Trapped Radials or Separate Radials? Measure the Complete Vertical

A radial is not a magic earth connection and it is not an isolated open-circuit stub. It is a driven conductor in the return side of a monopole system, coupled to the radiator, the other radials, soil, feed line and nearby metal. Add a trap and every one of those relationships can move.

Why the Radial Cannot Be Judged Alone

At a base-fed vertical, current leaving the radiator must be balanced by current in the intended return conductors and any unintended paths. The radial current is therefore set by the complete coupled structure. Its magnitude and phase depend on frequency, radial count, length, height, slope, symmetry, earth properties, the radiator and whatever the coax exterior is allowed to do.

A quarter-wave radial can be near resonance in a particular installation, but calling it an open-ended transmission-line stub skips the most important fact: there is no independent uniform two-conductor line. The radiator and every return path are electromagnetically coupled. The useful model is the complete antenna, not one wire viewed in isolation.

This matters especially on another band. A wire that is close to three-quarter wavelength may carry strong current in one geometry and very little in another. Harmonic length is a clue worth testing, not free gain.

What a Trap Adds

A common antenna trap is a parallel-resonant network inserted in series with a conductor. Around its parallel resonance it presents a high impedance, so the section beyond the trap carries less current and the inner section can act as the shorter-band conductor. Away from that frequency, the network is reactive and the outer section participates to a degree set by the trap and antenna impedances.

A trap is not lossless, but its existence does not prove that its loss is operationally important. A high-Q network at a low-current location may dissipate little power. A compact network carrying substantial circulating current may run hot. One manufacturer's antenna power rating cannot establish the limit of every trap-radial design because the components, placement, enclosure, waveform and thermal environment differ.

What N6LF Actually Found on Several Bands

Rudy Severns, N6LF, tested multiband radial systems rather than deriving their behaviour from wire length alone. His Ground Systems for Multiband Verticals, Part 6 used an adjustable vertical at 7.2, 14.2, 21.2 and 28.5 MHz with measured feedpoint impedance and relative transmission gain.

For elevated operation, a system with four quarter-wave radials cut for each band—16 wires connected together—performed close to selecting only the four radials for the active band. That is useful evidence for a conventional separate-radial solution. It is not evidence that two radials are always sufficient.

The same experiment provides a direct warning about harmonic shortcuts. Four 33-foot radials worked on 40 m, but on 20 m the configuration became unacceptable; on 15 m, where those wires were near three-quarter wavelength, four long radials also performed so poorly that Severns did not enter the data. Configurations with more 33-foot radials behaved differently. Length ratio alone did not decide the result.

In his earlier Ground System Performance for HF Verticals, Part 3, four nearly balanced elevated radials at 7.2 MHz produced almost the same field on the measured path as 64 surface radials. His later work also showed how sensitive sparse elevated systems can be to local current imbalance. Four carefully controlled radials can work extremely well; that does not make every sparse installation predictable.

Separate Radials: The Clear Baseline

Separate radials cut for each band avoid a resonant component in series with the return conductor. That makes them a clean reference against which any compact scheme should be compared.

• Current paths are easier to inspect. A clamp probe or calibrated current transformer can show which branches carry current on each band.
• Tuning interactions are easier to isolate. A branch can be disconnected or adjusted without opening a trap enclosure.
• Component stress is reduced. There is no lumped resonator in the radial branch, though high RF voltage can still exist near open ends.
• Failure modes are visible. Broken wires and corroded terminals are usually simpler to diagnose than a weather-shifted network.

The cost is wire, space and visual complexity. Four radials on each of four bands means 16 conductors if every band receives a dedicated symmetrical set. That may be perfectly reasonable on a roof or tower platform and impossible in a small garden. Practical constraints are part of the specification.

Where Trapped Radials Earn Their Place

A trapped radial can be attractive when the support points or azimuths are limited. One physical branch can provide different effective electrical lengths on different bands, preserving a more symmetrical layout than a dense fan that cannot actually be installed.

That is a defensible compromise when the design is treated as an antenna network rather than a construction recipe:

• characterise each trap under realistic loading, not only with a low-level dip test;
• model or measure the complete coupled structure on every operating band;
• verify radial-current magnitude and phase, including the section beyond each trap;
• provide voltage, current and thermal margin for the intended modulation and duty cycle;
• control feed-line common mode so the coax does not conceal a poor intended return path; and
• repeat measurements after weather exposure and mechanical movement.

A design that passes those checks is not invalid because it contains traps. It is simply a more coupled and component-dependent system than a set of plain wires.

Current Does Not Divide Equally by Permission

Connecting several radials to one plate does not make their currents equal. Current divides according to the complex impedance and mutual coupling of each path. A trap changes both, and nearby soil, gutters, mast sections, guy hardware and cables can break the intended symmetry.

Measure both magnitude and phase when the goal is symmetry. Equal clamp-meter magnitudes can still hide different phases, and one feedpoint SWR cannot reveal the branch distribution. A good match can coexist with loss, pattern distortion or a feed line carrying part of the return current.

The feed-line choke is therefore part of the experiment. Measure its common-mode impedance over the operating range and then measure current on the coax exterior below it in the final routing. A choke that works on one band can become capacitive or insufficient on another.

A Fair Comparison

Compare trapped and separate systems as complete antennas at the same site. Keep the radiator, feedpoint height, feed line, choke, support geometry and nearby objects unchanged as far as practical.

1. Document the as-built geometry. Record wire lengths, heights, slopes, azimuths, conductor type, trap position and nearby conductors.
2. Measure the reference plane. Move the VNA calibration or a verified de-embedding plane to the antenna feedpoint and record complex impedance across every band.
3. Measure branch currents. Use the same calibrated probe position on every radial, recording magnitude and phase if the instrumentation permits.
4. Control common mode. Record coax-exterior current below the choke and keep the feed-line route fixed.
5. Retune fairly. Adjust each configuration to its intended state without hiding matching-network loss.
6. Hold accepted power constant. Do not compare field strength at equal transmitter indication when mismatch or feed-system loss differs.
7. Check more than one direction. Several azimuths help separate a pattern change from a total-efficiency change.
8. Run the intended duty. Track trap and terminal temperature and inspect for drift, corona, tracking or arcing.
9. Repeat. Reconnect the reference configuration and publish repeatability and uncertainty, not only the best result.

Choose the Failure Mode You Can Manage

The Engineering Verdict

Separate band-specific elevated radials are the cleanest baseline. They remove the trap's component loss, thermal stress and tolerance from the problem, and they make branch-current measurements easier to interpret.

That does not prove that they always outperform a trapped system. A well-designed trapped radial can save space and can work well within verified limits. A badly installed set of plain radials can be unbalanced, lossy or dominated by the coax exterior. Topology alone does not certify the result.

The strongest version of “forget trapped radials” is therefore conditional: forget them when you have room for simpler separate conductors and do not need the extra network. Keep them on the table when geometry demands consolidation—and then measure the antenna as a coupled multiband system.

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