Choosing Between Near-Resonant and Non-Resonant End-Fed Wires
Choosing Between Near-Resonant and Non-Resonant End-Fed Wires
A wire placed near useful resonances can reduce the matching work on selected bands. A deliberately non-resonant wire can cover a wider set of frequencies through a suitable tuner. Neither description proves efficiency, pattern or an easy load. The useful comparison begins with the installed complex impedance, then follows every current path and every loss.
When I can shape a wire so that important operating bands fall near manageable antenna-side loads, I prefer to start there. It may reduce transformation range, circulating energy and retuning. That is a practical design preference—not proof that a near-resonant wire radiates better than a non-resonant one.
My comparison rule: declare the reference plane, measure resistance and reactance across every intended band, account for tuner, transformer, feedline and return-path loss, then compare the installed patterns. A topology wins only when those results support it.
What “Near Resonance” Actually Says
At a declared antenna-side reference plane, write the load as ZA = RA + jXA. Resonance occurs where the reactance XA passes through zero. “Near resonance” should mean that the remaining resistance and reactance are within a range the chosen network can handle with acceptable loss, voltage, current and temperature. It is not a universal SWR threshold.
A resonant resistance can be far from 50 Ω. A matching network can also present 50 Ω to the transmitter while the wire is not resonant. Resonance, transmitter match, low SWR and radiation efficiency are four different results. Keeping them separate prevents a convenient match from becoming an imaginary efficiency certificate.
On a multiband wire, the useful resonances are modes of the complete installed structure—not simply whole-number multiples printed beside a wire length. Height, slope, bends, conductor diameter, insulation, the return branch, coax exterior, soil and nearby metal all move the measured impedance.
What a Deliberately Non-Resonant Wire Offers
A non-resonant end-fed wire deliberately accepts a complex load and asks a tuner, sometimes together with a broadband transformer, to convert it to the transmitter's required impedance. That can be an excellent answer when supports limit the available length, one wire must cover many frequencies or a remote coupler can sit directly at the antenna terminal.
The word “non-resonant” does not mean unpredictable or inefficient. A fixed geometry produces repeatable impedance and current distributions in a stable environment. It does mean that the matching system must have enough range and margin for the actual loads, and that its losses belong in the result.
A remote tuner at the antenna feedpoint can keep a long coax run near its design impedance. A shack tuner may instead leave high standing-wave voltage and current on the transmission line between tuner and antenna. Real feedline then dissipates more power than its matched-loss figure alone suggests. This is an architecture question, not a verdict against non-resonant wires.
Transformer Ratios Are Starting Hypotheses
A nominal 4:1 or 9:1 impedance ratio describes an ideal transformation. It does not tune away arbitrary reactance, and it does not guarantee that the resulting load lies inside a tuner or transmitter range. The correct ratio follows the measured load region, not the wire's marketing category.
A real UNUN adds magnetizing impedance, leakage inductance, winding capacitance, copper resistance and core loss. High load resistance may create difficult voltage; low resistance may create difficult current; large reactance may raise circulating energy. Frequency, power, modulation duty cycle, enclosure and cooling decide whether a network remains within its electrical and thermal limits.
This is why a near-resonant wire can be easier on one selected band yet awkward on another, and why a non-resonant wire can be gentle when its coupler sees a favourable installed load. Measure the complete load sweep and test the network under representative complex loads before assigning a power or efficiency conclusion.
Count Loss at the Same Reference Plane
Use the same input boundary when comparing two systems. A useful loss ledger includes:
- Matching network: component resistance, dielectric loss and circulating current or voltage under the installed load.
- Transformer: winding, core, leakage and stray-field loss under frequency, power and thermal conditions.
- Feedline: matched attenuation plus the additional dissipation created by the actual standing-wave distribution.
- Return network: intentional counterpoise, coax exterior, mast, soil and capacitive coupling to surrounding structures.
- Radiator: conductor, loading, joints and nearby-material loss before accepted power becomes radiation.
Radiation efficiency is radiated power divided by accepted power at a stated antenna boundary. System efficiency may include upstream transformer, tuner and feedline losses. Both can be useful, but they are not interchangeable. A low SWR at the radio reports neither value.
| Design question | Near-resonant starting point | Non-resonant starting point |
|---|---|---|
| Matching burden | May be modest on selected measured modes; no guarantee of 50 Ω. | Usually expects a wider transformation range; actual burden follows the complex load. |
| Feedline choice | Coax may be practical where the matched load and loss are acceptable. | A feedpoint coupler or low-loss line may prevent a severe mismatch from occupying a long lossy coax run. |
| Band coverage | Chosen modes may simplify operation on priority bands. | A suitable coupler may reach frequencies between or beyond those modes. |
| Efficiency | Not established by resonance. | Not disqualified by non-resonance. |
| Pattern | Follows the installed current distribution on each band. | Follows the same physics; electrical length and environment can create several lobes and nulls. |
| Best evidence | Complex load, network loss and stress, exterior-current map, accepted-power pattern and repeatable field comparison. | |
Every End-Fed Wire Has a Return Branch
Current leaving one terminal must return to the other. A drawing that shows only the main wire has not removed the return path; it has left that path to the coax exterior, a dedicated counterpoise, mast, station wiring, soil and distributed capacitance.
That return structure changes the impedance, common-mode current and radiation pattern. It can be intentional: a declared section of coax exterior may form an antenna branch between transformer and choke. If so, the choke belongs at the planned current boundary and needs adequate measured common-mode impedance on every operating band. Placing a choke automatically at the transformer can remove a needed return branch; placing it only at the shack can let the whole feedline and station participate.
Transformation and common-mode suppression remain separate functions. The UNUN handles the differential impedance conversion for which it was selected. A separately specified choke controls exterior current at the chosen boundary. Neither function proves the other.
Pattern Is Not Promised by the Matching Category
At HF, the installed current distribution, electrical length, height, orientation, bends, ground and nearby structures form the azimuth and elevation pattern. A wire that is near one resonance may have a simple current distribution on that band and several current maxima on a higher band. A non-resonant wire can do the same as its electrical length changes.
Therefore “near resonant” does not guarantee a repeatable DX lobe, and “non-resonant” does not mean that lobes move randomly. Both patterns can be modelled when the complete geometry and environment are known. Both should be checked on every intended band rather than inferred from SWR.
Adding a long return wire to obtain another band changes the antenna into a different two-branch structure. It may alter impedance and pattern usefully, but no fixed added length guarantees top-band coverage, NVIS performance or efficient full-HF operation. Model and measure the resulting system as a new antenna.
Where Each Approach Earns Its Place
I favour the near-resonant starting point when the priority bands show manageable installed loads, the network operates with comfortable loss and stress margins, and the resulting patterns serve the intended paths. It can reduce boxes, tuning range and opportunities for loss—but only when the measurements say so.
I favour the non-resonant route when the site fixes the wire length, when continuous frequency coverage matters, or when a suitable feedpoint coupler and return structure can be installed. It can be just as serious an engineering solution. The tuner is then part of the antenna system, not an embarrassing accessory.
Convenience is also a valid design objective. Fast band changes, available supports, weather protection, access for maintenance and safe voltage clearance may decide the better antenna even when two systems have similar RF performance.
A Measurement-Led Selection
- Document the complete geometry: main wire, return branch, feedline route, choke, mast, bonds and nearby conductors.
- Measure the antenna-side complex impedance: record resistance and reactance across every intended band at a declared plane.
- Test the matching hardware: reproduce the expected loads and verify transformation, loss, voltage, current and temperature at representative power and duty cycle.
- Account for the line: use its length, frequency-dependent matched loss and actual load rather than a generic cable claim.
- Map exterior current: confirm which return branch is intentional and where the choke establishes the boundary.
- Check radiation evidence: use a complete-geometry model or controlled accepted-power A/B/B/A measurements, and restore the baseline when propagation changes.
Bottom line: starting near useful resonances can make a multiband end-fed system easier to engineer. It is not an automatic efficiency or pattern advantage. A non-resonant wire can be excellent when its tuner, feedline and return path are chosen as one measured system. Compare the complete installation, not two labels.
Primary technical references
- IEEE Std 145-2025 — Standard for Definitions of Terms for Antennas
- Recommendation ITU-R BS.705-2 — HF transmitting and receiving antenna characteristics and diagrams
- Keysight — Impedance Measurement Handbook, sixth edition
- Icom AH-730 — Official tuner range and antenna-condition specifications
- Recommendation ITU-T K.136 — RF electromagnetic fields along telecommunications conductors
Mini-FAQ
- Does resonance mean that an antenna is efficient? No. Resonance says that reactance is zero at a declared reference plane. Conductor, ground, transformer, tuner and feedline losses still decide efficiency.
- Does a near-resonant wire work without a tuner? Not necessarily. Resonant resistance may be far from 50 Ω, and the installed geometry can move both resistance and reactance.
- Is a 9:1 UNUN always correct for a non-resonant wire? No. The ratio must follow the measured complex load and the verified range, loss and stress of the complete matching system.
- Is a non-resonant wire inevitably lossier? No. A suitable low-loss feedpoint coupler and return structure can perform very well. Loss must be measured for the installed architecture.
- Should the common-mode choke always sit at the transformer? No. It belongs at the intended end of the antenna-side exterior-current branch, with sufficient measured impedance on each band.
- Which type is better for multiband operation? The one whose measured loads, matching loss, stress margins, return-current control and installed patterns best serve the required bands and paths.