UNUN or Current Balun? Let the Installed Load Decide

After years of measuring HF feed systems, I keep coming back to one practical architecture: when the installed load is unbalanced and its impedance calls for transformation, use an UNUN for that differential transformation and a separately characterised choke where the unwanted exterior-current path should end. But that is a measured default, not proof that current baluns are inherently inefficient.

First Ask: Efficient at Which Boundary?

“This transformer is efficient” sounds precise until we ask what power entered, what power left, which mode carried it and where the measurement plane was placed. Several different effects are routinely folded into one SWR reading:

A transformer can absorb power and make the input SWR look better. A low-loss transformer can also feed an antenna whose return current heats soil or flows on the station cable network. Neither observation belongs to the topology name alone.

Geometry Does Not Guarantee Balance—But It Does Not Forbid It

A centre-fed dipole, loop or open-wire-fed antenna may begin with symmetric geometry. Feedline route, unequal height, nearby metal, soil, support ropes, station wiring and different capacitance from each side to the surroundings can disturb that symmetry. Real installations often develop common-mode current even when the drawing looks balanced.

That does not mean every HF antenna is electrically unbalanced by nature. A carefully installed symmetric load with a controlled feeder can remain sufficiently balanced for a current balun to do its intended job. The term balanced should describe measured voltage and current relationships at declared ports—not our confidence in the sketch.

Roy Lewallen, W7EL, demonstrated the central mechanism in Baluns: What They Do and How They Do It. A balun at an antenna feedpoint is useful when it prevents the feedline from becoming an unintended third antenna conductor. The analysis does not say that equal-and-opposite current itself wastes energy.

A Current Balun Does Not Spend Power “Forcing Balance”

An ideal current balun passes the wanted differential current and presents high impedance to the unwanted common mode. It does not consume a fixed portion of transmitter power to make two currents equal. Real components have conductor, dielectric and magnetic loss, so they can dissipate power—but the amount depends on winding, material, frequency, load, common-mode voltage, drive and temperature.

If a choke or current balun intercepts common-mode current, part of the unwanted-mode energy may be reflected, redistributed or dissipated. That is not automatically a loss of useful radiated power: without the device, the exterior current may radiate in an unwanted direction, alter the intended pattern, couple noise into the receiver or put RF on station wiring.

The correct question is whether the complete installation delivers more of the accepted power into the intended field with acceptable heating and current distribution. A transformer temperature or one SWR trace cannot answer that alone.

What the UNUN-Plus-Choke Architecture Buys

An UNUN connects intentionally unbalanced ports and can provide the differential impedance transformation indicated by the measured load. A separate common-mode choke can then define where current on the outside of the coax, mast or another return conductor should stop.

That separation is useful because the two functions rarely have identical design windows:

• the transformer is tested for ratio, insertion loss, voltage/current stress and temperature with representative complex loads;
• the choke is tested for complex common-mode impedance, heating and voltage across the unwanted mode;
• the choke position follows the intended return or counterpoise boundary rather than a universal distance from the transformer; and
• installed current is measured before and after that boundary on every operating band.

This is why I favour the architecture for many real amateur installations. It remains understandable when the feedpoint is unbalanced, and it does not ask one winding to prove two different functions with one SWR curve.

It is not the only valid architecture. A deliberately arranged hybrid can combine transformation and choking in one assembly when both functions are independently verified. A current balun can be the better answer for a genuinely balanced installed load. “UNUN plus choke” is a robust design method, not a ban on other circuits.

Nominal Ratio Must Follow the Complex Load

A 4:1 or 9:1 label is the square of an intended ideal voltage ratio under stated conditions. It does not turn every 200 Ω or 450 Ω-looking point into 50 Ω across a band. Real HF antennas present a resistance and reactance that change with frequency, geometry, ground, return path and nearby conductors.

When reactance is large, internal voltage and current can exceed what a resistive calculation predicts. Magnetising impedance, leakage inductance, winding capacitance and the transmission-line impedance of the conductors all modify the transformation. Core loss also changes with frequency, flux, temperature and waveform.

For that reason, a transformer can be excellent with one load family and inefficient or overstressed with another. There is no defensible universal table assigning 0.05–0.2 dB to all UNUNs, 0.1–0.3 dB to all current baluns and 1–2 dB to “real” balanced antennas. Those numbers require the exact devices, fixtures, loads, frequencies, drive and uncertainty.

Commercial Catalogues Do Not Prove an Antenna Rule

RF component catalogues contain unbalanced-to-unbalanced transformers, balanced-to-unbalanced transformers, hybrids, splitters, combiners and devices intended for differential semiconductor ports. The number of products in each category reflects its markets, frequency ranges and naming conventions. It does not prove that one topology is always more efficient for HF antennas.

Broadcast arrays, test probes, mobile radios and receiver front ends also solve different port and current-path problems. A tapped network in a transmitter or a transformer beside an integrated circuit cannot be copied into an antenna argument without carrying along its source impedance, load, balance requirement, shielding and operating frequency.

I would rather use one measured antenna load than a hundred catalogue entries. The installed current path is the evidence we need.

A Fair Comparison on the Bench and at the Antenna

1. Freeze the objective. State whether the comparison concerns differential insertion loss, common-mode suppression, temperature, accepted power or realised field in a direction.
2. Measure the installed load. Record complex impedance at the proposed transformer plane on every band, with the intended return structure connected.
3. Use equal reference planes. Calibrate or characterise fixtures so that connector and line changes do not masquerade as transformer loss.
4. Test representative loads. Include the resistance and reactance the device will actually see, not only a convenient 50 Ω termination.
5. Separate mismatch and dissipation. A two-port or calibrated power method should distinguish reflected power from power absorbed inside the transformer.
6. Measure balance and common mode separately. Record output amplitude/phase or load currents, plus current around the complete feedline exterior.
7. Apply realistic drive. Check temperature, voltage and current at the intended waveform and duty cycle after thermal equilibrium.
8. Compare the installed field. Use the same accepted power, antenna geometry and cable route with A/B/A restoration; repeat on every relevant band and direction.

The Fair-Rite technical catalogue shows why a broadband transformer has distinct low-, mid- and high-frequency loss regions and why ferrite data, winding and parasitics belong together. Ruthroff's original Some Broad-Band Transformers likewise combines circuit analysis with measured response. That is the standard I want for any efficiency claim.

Primary Sources

• Roy Lewallen, W7EL — Baluns: What They Do and How They Do It: original analysis and experiments on feedline imbalance and voltage- versus current-balun behaviour.
• C. L. Ruthroff — Some Broad-Band Transformers: original circuit analysis and measured transmission-line-transformer response.
• Gustav Guanella — New Method of Impedance Matching in Radio-Frequency Circuits: foundational transmission-line-transformer method and series/parallel connection framework.
• Fair-Rite Products — 17th-edition technical catalogue: manufacturer treatment of broadband-transformer equivalent circuits, ferrite behaviour and insertion-loss regions.

Joeri's Bottom Line

I do not ask a transformer name to decide whether an antenna is balanced. I map the installed current. If the load and return structure are unbalanced, I usually choose the UNUN ratio from the measured differential impedance and specify the choke as a separate component at the current-path boundary.

That approach covers a large part of real amateur practice because it keeps the two jobs visible. It does not make a current balun wasteful. On a genuinely balanced installed load, a properly designed current balun can transfer the wanted power with very low loss while keeping the feedline out of the antenna. The winning architecture is the one that proves its loss, stress and current distribution in the system where it will operate.

Questions, antenna-factor records or height trials to share? Contact RF.Guru.