RF Antenna Engineering

How to Select an RF Antenna: Frequency, Gain, Pattern, Polarization, Match, Power and Installation

A practical antenna-selection method that turns operating band, angular coverage, gain pattern, polarization, port match, waveform power, mounting environment and measurement evidence into an RFQ and acceptance plan.

Published: Jul 14, 2026
Reading time: 10 min

People-free microwave antenna, radome, precision rotary positioner and RF feed hardware in an unattended antenna verification range

Define the installed radiation requirement before comparing antennas

The correct antenna covers the required directions, rejects the directions that matter, preserves polarization and match, accepts the real waveform and still meets its limits in the installed configuration. Start with a coordinate system and angular coverage map. Add continuous frequency coverage, port and reference-plane definitions, gain and pattern envelopes, polarization, accepted power, mechanical interfaces and environmental states. Only then compare antenna families, because maximum gain and a single VSWR value do not describe the complete radiating boundary.

Begin with spatial coverage and an installed coordinate system

Write the service volume in azimuth and elevation before naming an antenna type. A point-to-point link may need a narrow main beam and controlled sidelobes; a mobile platform may need coverage through a large elevation range; a navigation receiver may need an upper-hemisphere pattern while rejecting low-elevation interference; an array may need a scan volume rather than one boresight direction. Mark regions that must be served, regions where gain may fall, and regions that require a null, sidelobe ceiling or front-to-back limit.

Define zero angle, positive rotation, antenna boresight, platform axes, polarization basis and installed orientation. State whether the pattern is required in principal cuts or over a three-dimensional grid and whether values are realized gain, gain, directivity or normalized pattern. Without those definitions, two technically correct plots can describe different quantities and cannot be used for acceptance. Mechanical datums and the RF coordinate system should be traceable to the same drawing.

Freeze continuous band coverage and the conducted reference plane

List every transmit and receive range, guard band and operating state as continuous intervals. A nominal center frequency does not reveal a notch, split band or scan-state limitation. If one antenna serves several services or ports, identify simultaneous operation, diplexing, bias networks, integrated filters and any band-specific cable or transition. The declared antenna band is the range over which all required limits remain valid, not merely the range where some energy is radiated.

Name the plane where impedance, return loss, VSWR and accepted power apply: connector face, waveguide flange, cable end, board launch or antenna element feed. State which adapter, feed line, balun, matching network, bias tee and transition losses are included. Return loss and VSWR describe reflection at that plane; they do not prove radiation efficiency, pattern or polarization. Preserve complex S-parameters when phase or multiport coupling matters.

Specify a gain-pattern envelope rather than one maximum-gain number

Maximum gain is one point on one pattern at one frequency and polarization. A usable specification may also need minimum gain inside the service region, ripple, half-power beamwidth, first-null beamwidth, sidelobe envelope, null depth, front-to-back ratio and cross-polar level. Apply each limit over frequency and, for arrays, over scan angle. A broad coverage requirement can conflict with high peak gain, so the system must decide which angular regions carry the link or sensing budget.

Directivity describes how concentrated the radiation is relative to an isotropic source; gain also includes dissipative efficiency. Realized gain additionally includes mismatch at the declared port. These quantities are related but not interchangeable. If a proposal reports only directivity, the accepted power and radiated result remain unknown until conductor, dielectric, matching and feed losses are accounted for. Ask for the quantity, polarization, frequency, coordinate convention and measurement method with every pattern file.

Treat polarization, ports and array behavior as first-class requirements

For linear polarization, define orientation and allowable cross-polar response over the service angles. For circular polarization, define right- or left-hand sense using the agreed viewing convention and set axial-ratio limits over frequency and angle. A circularly polarized label at boresight does not guarantee useful polarization at the edge of coverage. Polarization mismatch belongs in the system budget and should not be hidden inside a generic margin.

Dual-polarized, diversity, MIMO and array antennas add port isolation, envelope correlation, embedded-element pattern, amplitude and phase balance, scan loss and coupling. Specify whether patterns are measured with unused ports terminated and whether active elements are driven individually or as a beam state. For a calibrated array or navigation antenna, phase center and group-delay behavior may matter as much as peak gain.

Translate the real waveform and environment into accepted port stress

State CW, average and peak power, pulse width, duty cycle, modulation crest factor, number of simultaneous carriers, source and load mismatch, reverse power and temperature at each port. A single power rating cannot represent heating, electric-field peaks and breakdown risk under every waveform. High-power limits can depend on frequency, geometry, spacing, pressure, contamination, connector condition and mismatch; pulse and altitude conditions therefore belong in the RFQ and test plan.

Separate electrical survival from performance retention. An antenna may survive a high-power pulse yet detune, heat, generate passive intermodulation or change polarization. Environmental requirements can include temperature, altitude, pressure, humidity, salt, icing, wind, vibration, shock, ingress, vacuum or solar loading. Define which RF parameters are checked during exposure and which are remeasured afterward.

Make the platform, radome and nearby structures part of the antenna configuration

Ground plane size, mast, vehicle roof, fuselage, enclosure, cable routing, fasteners and nearby conductive equipment can alter match, pattern, efficiency, phase center and cross-polarization. A radome can add loss, reflection, beam shift and polarization error, especially as frequency, incidence angle, moisture or temperature changes. If the delivered system uses a radome, acceptance evidence should represent that radome, its spacing, material state and attachment.

Freeze mounting datums, keep-out zones, cable exit, connector torque, grounding, bonding, drainage and allowable installation tolerances. Decide whether supplier acceptance is for the stand-alone antenna, an agreed representative ground plane, a supplied radome assembly or the final platform. When the final platform cannot be placed in the range, define a correlation plan using simulation, representative fixtures, near-field scanning or installed check measurements.

Design the measurement and evidence package before purchase

Conducted tests can establish port match, isolation and feed-network behavior, but radiated gain and pattern require an over-the-air method. State whether substitution gain, comparison gain, near-field transformation, compact-range or far-field measurement is acceptable. Control range reflections, quiet-zone quality, alignment, cable movement, positioner accuracy, receiver linearity, dynamic range, reference-antenna calibration and mismatch. The uncertainty statement must match the frequency, polarization and angular metric being accepted.

Request native pattern data rather than screenshots alone. Preserve frequency, polarization, port state, beam state, angle convention, angular step, units, normalization, calibration identity, uncertainty, fixture, radome and environmental condition. Tie files to serial number and revision. A plotted curve without this metadata is difficult to compare, reproduce or use in a system model after the hardware changes.

Decision boundary	Requirement to freeze	Reject the proposal when
Coverage and coordinates	Installed axes, service and exclusion angles, pattern cuts or 3D grid	Only an unreferenced beamwidth is supplied
Band and ports	Continuous ranges, port states, impedance and named conducted plane	Only a center frequency or connector name is shown
Gain and pattern	Realized-gain envelope, ripple, sidelobes, nulls and front-to-back limits	One typical maximum-gain value represents the antenna
Polarization	Orientation or sense, axial ratio, cross-pol, port isolation and angle range	The proposal says only linear or circular
Power and environment	CW, average, peak, pulse, duty, mismatch, pressure, temperature and exposure	A power number has no waveform or environmental boundary
Installed configuration	Ground plane, platform, radome, spacing, cable route, datum and tolerance	Free-space data is presented as installed performance
Acceptance evidence	Method, calibration, uncertainty, raw data, metadata and revision traceability	Only a plot image or simulated peak is offered

Worked check: a 2:1 VSWR is not a complete antenna efficiency result

For a 2.0:1 VSWR at the declared port, the reflection-coefficient magnitude is (2-1)/(2+1) = 0.333. Reflected power is therefore about 11.1%, equivalent to a return loss of about 9.54 dB and a mismatch loss of about 0.51 dB. That arithmetic describes mismatch only. If conductor, dielectric, feed and radome losses total another 1.2 dB, realized gain can be roughly 1.7 dB below directivity before installation effects. A favorable VSWR can still coexist with low radiation efficiency or a distorted pattern, so match, gain and pattern require separate evidence.

Text-free RF antenna engineering diagram linking the conducted reference plane, free-space pattern and polarization, and the installed radome and platform boundary — Read matching, free-space radiation and installed behavior as one evidence chain; a good port match or peak-gain value alone does not establish installed coverage.

Verify from the conducted port to the installed radiation boundary

Approve continuous frequency ranges, port states, waveform power, coordinate system, service volume and exclusion regions.
Freeze the antenna, feed, cable, adapters, mounting datum, ground plane, radome and platform configuration by revision.
Calibrate conducted fixtures to the named plane and measure complex return loss, isolation and coupling over frequency and environment.
Measure co-polar and cross-polar patterns with declared angular steps, frequency points, port terminations and beam states.
Determine gain or realized gain with a traceable reference antenna and an uncertainty budget appropriate to the band.
Verify sidelobes, nulls, front-to-back ratio, beamwidth, axial ratio and scan loss over the required angular region rather than boresight alone.
Apply approved CW or pulsed stress with real duty, crest factor, mismatch, pressure and temperature; inspect detuning, heating and damage.
Repeat critical measurements with the radome and representative platform, then correlate any final installed check to the range result.
Trend unit-to-unit and remount repeatability where production, calibration or interchangeability depends on the antenna.
Retain raw S-parameters, pattern grids, calibration identities, uncertainty, coordinates, photos without people, serials, revisions and deviations.

Antenna-selection failures that survive a catalog comparison

Selecting maximum gain before defining the angular coverage volume
Using directivity, gain and realized gain as if they were the same quantity
Accepting a single center-frequency plot for a continuous-band requirement
Treating good VSWR as proof of radiation efficiency or pattern quality
Specifying circular polarization without sense, axial ratio and angle range
Ignoring cross-polarization, sidelobes or null depth outside boresight
Applying one nominal power rating to a different pulse, mismatch, pressure or temperature
Testing the antenna without the radome, ground plane or nearby structure used in service
Comparing simulated and measured plots with different normalization or coordinate conventions
Keeping screenshots while discarding native pattern data, calibration and revision metadata

Information required for an RF or microwave antenna RFQ

Use case, installation, platform axes, service volume, exclusion regions and pointing or scan behavior
Continuous transmit and receive ranges, guard bands, port count, simultaneous states and impedance
Realized gain or gain limits by angle and frequency, beamwidth, ripple, sidelobes, nulls and front-to-back ratio
Linear orientation or circular sense, axial ratio, cross-polar discrimination and port isolation
Conducted reference plane, connector or waveguide, cable, balun, matching network and included losses
CW, average, peak and pulse power, duty, crest factor, carriers, mismatch, reverse power and temperature
Ground plane, radome, enclosure, keep-out zone, mounting datum, envelope, mass, cable exit and grounding
Temperature, altitude, pressure, humidity, salt, ice, wind, vibration, shock, ingress, vacuum or contamination
Single-port, diversity, MIMO or array states, beam steering, phase center, coupling and calibration needs
Conducted and radiated test methods, frequency points, angle grid, polarization basis and acceptance limits
Calibration traceability, measurement uncertainty, native data format, serial linkage and change control
Required samples, first-article evidence, production sampling, remount repeatability and installed correlation

Continue the antenna engineering decision

Related FAQ

What is the difference between antenna gain, directivity, beamwidth and sidelobes?

Directivity compares radiation in a direction with an isotropic radiator using the same total radiated power. Gain includes dissipative efficiency; realized gain also includes impedance mismatch at the declared feed. Beamwidth describes an angular width of the main beam, commonly between half-power points, while sidelobes are local radiation maxima outside that main lobe.

A high peak gain does not guarantee adequate coverage. The service region may need a minimum realized-gain envelope, ripple limit, sidelobe ceiling, null depth, front-to-back ratio and cross-polar limit over frequency. Beamwidth and gain also trade against each other for a given effective aperture.

Ask which gain quantity is reported, at which frequency and polarization, in which coordinate system, with what normalization and measurement uncertainty. Use the full pattern grid for system analysis rather than accepting one boresight value.

How are antenna VSWR, return loss, reflection coefficient and mismatch loss related?

At a named reference plane, reflection coefficient Γ is the reflected-to-incident voltage-wave ratio. Return loss is -20 log10|Γ| in dB, and VSWR is (1+|Γ|)/(1-|Γ|). Reflected power fraction is |Γ| squared; mismatch loss is -10 log10(1-|Γ| squared). A 2:1 VSWR corresponds to |Γ| about 0.333, return loss about 9.54 dB and mismatch loss about 0.51 dB.

These relations apply at the same impedance and plane. A cable, adapter, balun or matching network can transform the observed result, so state whether it is included and preserve calibration details. Wideband acceptance should use limits over the continuous band, not one marker.

Good match does not prove that accepted power is radiated efficiently or in the required directions. Review match, realized gain, efficiency, pattern and polarization as separate evidence, then evaluate installed effects.

How should linear, circular, axial-ratio and cross-polarization requirements be specified?

For a linearly polarized antenna, state the orientation relative to the installed coordinate system and the permitted cross-polar response. For circular polarization, state RHCP or LHCP with the agreed viewing convention. Axial ratio indicates how close the polarization ellipse is to a circle; it must be limited over the required frequency and angular region, not only at boresight.

Cross-polar discrimination compares desired and orthogonal polarization responses under a declared basis. It is distinct from isolation between two feed ports. Dual-linear and dual-circular designs may require both port isolation and pattern-level cross-polar limits.

Polarization mismatch reduces received power and can vary with platform attitude, reflection and radome effects. Include the operational orientation, angular coverage, radome and representative installation in verification.

Why can an installed antenna pattern differ from the free-space data sheet?

An antenna interacts electromagnetically with its surroundings. Ground-plane dimensions, mast or vehicle structure, enclosure, fasteners, cable route and nearby conductors can change current distribution and scattering. The result can be different return loss, efficiency, beam pointing, sidelobes, nulls, cross-polarization and phase center from a free-space or reference-ground-plane measurement.

A radome can add insertion loss, reflection, beam shift and polarization error. Those effects depend on material, thickness, seams, spacing, incidence angle, moisture, ice, temperature and frequency. If the product is used with a radome, test or correlate the declared radome configuration rather than treating it as cosmetic hardware.

Freeze the installed assembly by drawing and revision. Use a representative platform or a documented correlation method, then retain both baseline range data and installed check data so future mounting or material changes can be assessed.

Engineering inquiry

Share your RF requirement

Share the product, operating requirements and project context. Our engineering team will route your request to the right specialist.