Test a broadband waveguide antenna in two parts: measure input match at its feed interface, then measure gain, radiation pattern and polarization on a characterized antenna range. Repeat the radiated measurements across the operating band. A low VSWR at one frequency does not prove broadband gain or pattern performance.
This guide covers passive antennas at small-signal test levels. Power handling, thermal behavior and environmental qualification need separate procedures and acceptance limits.
Define the acceptance criteria before setting up
| Quantity | Test result | Condition to specify |
|---|---|---|
| Input match | S11, return loss or VSWR versus frequency | Feed reference plane, waveguide mode and included adapters |
| Gain or realized gain | Absolute level in dBi versus frequency and direction | Reference antenna data, loss and mismatch corrections |
| Radiation pattern | Co-polar and cross-polar angular cuts or a full scan | Coordinate system, polarization convention and normalization |
| Beam shape | Peak direction, half-power beamwidth and sidelobes | Frequency, angular coverage, sampling and dynamic range |
| Broadband consistency | Band-edge behavior, gain variation and beam changes | Frequency grid and limits for each measured quantity |
Use the requirements for the actual antenna and application. Values such as “gain above 12 dBi” or “VSWR below 1.5” are possible purchase requirements, not universal pass limits for every broadband horn.
Match the feed hardware to the band
Check the antenna drawing, waveguide designation, flange, connector and specified operating band. Standard rectangular WR90 and WR28 interfaces serve different frequency ranges; a WR90 calibration kit is not appropriate for a 26.5–40 GHz WR28 feed. Dolph’s calibration-kit datasheet, for example, lists separate WR90 and WR28 models.
For a double-ridged broadband feed, match the WRD interface and its full band instead of choosing a rectangular-guide standard by frequency alone. Include every adapter, cable and frequency extender in the test-path drawing.
Use waveguide-to-coaxial adapters within their specified band and power limits. Mount the antenna without bending the feed assembly, and route moving cables so rotation does not change the connection or load the flange.
Measure input match at the antenna reference plane
Calibrate the VNA at the antenna’s feed plane using suitable standards and definitions. A coaxial calibration upstream of an adapter leaves that adapter in the measured response unless it is otherwise corrected. The waveguide calibration procedure explains reference-plane selection and verification.
Keep the aperture clear of nearby objects during the match measurement; objects in front of the antenna can change the measured response. Record the mounting and surrounding conditions so later tests are comparable.
For a passive single-port antenna, S11 log magnitude is negative for partial reflection, while return loss is conventionally reported as a positive number. An ideal short reflects strongly: it does not verify a calibration by producing very low S11.[1]
| S11 log magnitude | Return loss | VSWR, approximately | Reflected power |
|---|---|---|---|
| −10 dB | 10 dB | 1.925:1 | 10% |
| −15 dB | 15 dB | 1.433:1 | 3.16% |
| −20 dB | 20 dB | 1.222:1 | 1% |
These are calculated conversions, not product specifications: |Γ| = 10S11(dB)/20, VSWR = (1 + |Γ|)/(1 − |Γ|), and reflected-power fraction = |Γ|². Thus, −10 dB S11 and 1.5:1 VSWR are not equivalent limits.
Select a valid radiated-measurement range
For an initial far-field distance estimate, use R ≥ 2D²/λ, where D describes the antenna’s radiating extent and λ is free-space wavelength. The distance must also suit the other range antenna and the required uncertainty. A fixed distance of “15 wavelengths” is not sufficient for every aperture.[2]
For a calculated example with D = 0.15 m, the estimate is about 1.5 m at 10 GHz and 6 m at 40 GHz. For fixed aperture size, the higher frequency gives the longer distance. This estimate does not certify a chamber’s quiet zone or reflection performance.
If the available range is too short, use a validated near-field system with the required spatial sampling, phase information and probe correction, or a suitable compact antenna test range. Relabeling a short-range angular scan as a far-field pattern does not correct it. NIST describes near-field scanning as a measurement-and-transformation method.[3]
Measure gain with a characterized reference
Compare the antenna under test with a reference antenna whose gain data cover the test frequencies. Preserve the source, range geometry and receive path, and apply corrections for changes in cable loss, mismatch, distance or polarization. Define whether the reported quantity is gain or realized gain; the latter includes input mismatch.
For a free-space, polarization-aligned link, a useful power-budget check is:
Pr(dBm) = Pt(dBm) + Gt(dBi) + Gr(dBi) − 20 log10(4πR/λ) − L(dB).
Antenna gains add to the received-power budget. This is the logarithmic form of the Friis relation; it assumes appropriate far-field conditions and aligned antennas.[4]
As a calculation example, at 10 GHz and 10 m the free-space path loss is approximately 72.45 dB. With 0 dBm source power, 20 dBi transmit gain, 15 dBi receive gain and 2 dB total feed loss, the predicted received power is −39.45 dBm. This checks the setup budget; it is not a substitute for gain calibration.
When selecting a standard gain horn, confirm the available gain characterization, polarization and interface. Do not treat a nominal catalog gain as an uncertainty-qualified calibration result.
Measure pattern and polarization across the band
- Align and define axes: record the antenna orientation, rotation center and zero-angle direction.
- Sample the beam: choose angular spacing that resolves the narrowest expected beam and sidelobe structure; repeat with finer spacing if peaks or beamwidth change materially.
- Measure both polarizations: use a stated co-/cross-polar basis and characterize the receive channels. Keep the cross-polar trace referenced to the co-polar level when reporting discrimination.
- Check dynamic range: identify receiver noise and chamber reflections. Do not report a sidelobe or cross-polar level below the demonstrated measurement floor as an exact value.
- Repeat by frequency: include band edges and any frequencies where matching or gain changes sharply. Add intermediate samples until the behavior is adequately resolved.
Two principal-plane cuts do not constitute a full 3D pattern. The radiation-pattern measurement guide covers angular scans, normalization and extracting beam metrics.
Report the result with its limits
Deliver raw S-parameters, frequency-tagged pattern data, gain results, the coordinate convention and the test configuration. State which losses were corrected and whether patterns are normalized or absolute. Include the reference antenna’s identification and the applicable calibration information.
The uncertainty budget should address the reference, alignment, mismatch, range reflections, receiver behavior and connection repeatability. NIST recommends estimating the separate contributions and combining them into the overall measurement uncertainty.[5]
A result close to a limit needs the agreed acceptance rule and uncertainty, not just a green marker on a trace. For a purchase inquiry, send Dolph the operating band, feed drawing, polarization and required gain/pattern limits, together with the test data expected at delivery.
