Why You Can’t Measure Antenna Efficiency with a VNA

What You Can Measure Instead

A VNA (Vector Network Analyzer) is excellent at measuring impedance and matching. But antenna efficiency is a different question entirely.

People often connect a VNA to an antenna, see a beautiful 1.1:1 SWR, and assume the antenna must be “efficient.” That assumption is the trap:

• A 50 Ω dummy load can show an excellent match (SWR ≈ 1:1) ... and have 0% radiation efficiency.
• A very lossy antenna can be matched to 50 Ω and show a great SWR while quietly turning most of your RF into heat (ground loss, coil loss, resistive loss in joints, nearby objects, etc.).

So: S11 / SWR is not efficiency. It only tells you how well power is accepted by the load seen at the feedpoint ... not how much of that accepted power becomes radiated electromagnetic energy.

What “antenna efficiency” actually means

A practical definition is radiation efficiency:

η = Pradiated / Paccepted

At the feedpoint (near resonance, where reactance is near zero), you can think in terms of resistances:

• Rrad = radiation resistance (the “good” part)
• Rloss = loss resistance (the “bad” part: ground, conductor loss, coil/trap loss, absorption by nearby objects, etc.)

At resonance:

Rin = Rrad + Rloss

Then:

η = Rrad / (Rrad + Rloss)

What a VNA can measure on an antenna

Used correctly, a VNA is fantastic for electrical characterization. Here’s what it can reliably tell you.

Complex impedance vs frequency

• Resistance (R) and reactance (X) across a band
• Resonant frequency (where X crosses zero)
• Whether the antenna looks capacitive or inductive above/below resonance

(Best practice: calibrate at the measurement plane ... ideally at the feedpoint, or de-embed the feedline.)

SWR / return loss / reflection coefficient

SWR is derived from S11. It’s a matching metric:

• It tells you how much power reflects back because of mismatch.
• It does not tell you how much of the accepted power is radiated.

Bandwidth and “loaded Q” (with caution)

You can infer how quickly impedance changes with frequency and estimate a loaded Q for traps/matching networks/small antennas. But bandwidth is influenced by many things (including loss) ... so it’s not a clean efficiency metric by itself.

Network/component behavior (where a VNA truly shines)

A 2-port VNA is ideal for components and networks:

• Coax insertion loss (S21)
• Matching network transfer behavior
• Balun/choke behavior (with the right fixture) and frequency-dependent impedance trends

Why S11 can’t tell you efficiency

S11 answers: “How much power reflects because of mismatch?”

Efficiency asks: “Of the power that didn’t reflect, how much got radiated?”

Those are different. A lossy antenna can be a perfect match.

Why VNAs often “lie” more easily on real outdoor antennas

Antennas are receivers ... and HF is a hostile measurement environment

On HF, an antenna can pick up very strong off-air signals (AM broadcast, shortwave broadcasters, nearby transmitters, local noise sources). Many VNAs (especially compact/low-cost units) have wideband, lightly filtered receiver front-ends that can overload more easily.

When that happens, the VNA is no longer measuring only its own stimulus and reflection. External RF can get into the measurement and you’ll see:

• unstable traces
• ripple that changes with time
• return loss that looks mysteriously “too good” or “too bad”
• nonsense impedance in certain parts of the spectrum

Calibration and reference-plane mistakes are common

If you calibrate at the instrument and then add 20 m of coax, the coax transforms the impedance. You might get “correct” S11 at the instrument plane ... but you are no longer looking at the antenna feedpoint unless you de-embed or calibrate at the end of the coax.

(This is why many field-focused antenna analyzers feel “easier” ... they’re built around a one-port antenna workflow and tolerate real-world environments better.)

So what can you measure instead, if you care about efficiency?

To estimate efficiency, you must separate radiated power from loss power. In practice, there are two useful buckets:

• Loss accounting (where power becomes heat)
• Far-field comparisons (measure what actually gets radiated)

Loss accounting with I²R ... where the power becomes heat

If you can estimate losses, you can build efficiency from them:

Ploss = Irms2 · Rloss

Then:

η = 1 - (Ploss / Paccepted)

What makes this hard: you need the RF current where losses occur (often very high near loading coils, trap regions, and at the base of verticals) ... and you need the RF loss resistance (not the DC resistance).

Common real-world loss contributors:

• Ground loss (verticals, low antennas, insufficient radials)
• Loading coil loss (shortened antennas)
• Conductor loss (thin conductors, small diameter, rough surfaces)
• Joints and connections (oxidation, loose clamps, dissimilar-metal corrosion)
• Matching network loss (component ESR, ferrites heating, saturation under power)

A practical workflow that actually works:

• Measure input resistance at resonance (with an antenna analyzer, or a carefully used VNA at the feedpoint). This gives Rin = Rrad + Rloss.
• Use theory/modeling (or known approximations for that geometry) to estimate Rrad.
• Infer Rloss = Rin - Rrad and compute efficiency.

Loading coils and Q ... why coil quality matters so much

Short antennas force high current through loading components. Even a small loss resistance becomes big heat.

If you know coil reactance XL and coil Q:

Q = XL / Rcoil

So:

Rcoil = XL / Q

High current + modest Rcoil = real power loss. This is why physically larger coils, good conductors, sensible spacing, and robust construction matter ... especially on low bands and shortened antennas.

Far-field comparisons ... measure what actually gets radiated

Absolute gain measurement is hard. Relative comparison is much easier and often more honest.

A practical A/B approach:

• Use a reference antenna with known/stable behavior (for example a clean half-wave dipole in a consistent configuration).
• Transmit a stable carrier at fixed power on a fixed frequency.
• Measure received power at a distant receive site (or a calibrated field-strength setup).
• Swap to the antenna under test without changing location, height, power, or feedline routing.
• The difference in received power is approximately the difference in realized gain.

(On HF, “near field” is physically large. A house 2 meters away is not “clear” ... it’s usually deep inside the reactive near-field where coupling, detuning, absorption, and pattern distortion dominate.)

Ground, height, and why “radials” are not a detail

Vertical antennas are the classic case: the ground return is part of the system. Current is maximum near the base, and return current flows in your radial system and/or the soil. Soil is not a perfect conductor ... it behaves like a resistor at RF, which becomes I²R ground loss.

A dense radial field provides a low-resistance path so current doesn’t have to flow through lossy soil. That’s why serious radial systems dramatically improve real-world performance.

Important nuance: even with excellent radials, nearby conductive structures can still absorb energy and distort patterns. Radials solve the ground-conductivity problem ... not the entire site-coupling problem.

Tube thickness vs tube diameter ... the RF “gotcha”

At HF, skin effect means most RF current flows in a thin surface layer (skin depth). Once your wall thickness is more than a few skin depths, making it thicker doesn’t reduce RF resistance much.

What usually matters more is surface area: larger diameter means more circumference, lower RF resistance, and often less loss where currents are high (base of verticals, loop conductors, loading coils, traps).

In real antennas, joints often dominate losses. A slightly oxidized, loose, or dissimilar-metal joint can add more RF loss than the difference between thin-wall and thick-wall tubing.

A realistic field workflow that won’t fool you

• Use impedance tools for what they are good at: tune resonance, measure R and X, verify SWR bandwidth.
• Control feedline behavior: use a proper choke/balun where needed so the coax doesn’t become part of the antenna.
• Don’t measure in a storm of RF: avoid measurements while strong transmitters are active nearby.
• If using a VNA at an antenna: add an inline attenuator (10–20 dB), consider band-pass/notch filtering, keep leads short, and calibrate at the end of the cable.
• Estimate losses: infer Rloss from resonance resistance + modeled/known Rrad, and pay attention to ground, coils, and joints.
• Validate with an A/B far-field comparison: same power, same geometry, stable receive site, reference antenna vs antenna under test.

Interested in more technical content? Subscribe to our updates for deep-dive RF articles and lab notes.

Questions or experiences to share? Feel free to contact RF.Guru.