Reducing Coupling in Quarter-Wave Fan Verticals

The attractive fan-vertical story says that the branch nearest resonance accepts the current and the other wires politely wait for their bands. That can be a useful first approximation. It becomes dangerous when we use it as proof. Closely spaced branches share electric and magnetic fields, so trimming the feedpoint back to a comfortable impedance can conceal a different current division—and a different antenna.

Every Branch Belongs to One Coupled Network

Each radiator has a self-impedance. Every pair also has mutual impedance because current on one conductor produces fields that induce voltage and current on another. A compact statement for the feed-region network is:

The current on branch i therefore depends on its own impedance and on the magnitude and phase of current on all neighbouring branches. Spacing, angle, length, conductor diameter, loading, support dielectric, feedpoint structure, mast, radial system, feedline and nearby metal all change that result.

This is more complete than saying that parallel wires form a capacitance near their voltage maxima. Electric-field coupling can be important near high-voltage regions, but magnetic coupling follows current, and the integrated mutual impedance belongs to the entire geometry. A single “capacitive coupling” label cannot predict branch currents or pattern.

A Match Can Be Recovered After the Pattern Has Changed

A VNA at the feedpoint reports the impedance of the complete network at its calibration plane. It does not say which conductor carries the current, whether induced current is in a useful phase, how much power is dissipated or where the far-field energy goes.

Adding a branch can pull several resonances. Trimming those wires may return the feedpoint to a low SWR. A tuner can transform the same changed load into something the transmitter accepts. Neither action restores the earlier current distribution automatically.

Nearby Bands Increase the Verification Burden

When branch resonances are widely separated, one conductor may dominate while the others present relatively high impedances. As the intended bands move closer, the branches can have more comparable electrical lengths and impedances. Their resonances, current distributions and end effects then become more strongly entangled.

No wavelength range is immune, and the problem is not automatically worst from 15 to 40 metres. Physical spacing must be normalized to wavelength, while electrical length, frequency ratios, support geometry and the radial/mast environment also matter. A fixed separation in centimetres has very different meaning on 40 metres and 10 metres.

Harmonic relationships can create another interaction. A wire that is far from its fundamental can still support a higher-order current mode near another band. This is why a branch plan should be modelled across every required frequency, not only at the frequency printed on each wire label.

Fewer Branches Can Be the Better Multiband Design

One dedicated quarter-wave radiator per band is an option, not a rule. It can work well when the geometry leaves enough separation and the complete current system is verified. It can also create a dense structure in which every adjustment moves several bands.

Before adding another branch, ask what it buys:

• Does it provide bandwidth that an existing branch cannot cover?
• Does it avoid unacceptable tuner or feedline loss?
• Does it preserve a required current distribution and elevation pattern?
• Does it reduce switching or deployment complexity enough to justify the interaction?
• Can its effect be measured and reproduced after wind, rain and maintenance?

If the answer is only “it produces another low-SWR dip,” the new conductor has not yet earned its place. A narrower-band tuner-assisted operating range may be a better compromise than another close resonator, provided tuner loss, component stress, feedline current and pattern remain acceptable.

The opposite can also be true. A deliberately staggered group of nearby conductors may create useful bandwidth when its combined current and pattern have been designed for that purpose. The point is not that dense fans fail; it is that their operating unit is the coupled set, not one labelled wire at a time.

A Loading Coil Does Not Make Coupling Disappear

A series loading coil adds inductive reactance and can shorten a radiator physically. That may reduce the length running beside adjacent branches. It does not isolate the loaded wire by definition. The coil has electric and magnetic fields, distributed capacitance, conductor loss and coupling to nearby wires and metal.

The magnetic field is not “self-contained,” and its relevance is not decided merely by calling it orthogonal to a neighbouring element. Coil axis, spacing, winding dimensions, shield or enclosure, nearby conductors and the current through both structures determine mutual coupling.

Coil position is a trade rather than a universal percentage:

• Closer to a current maximum: less inductance may produce the required electrical shortening, but coil current and I²R loss can be larger.
• Farther toward a current minimum: more inductance is normally required, with higher local voltage and greater sensitivity to winding capacitance and surroundings.
• At any position: Q, self-resonance, voltage spacing, weather, support mechanics and the resulting full-wire current taper must be checked.

A 60–75% height prescription cannot balance those variables for every band, wire length and coil. Choose position in the complete model, build the real coil, measure its R+jX and temperature, then put that measured network back into the antenna model.

Linear Loading Is a Geometry, Not a Defect

A folded or parallel return section creates distributed inductance, capacitance and mutual coupling. It can shorten a radiator and may be useful in a constrained installation. It can also introduce loss, high voltage, current cancellation or strong sensitivity to spacing.

Calling linear loading “theoretically inductive” and then blaming incidental capacitance is too simple. Both field mechanisms are present from the beginning. The result depends on conductor separation, overlap, fold orientation, wire diameter, insulation, support and proximity to the other fan branches.

There is no universal 15 cm threshold. Record separation as a physical dimension and as a fraction of wavelength, then model and measure the assembled geometry. A compact coil and a folded conductor are alternatives with different current, voltage, loss and coupling profiles—not a clean solution and a dirty one.

Spacing and Symmetry Help Only When the Measurements Agree

Increasing branch separation or changing branch angles often reduces some mutual terms, but “as wide as practical” is not a design value. A star layout can distribute conductors mechanically and can make one aspect of the geometry symmetric. It can also alter the azimuth pattern or create several similar coupling paths.

Supports and guys matter. A conductive mast or wet rope can become part of the coupled system. Radial currents can be unequal, and current on the coax exterior can add another radiating branch. Include those conductors in the model and current map rather than assuming radiator symmetry guarantees system balance.

Low Takeoff Angle Must Be Demonstrated

A ground-mounted quarter-wave-like vertical can favour low-elevation radiation in an appropriate installation. A fan label does not preserve that result automatically. Induced current on adjacent branches can lengthen the effective structure, introduce higher-order modes or move current toward lossy regions. Ground and radial loss, mast current, feedline common mode and surrounding objects change the pattern again.

Input impedance and far-field pattern are different outputs. LLNL's NEC reports conductor current and radiation patterns separately for exactly this reason. A credible low-angle claim needs a complete-geometry model with the installed ground and terminations, or a calibrated field/pattern comparison at equal accepted power.

A Measurement Sequence That Exposes Coupling

1. Freeze the baseline geometry. Record every branch, loading section, radial, mast, feedline route, choke and nearby conductor.
2. Calibrate at a declared plane. Save complex S₁₁ or R+jX across every operating band rather than only SWR minima.
3. Map branch current. Use repeatable probes and marked positions near the feedpoint and farther along each accessible conductor.
4. Record magnitude and phase where possible. Equal current magnitudes do not establish constructive far-field addition.
5. Change one branch. Remove, shorten, detune or reposition only the questioned element while keeping cable and soil conditions stable.
6. Restore A/B/A. The second A result must reproduce the first within measurement uncertainty before accepting the B difference.
7. Compare the field result. Use a validated model or calibrated A/B/A field measurement at equal accepted power for pattern or low-angle claims.
8. Repeat under operating stress. Record coil and tuner temperature, impedance drift, weather condition, waveform, duty cycle and test duration.

Current clamps around individual radiator wires need a documented transfer response and repeatable placement. Measuring the complete coax with a suitable current probe gives exterior/common-mode current; it does not reveal the intended equal-and-opposite differential currents inside the line. Keep those measurements conceptually separate.

Primary and Authoritative Technical Sources

• ARRL Laboratory — Multiband Dipoles Compared
• NASA Technical Reports Server — Delay-Line and Mutual-Coupling Considerations for Antenna Arrays
• Lawrence Livermore National Laboratory — Numerical Electromagnetic Code v5
• David Birnbaum, K2LYV — Design of a Two-Band Loaded Dipole Antenna
• Coilcraft — Testing Inductors at Application Frequencies
• IEEE 149-2021 — Recommended Practice for Antenna Measurements
• National Bureau of Standards Circular 598 — Techniques for Accurate Antenna-Gain Measurement
• ICNIRP — RF Exposure Guidelines from 100 kHz to 300 GHz

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