End-Fed SWR Measurement: Define the Return Path and Reference Plane
End-Fed SWR Measurement: Define the Return Path and Reference Plane
When an end-fed antenna deliberately uses part of the coax exterior as its return conductor, the feedline is no longer only a transport line. Its route, length and common-mode boundary become part of the installed antenna—and the SWR reading changes when either the circuit or the measurement plane changes.
I often see one SWR curve treated as if it describes the wire, transformer, coax and return path independently. It does not. An instrument reports reflection at its calibrated plane. To understand an end-fed system, draw every conductor that carries RF, identify the intended return boundary and compare measurements without silently changing that boundary.
My practical rule: transformation and common-mode suppression are separate jobs. An UNUN may transform an intentionally unbalanced port, while a separately specified choke defines where the chosen coax-exterior return section ends. Neither the ratio nor the choke position comes from one universal fraction of wavelength; the installed complex load, current map, loss and stress decide.
An End-Fed Wire Still Needs a Complete RF Circuit
Current cannot leave one transformer terminal and disappear. The return path may include a dedicated counterpoise, radial or capacitive structure, the exterior of the coax shield, nearby conductors and displacement current through the surroundings. Several paths can operate at once.
Inside a coaxial cable, the wanted transmission-line mode carries equal and opposite currents on the centre conductor and the shield's inner surface. Current on the outside of the shield is a different mode. It can radiate, receive local noise and couple to the mast, ground, equipment and station wiring.
If a defined length of shield exterior is intentional, draw it as an antenna conductor. The transformer is then not the far edge of the installed radiating system. The common-mode choke establishes the far boundary of that chosen exterior section—provided its impedance is adequate at the installed frequency, current, voltage and environment.
SWR Belongs to a Declared Reference Plane
A one-port vector network analyser measures the complex reflection coefficient at its calibration plane. From that value it can display return loss, SWR and impedance. Move the instrument through a cable without mathematically removing the cable, and the reference plane changes.
Reflection coefficient at a load plane: Γ = (ZL − Z0) / (ZL + Z0)
Lossless-line SWR: SWR = (1 + |Γ|) / (1 − |Γ|)
On an ideal uniform lossless line, the magnitude of Γ and therefore SWR stay constant along the line, while reflection phase changes. Resistance and reactance at the radio can therefore differ greatly from the impedance at the antenna even though the ideal-line SWR is unchanged.
Real cable has attenuation. The reflected wave makes a round trip before a source-end instrument sees it, so line loss reduces the displayed reflection. A longer or lossier cable can make source-end SWR look closer to 1:1 while delivering less power to the antenna. That is attenuation masking mismatch, not an improved load.
Choose the Plane That Answers the Question
| Measurement plane | What the result includes | What it cannot prove alone |
|---|---|---|
| Wire-side transformer port | The high-impedance antenna-side load at small-signal test level, if the intended return network is connected | Transformer input match, feedline loss, operating-power stress or realised radiation efficiency |
| Coax-side transformer port | Wire, return network and the transformer's small-signal input behaviour at that plane | The impedance presented at the radio after the feedline, or whether shield current stays inside its intended boundary |
| Radio or tuner end | The source-end reflection after line transformation and attenuation, with the installed cable route | The antenna-terminal SWR, transformer loss, radiated power or the cause of any change |
| De-embedded antenna plane | An estimate after mathematically removing a characterised fixture or cable | Accuracy beyond the cable model, calibration, connector repeatability and instrument uncertainty |
“Measure at the UNUN” is therefore incomplete. State which port, which return conductor is attached, where the choke sits, whether the feedline is present and where calibration was performed. A wire-side measurement with the intended shield-exterior return disconnected is a measurement of a different circuit.
Coax Transformation and Shield Current Are Different Mechanisms
A mismatched differential signal inside the coax follows ordinary transmission-line behaviour. Its input impedance changes with the line's complex propagation constant, characteristic impedance and electrical length. The manufacturer's velocity factor describes that internal mode for the specified cable and frequency range.
The exterior-current path is not automatically governed by that same velocity factor. Its effective propagation depends on the jacket, soil, routing, mast, building, nearby conductors and how the cable is terminated in common mode. A quarter wavelength calculated with the internal coax velocity factor does not locate a universal exterior-current maximum or a universal choke position.
Changing coax length can therefore produce two different experiments:
- If the exterior-current boundary and route remain unchanged, added cable mainly changes differential transformation and loss between two planes.
- If the added section lies inside the intentional exterior return or moves its choke boundary, the installed antenna itself changes.
- If the new cable route couples differently to soil, a mast or station wiring, the common-mode network changes even when physical length is nominally unchanged.
These effects can occur together. That is why one SWR minimum cannot identify the cause.
A Choke Defines a Current Boundary, Not a Magic Distance
A suitable common-mode choke adds impedance to the unwanted exterior-current path while passing the wanted internal coaxial mode. Its useful impedance is complex and frequency-dependent. Core material, winding, stray capacitance, load, temperature and installation all matter.
Place the first choke where the intended return conductor should end. That can be close to the transformer when a dedicated return structure is already present, or farther down the cable when a chosen exterior section is deliberate. Measure current around the complete coax at several marked positions on every band, then repeat after the choke is installed.
A second choke near the station can be useful when another current path exists through equipment, mains, control or bonding conductors. It is not automatically required, and a quiet station end does not prove that the upstream coax exterior has stopped carrying current or changing the pattern.
The practical choice is not “choke or no choke.” It is a defined current boundary, enough common-mode impedance across the required bands, acceptable differential insertion loss and acceptable voltage, current and thermal stress.
Transformer Ratio Does Not Predict the Installed Match
An ideal transformer with turns ratio n transforms impedance by n2. A ratio label such as 4:1, 9:1 or 49:1 does not establish the complex antenna-side impedance, guarantee 50 Ω at the coax port or prove suitability on every band.
A real broadband transformer adds magnetising impedance, leakage, winding capacitance, conductor loss and core loss. Its result depends on frequency, complex load, construction, drive waveform, voltage, current, power, duty cycle and temperature. A small-signal VNA trace does not certify operating-power loss or survival.
Measure transformer transmission and reflection with a representative complex load over the required frequency range. At operating power, monitor temperature and verify voltage/current limits. Keep this separate from exterior-current suppression: a favourable SWR can coexist with transformer heating, and a high common-mode impedance does not establish low differential loss.
Low SWR Is Not Radiation Efficiency
SWR describes mismatch at a stated plane and reference impedance. It does not divide accepted power into useful radiation and loss. A dummy load has excellent SWR and deliberately poor radiation.
For an installed end-fed system, the accepted-power balance can include:
- radiation from the intended wire;
- radiation from an intentional or unintended return conductor;
- feedline attenuation under the installed standing-wave distribution;
- transformer and choke dissipation;
- conductor, connection, ground and dielectric loss; and
- coupling into nearby structures and station wiring.
A changed SWR curve may be useful diagnostic evidence, but it is not a direct measurement of radiation efficiency, field strength, pattern or received SNR. Those require separate calibrated measurements or a validated model of the complete installation.
A Controlled Measurement Sequence
- Draw the installed conductors. Include the wire, transformer ports, intended counterpoise or radial, coax exterior to the first choke, mast, bonding and station connections.
- Name the question and plane. Decide whether you need wire-side impedance, transformer input match, radio-end match, line loss, common-mode current or radiated performance.
- Calibrate at that plane. Perform the analyser calibration with the same adapters and test lead arrangement that will remain in the measurement.
- Record the vector result. Save frequency, R + jX or complex S11, not only the lowest SWR number.
- Characterise the feedline. Record cable type, physical length, connectors, attenuation and internal-mode velocity factor; de-embed only within the quality of that data.
- Map exterior current. Use a suitable RF current probe around the complete cable at repeatable points before and after the intended boundary.
- Change one variable. Move one choke, add one known cable section or change one return conductor while holding wire geometry and cable route fixed.
- Restore the baseline. Use A/B/A repetition to expose connector repeatability, environmental drift and accidental geometry changes.
- Repeat band by band. The wire mode, transformed impedance, exterior-current distribution and choke impedance all change with frequency.
- Qualify at operating power. With the analyser safely removed, check transformer, choke, cable and connector heating and station RF under the intended power and duty cycle.
Never connect a VNA or antenna analyser to a live transmitter. Discharge static safely, follow the instrument's maximum-input limits and keep people clear of the antenna during transmit tests.
Read Changes as Evidence, Not Instant Diagnosis
| Observation | Possible mechanism | Useful next check |
|---|---|---|
| Radio-end R + jX changes after adding cable, while the current map and return boundary remain stable | Ordinary differential transmission-line transformation | Compare with a characterised line model or shift the calibrated plane |
| Longer cable shows lower source-end SWR | Round-trip attenuation may be hiding reflection | Measure cable loss and estimate or de-embed the antenna-plane reflection |
| Rerouting the cable or moving the choke changes both S11 and exterior current | The shield exterior is participating in the installed antenna | Map current at more positions and restore the original route in an A/B/A test |
| Touching the analyser, cable or transformer changes the trace | The instrument body or operator may have joined an undefined return path | Define the return conductor, add a measured current boundary and repeat remotely |
| Low SWR coincides with transformer or choke heating | Power may be dissipated in the matching or suppression network | Measure loss with a representative load and repeat the thermal test at declared duty cycle |
| One band improves while another moves | The multiband current distribution and common-mode network differ by frequency | Repeat the full impedance and current map on every required band |
Primary Measurement and Transmission-Line References
- Keysight, S-Parameter Design—travelling waves, reflection coefficient, characteristic impedance and two-port network representation.
- Keysight, Techniques for Precise Cable and Antenna Measurements in the Field—calibration, return loss, VSWR, insertion loss and cable effects.
- Rohde & Schwarz, VNA Calibration Methods and Standards—systematic-error correction and locating the calibration reference plane at the device interface.
- ARRL, Let's Talk Transmission Lines—standing waves, impedance changes along a mismatched line and added line loss.
- Bockelman and Eisenstadt, “Combined Differential and Common-Mode Scattering Parameters”—a rigorous separation of differential and common-mode network behaviour.
- TDK, Measuring Common-Mode and Differential-Mode Choke Impedance—distinct test connections for the two modes.
- Tom Rauch, W8JI, Common-Mode Current—installed coax-exterior current paths, balance and location-dependent suppression.
Joeri's Bottom Line
For an end-fed antenna, I want the SWR trace tied to a circuit drawing. If the coax exterior is the chosen return conductor, I treat the section up to the choke as part of the antenna. If it is not intended to carry current, I establish a measured boundary where that current should stop.
Then I compare the same installed geometry at declared reference planes. That tells me whether a change came from differential line transformation, attenuation, the transformer or the exterior-current path. Only after those mechanisms are separated does an SWR curve become useful engineering evidence.
Mini-FAQ
- Why does end-fed SWR change when coax length changes? The added cable transforms impedance and adds loss. If its exterior is inside the intentional return boundary, changing its length also changes the installed antenna.
- Is the SWR at the radio the antenna's true SWR? It is the SWR at the radio-side reference plane. Cable transformation and attenuation separate it from the reflection at the antenna terminals.
- Should the first choke be one quarter wavelength from the transformer? Not as a universal rule. Place it at the intended exterior-current boundary and verify current around the complete coax on each operating band.
- Does a low SWR prove that the transformer is efficient? No. A matched input can coexist with transformer, choke, cable or ground loss. Transmission, temperature and stress need separate checks.
- Can I measure an end-fed wire directly at the transformer? Yes, for a clearly defined plane and low-power test, but the intended return structure must be present. Disconnecting it creates a different circuit.
- What should I save besides the minimum SWR? Save frequency, R + jX or complex S11, calibration plane, cable data, geometry, choke position, exterior-current map and an A/B/A restoration result.