SWR Cannot Validate Gain, F/B or Radiation Efficiency
Why the field screenshots do not confirm the model’s principal performance claims.
A low SWR is useful evidence. It shows that an antenna system presents a reasonably convenient impedance at the analyser’s reference plane over the displayed frequency range. It can confirm that an adjustment produced a match and that little incident power is immediately reflected there.
It is not a measurement of radiation.
Greg Mihran’s July 2026 antenna primer repeatedly follows precise 4NEC2 predictions of gain, radiation efficiency, takeoff angle, beamwidth and front-to-back ratio with photographs of low-SWR RigExpert sweeps. In the clearest examples, the field screenshot and the modelled far-field quantities appear on the same slide. That presentation creates an impression of experimental confirmation which the measurement cannot supply.
The screenshots need not be false for the inference to be false. They answer a different question.
SWR tells us how much power is reflected at one electrical port. Gain, front-to-back ratio and radiation efficiency tell us what happens in the electromagnetic field after power has been accepted. They require different measurements.
Give SWR credit for what it really measures
SWR is not meaningless and a VNA is not the wrong instrument. A properly calibrated one-port measurement can establish the complex reflection coefficient, input impedance, return loss and mismatch bandwidth at a named reference plane. It is one of the first measurements an antenna engineer should make.
For a real reference impedance Z0, the complex reflection coefficient is:
Γ = S11 = (Zin − Z0) / (Zin + Z0)SWR = (1 + |Γ|) / (1 − |Γ|)Preflected / Pincident = |Γ|2Paccepted / Pincident = 1 − |Γ|2
The Keysight Network Analyzer Basics application note gives these same relationships and distinguishes input reflection S11 from forward transmission S21. A RigExpert screen displayed in SWR format shows the magnitude of a one-port reflection result versus frequency. The instrument may offer other displays, but that screenshot publishes no far-field data.
If measured and modelled SWR agree, the model has passed one useful port-impedance check. If they disagree materially, the model or physical configuration deserves investigation. Agreement is positive evidence—but only for the observable that was compared.
The primer’s “99.8%” arithmetic is not efficiency
Slides 51, 58, 104, 109, 128, 134, 140 and 144 repeat a box labelled “SWR Matching”:
| SWR | |Γ| |
Reflected power | Accepted power |
|---|---|---|---|
| 1.10:1 | 0.0476 | 0.23% | 99.77% ≈ 99.8% |
| 1.20:1 | 0.0909 | 0.83% | 99.17% ≈ 99.2% |
| 1.30:1 | 0.1304 | 1.70% | 98.30% |
| 1.40:1 | 0.1667 | 2.78% | 97.22% |
| 1.50:1 | 0.2000 | 4.00% | 96.00% |
The arithmetic is correct. The label must remain mismatch acceptance, not radiation efficiency. Slide 155 is more careful when it says that 1.5:1 implies 96% power delivered to the antenna structure. The unresolved question is what the structure does with those accepted watts.
Paccepted = Pradiated + Plossηradiation = Pradiated / Pacceptedηrealised = (1 − |Γ|2) ηradiation
At 1.5:1 SWR, 96% of incident power is accepted. If the antenna structure is only 1% radiation-efficient, 0.96% of incident power becomes radiation. If it is 95% radiation-efficient, 91.2% becomes radiation. The SWR is identical.
NIST explicitly separates reflected power from radiation efficiency. Its examples show why an accepted-power definition and an incident-power definition must not be mixed.
The dummy-load test ends the argument
Connect a good 50 Ω dummy load. The analyser reports approximately 1:1 SWR. Nearly 100% of the incident power crosses the reference plane. Almost all of it becomes heat, not useful radiation.
Now connect a carefully designed matched antenna. It can show the same 1:1 SWR while turning most accepted power into radiation.
The port measurement cannot distinguish the two because it measures the wave coming back through the port, not the full set of fields leaving the structure. The same numerical input impedance can result from countless combinations of radiation, ohmic loss, ground loss, transformer loss and current distribution.
The reverse example matters too. An otherwise lossless antenna with 2:1 SWR can still have 100% radiation efficiency relative to accepted power. It accepts 88.9% of the incident power, then radiates all of that accepted power. Poorer SWR has reduced realised performance through mismatch; it has not proved internal dissipation.
Therefore:
- low SWR does not prove high radiation efficiency;
- high SWR does not prove low radiation efficiency;
- SWR supplies the mismatch term, and nothing more.
Why one port cannot reveal an angular pattern
Antenna pattern is a function of direction. If U(θ,φ) is radiation intensity:
Pradiated = ∫4π U(θ,φ) dΩD(θ,φ) = 4πU(θ,φ) / PradiatedG(θ,φ) = 4πU(θ,φ) / Paccepted = ηradiationD(θ,φ)Grealised(θ,φ) = (1 − |Γ|2)ηradiationD(θ,φ)
NIST Technical Note 1551 defines antenna gain as directivity multiplied by radiation efficiency and defines radiation efficiency as radiated power divided by accepted power. SWR provides only the final mismatch factor used when calculating realised gain from incident power.
Notice what is absent from S11: there is no azimuth, elevation, polarisation, beamwidth, front direction, back direction or radiated-power integral. A scalar SWR sweep cannot contain an unmeasured three-dimensional field.
A useful thought experiment is to rotate a complete directional antenna by 180° in a clear, uniform test environment. Its terminal impedance and SWR remain essentially the same. Its forward lobe now points the opposite way. The port did not report that change in geographic direction because that is not the information it measures.
What the primer pairs together
| Primer section | Modelled performance claim | Published field evidence | Scientifically justified conclusion |
|---|---|---|---|
| Slides 42–51: PERformer | 90.8% average efficiency, +0.41 dBi average gain, 3.09 dB F/B, 24° peak angle and 44° beamwidth. | Slide 51 shows six RigExpert SWR photographs. | The tested system had low analyser-plane reflection over the displayed sweeps. None of the far-field quantities was measured. |
| Slide 58: PERformer NVIS | +0.67 to +4.15 dBi, an 86° peak angle, 75°–131° beamwidth and 85.5%–87.7% efficiency. | Six field SWR photographs appear on the same slide beside the table headed “Computer Model Specifications”. | This is the clearest category error: the photographs support input match, not “great gain @ 86°”, beamwidth or efficiency. |
| Slides 60–69: parasitic PERformerArray | Average +2.7 dBi gain, 7.7 dB F/B and 87% efficiency; 20 m claim of +1.9 dBi and 8.2 dB F/B. | Deployment photographs and one 20 m RigExpert sweep near 1:1. | The driven element and nearby parasitic element can be matched. The induced reflector current and resulting pattern remain unmeasured. |
| Slides 75–79: phased PERformerArray | Modelled gains and up to 22 dB F/B. | 4NEC2 SWR plots labelled [50 ohm] [Src 1], not a measured combined physical feed-network port. |
A source-1 match does not verify element-current amplitude and phase, feed-network loss or the array pattern. |
| Slides 95–109: Challenger and Dominator | 94.3% and 99.5% structural efficiency, modelled gains, angles and beamwidths. | Field SWR photographs after practical matching transformers and chokes. | The transformed input is matched. The screens do not confirm structural efficiency, transformer efficiency, gain or angle. |
| Slides 121–122: ChallengerArray | +1.40 dBi, +1.76 dB over one element, 8.38 dB F/B and 88.1% structural efficiency. | Slide 122 overlays a RigExpert screenshot on a 200 Ω model SWR plot beside modelled elevation and azimuth patterns. | Visual proximity is not cross-validation. A 50 Ω-side screenshot cannot establish the plotted pattern, especially without the complete 4:1 network in the comparison. |
| Slides 127–128: DominatorArray | +3.92 dBi, 16° peak angle, 24° beamwidth, 8.54 dB F/B and 99.5% efficiency at 28.4 MHz. | Four low-SWR field screenshots after the high-ratio matching system. | The transformer-input match and bandwidth were measured. The headline far-field results were not. |
| Slides 131–144: Marauder, Thruster and Hammer | Modelled gain, directional ratios and takeoff angles for several further designs. | More sets of low-SWR analyser photographs. | The same evidence boundary applies: a match sweep is not a gain, efficiency or pattern sweep. |
The fair conclusion is not that the modelled values must be wrong. It is that the field screenshots do not validate them. The distinction between “predicted” and “measured” must remain visible.
Slide 58: six measurements, but only one kind of measurement
Slide 58 is titled “PERformer NVIS is Very Broadbanded with Great Gain @ 86°”. It places six real analyser photographs beside a model table containing SWR, reflection coefficient, gain, radiation angle, F/B, beamwidth, efficiency and impedance.
The photographs provide six bands of evidence about match. They do not provide six independent confirmations of every column in the table.
- “Very broadbanded”: supported only as impedance/SWR bandwidth over the displayed spans.
- “Great gain”: needs a calibrated radiated-field comparison.
- “86°”: needs an elevation-pattern measurement.
- 85.5%–87.7% efficiency: needs radiated power relative to accepted power.
- Beamwidth: needs angular samples through the relevant plane.
One photograph can be legitimate evidence and still be irrelevant to most words in its headline.
The reference-plane problem is not a footnote
The Challenger model plots are referenced to 200 Ω. The physical antenna is measured by a nominally 50 Ω analyser after a 4:1 unun, choke and any intervening feedline. The Dominator model plots are referenced near 2,450 Ω, while its field screens are taken after a 49:1 or 56:1 transformer and choke.
Those are different electrical planes. A direct comparison requires the real matching network to be included in the model or measured and de-embedded. It requires the calibration plane, cable length, transformer configuration and choke to be disclosed.
A lossy path can even improve the apparent source-end SWR because the reflected wave is attenuated on both trips:
Γinput = Γload e−2(α + jβ)l|Γinput| = |Γload| e−2αl
That is not an accusation that these particular sweeps are dominated by loss. It is why the measurement boundary must be stated before a good match is used as evidence about the radiator. Slide 91 itself lists approximately 0.51–1.08 dB transformer-plus-choke loss for the Dominator alternatives; that loss is real system performance even when SWR is excellent.
Why model-to-field SWR agreement is only partial validation
Suppose a model predicts 1.05:1 and the field analyser measures 1.05:1. That is encouraging. It suggests the net complex impedance, after any required transformation, is close enough to produce the same reflection magnitude.
It does not prove that the model obtained the right answer for the right physical reasons. Multiple combinations of radiation resistance, loss resistance and reactance can produce the same total input impedance. Multiple current distributions can also present similar terminal impedance while producing different azimuth and elevation patterns.
For a parasitic or phased array this non-uniqueness is critical. The front-to-back ratio depends on the complex element currents. A mistuned parasitic element or an erroneous phase network can leave the driven-port SWR looking good while degrading or reversing the intended pattern. The missing validation quantity is the current and field, not another photograph of the port.
A scientifically sound statement would be:
What actually measures gain?
Gain is an absolute far-field quantity in a stated direction. It combines directivity with radiation efficiency. A credible measurement therefore needs a calibrated reference, known geometry, polarisation, accepted power and propagation environment.
Recognised approaches include gain substitution against a calibrated antenna, two- or three-antenna methods and near-field measurements transformed to the far field. NIST’s extrapolation-range work describes precision gain and polarisation measurements, while IEEE 149-2021 provides the broader antenna-measurement framework.
For an HF field comparison, a carefully controlled A/B test can establish relative realised gain in selected directions:
- Use the same transmitter, frequency, location and receiver.
- Switch rapidly between the test antenna and a known reference.
- Correct to equal accepted power at the antenna reference planes.
- Control polarisation, feedline common mode and geometry.
- Repeat across azimuths, elevation-sensitive paths and time.
- Publish raw field levels, calibration and uncertainty.
One directional A/B value is not a full gain pattern, but it is radiated-field evidence. SWR is not.
What actually measures front-to-back ratio?
Front-to-back ratio compares radiation intensity in two directions:
F/B = 10 log10(Ufront / Uback)or, at equal distance and for the same field component,
F/B = 20 log10|Efront / Eback|
The antenna must be rotated in a controlled field, or a calibrated probe must sample the relevant angular pattern. The report must define whether “back” means exactly 180° from the forward maximum, the largest rear lobe or a rear-sector average. It must state frequency, elevation, polarisation, angular resolution and uncertainty.
An SWR reading has no front and no back. It cannot measure 8.54 dB, 15 dB or 22 dB F/B.
This matters because deep rear nulls are fragile. Ground reflections, nearby objects, feedline radiation and small element-current errors can fill them. A modelled 22 dB null deserves a measured pattern, not a one-port trace.
What actually measures radiation efficiency?
Radiation efficiency requires the total radiated power and the power accepted at the defined antenna port. Established approaches include:
- measuring absolute gain and the full three-dimensional pattern, deriving directivity, then using
ηrad = G / D; - integrating calibrated radiated power over the full sphere and dividing by accepted power;
- using a calibrated reverberation-chamber method;
- using a Wheeler-cap method where the antenna size, frequency and assumptions make that method appropriate.
NIST’s reverberation-chamber work explicitly treats radiated power as the missing observable needed to determine antenna efficiency. These methods are substantially more involved than a one-port field sweep because they answer a substantially different question.
A calibrated relative field-strength comparison can demonstrate that one configuration places more power in a tested direction. It cannot by itself separate increased total efficiency from pattern redistribution. To claim total radiation efficiency, the entire angular power distribution or another recognised efficiency method is required.
Broad SWR bandwidth is only impedance bandwidth
The primer repeatedly describes the antennas as “very broadbanded” based on SWR sweeps. That wording needs a named performance criterion.
A low-SWR band is an impedance bandwidth. Across that same band:
- gain can change;
- the main lobe can move;
- F/B and null depth can collapse;
- matching-network loss can increase;
- common-mode current can change;
- radiation efficiency can rise or fall.
The NTIA Antenna System Guide describes VSWR as a practical measure of terminal-impedance bandwidth while noting that antenna bandwidth can instead be defined by acceptable gain, pattern or another performance characteristic. Those bandwidths are not automatically identical.
Loss can also broaden a match. A wide, smooth SWR curve is therefore not intrinsically a badge of efficient radiation. It may accompany an excellent antenna, a lossy antenna or both at different frequencies.
Common-mode current can pass the SWR test
A matching network, sparse counterpoise or imperfect choke can allow RF current onto the outside of the coax, mast or other conductors. That current can contribute to radiation, dissipation and pattern distortion while also changing the measured input impedance.
The analyser still sees one total reflection coefficient. It does not label which current flowed on the intended radiator and which flowed on the feedline exterior. A low SWR can coexist with an installation-sensitive pattern and RF in the operating position.
NIST’s review of antenna-measurement challenges discusses unwanted common-mode currents as a source of measurement error. For portable HF systems, a clamp-on RF current probe, documented choke impedance and coax-route sensitivity tests are valuable companions to the VNA.
A reproducibility matrix for the primer’s claims
| Claim | Minimum suitable evidence | What must be reported |
|---|---|---|
| Input match and SWR bandwidth | Calibrated S11 at the antenna-system port. |
Raw Touchstone data, calibration plane, cable, choke, network, geometry and environment. |
| Absolute or relative gain | Calibrated substitution, two-/three-antenna method or controlled A/B field measurement. | Accepted power, reference antenna, range, polarisation, direction, propagation correction and uncertainty. |
| F/B and beamwidth | Azimuth/elevation pattern sweep or suitable near-field scan. | Angular convention, resolution, field component, reference direction, multipath control and raw levels. |
| Radiation efficiency | Radiated-versus-accepted power, gain/directivity, reverberation chamber or applicable Wheeler method. | System boundary, mismatch treatment, feed-network loss, full pattern or chamber procedure and uncertainty. |
| Array operation | Complex element currents plus a measured pattern. | Amplitude, phase, reference directions, feed-network loss, mutual coupling and common-mode current. |
| Model validation | Agreement across independent observables, not only impedance. | Source files, solver, ground, losses, convergence, currents, power budget and comparative measurements. |
Takeaways you can trust
- SWR is a valuable one-port mismatch measurement.
- The reference plane must be stated.
- The primer’s 99.8% to 96.0% values are accepted-power percentages, not radiation efficiencies.
- Accepted power can become radiation or heat.
- A 1:1 dummy load proves that perfect SWR need not imply useful radiation.
- Radiation efficiency is radiated power divided by accepted power.
- Gain combines radiation efficiency with directivity.
- Realised gain additionally includes mismatch.
- F/B, beamwidth and radiation angle require angular field measurements.
- A model-to-field SWR match partially validates port impedance, not the full model.
- Transformer, choke and feedline loss can separate the analyser plane from the radiator plane.
- Broad SWR bandwidth is not automatically broad gain, pattern or efficiency bandwidth.
- Common-mode current can alter pattern while SWR remains low.
- Precise gain, F/B and efficiency claims require calibrated radiated-field or power-balance evidence.
In Summary
The primer contains many genuine field measurements. Its RigExpert photographs show that the tested antenna systems could be adjusted to present low reflection over the displayed spans. That is useful practical information.
But the principal performance claims—gain in dBi, front-to-back ratio, radiation angle, beamwidth and radiation efficiency—come from computer models. The field screenshots do not measure those quantities, and placing the two evidence types together does not make them interchangeable.
The correct scientific statement is modest and strong: measured SWR can corroborate the model’s input match at a carefully aligned reference plane. Gain requires a calibrated field comparison. F/B requires a pattern. Efficiency requires radiated power relative to accepted power.
A good match may accompany a good antenna. It is not evidence that the antenna is good.
Mini-FAQ
- What does SWR actually measure? SWR describes the magnitude of impedance mismatch at a particular reference plane. It indicates how much power is reflected there, but not what happens to the power accepted by the antenna system.
- Can an inefficient load have a perfect SWR? Yes. A dummy load can present a nearly perfect 1:1 SWR while converting almost all accepted power into heat. A lossy antenna or matching network can likewise provide a convenient match without radiating efficiently.
- Why can’t SWR determine radiation efficiency? Radiation efficiency is radiated power divided by accepted power. SWR identifies reflected power, not how the accepted power is divided between useful radiation and conductor, transformer, feedline or ground loss.
- Can low SWR prove antenna gain? No. Gain combines radiation efficiency with directivity. It must be established through calibrated pattern or comparison measurements that include accepted power, geometry, polarisation and a suitable reference antenna.
- Can SWR establish front-to-back ratio? No. Front-to-back ratio compares radiation or reception in two directions. It requires directional field or pattern measurements. A one-port impedance reading contains no information about where the energy travels.
- What evidence would support gain, F/B and efficiency claims? Use controlled far-field pattern measurements, equal accepted power, a calibrated reference antenna, stable geometry, appropriate common-mode suppression and a recognised efficiency-measurement method. SWR remains useful, but it answers only the matching question.
Want more technical RF content? Subscribe for new deep dives and lab notes.
Have a question or field observation? Contact RF.Guru.