B. Γ = (Z_in − Z₀)/(Z_in + Z₀)

Question 2

An antenna has maximum dimension 0.30 m and operates at 3 GHz. Using R_ff ≥ 2D²/λ, what is the minimum far-field distance?

A. 0.18 m

B. 1.8 m

C. 18 m

D. 180 m

Correct Answer: B. 1.8 m

Question 3

Which quantity is most directly measured by a one-port VNA test of an antenna?

A. S₁₁

B. Axial ratio only

C. Radiation efficiency only

D. Far-field sidelobe level

Correct Answer: A. S₁₁

Question 4

A measured gain is uniformly low at all angles. Which cause is most consistent with this symptom?

A. A coordinate rotation of the AUT only

B. A cable-loss or reference-gain error

C. A narrow angular sampling step

D. A change in sidelobe phase only

Correct Answer: B. A cable-loss or reference-gain error

Question 5

The half-power beamwidth is defined by the angular separation between points at which power is:

A. Equal to the null level

B. One-half of maximum, or −3 dB

C. One-tenth of maximum, or −20 dB

D. Twice the maximum

Correct Answer: B. One-half of maximum, or −3 dB

Question 6

Which method uses three pairwise antenna measurements to determine antenna gains without requiring a calibrated standard-gain antenna?

A. Three-antenna method

B. Wheeler-cap method

C. SOLT calibration

D. Time-domain gating

Correct Answer: A. Three-antenna method

Question 7

If |Γ| = 0.333, the approximate VSWR is:

A. 1.5:1

B. 2:1

C. 3:1

D. 4:1

Correct Answer: B. 2:1

Question 8

What is the principal purpose of a near-field-to-far-field transformation?

A. To remove all antenna losses physically

B. To infer far-field behaviour from measured close-range fields

C. To convert S₁₁ directly into efficiency without assumptions

D. To eliminate the need for spatial sampling

Correct Answer: B. To infer far-field behaviour from measured close-range fields

Question 9

Which statement about realized gain is correct?

A. It excludes mismatch loss by definition

B. It includes mismatch loss under the stated convention

C. It is always equal to directivity

D. It is independent of reflection coefficient

Correct Answer: B. It includes mismatch loss under the stated convention

Question 10

In an uncertainty budget, independent standard uncertainties are commonly combined using:

A. Simple arithmetic addition in all cases

B. The root-sum-square method

C. The largest value only

D. The product of all uncertainties

Correct Answer: B. The root-sum-square method