Open vs Closed Antennas — Resonance vs Traveling Wave
Open vs Closed Antennas — Resonance vs Traveling Wave
Open-ended wires, resonant loops and resistively terminated antennas are useful design families. Their real behaviour still comes from boundary conditions, current distribution, loss, ground and the installed electromagnetic environment.
The familiar three-way comparison is worth keeping: an open-ended resonant wire, a resonant loop and a resistively terminated antenna are not the same structure. But “open” and “closed” do not decide efficiency, noise, bandwidth or radiation pattern. They only begin the current-path description.
Joeri’s short version: start with the conductor geometry, then solve or measure the complex current along it. Resonance and traveling-wave behaviour are properties of that complete boundary-value problem—not rewards assigned to an antenna because its outline is open or closed.
The Three Families Are Design Intent, Not Laws of Nature
| Useful family | Typical examples | First question to ask |
|---|---|---|
| Open-ended resonant wire structure | Dipole, inverted V, monopole with its return system, parasitic Yagi element | Where do feed, open-end, coupling and return-path boundaries place the current maxima, minima and phase changes? |
| Resonant continuous loop | Full-wave delta, quad or horizontal loop; electrically small loop in a different operating limit | Which loop mode is excited, and how do circumference, feed position, conductor, shape and surroundings set its current distribution? |
| Resistively terminated antenna intended for traveling-wave operation | Beverage, terminated rhombic, terminated folded dipole | How much incident wave reaches the termination, how much reflects, how much radiates, and how much power is lost in the resistor, ground and conductors? |
A Beverage is not usefully classified as “closed” in the same geometrical sense as a loop. A quarter-wave monopole is not a one-conductor RF circuit merely because its radiator has an open end. A terminated folded dipole can support both forward and reflected components. The family name tells us the intended operating mechanism; the installed fields tell us what actually happened.
Boundary Conditions Create the Current Distribution
At the end of an ideal thin open conductor, longitudinal conduction current must fall toward zero. Charge accumulates and the electric field becomes strong near the end. That boundary helps establish a standing-current distribution, but the picture is not a packet of RF bouncing from a perfectly marked tip. Radiation, conductor diameter, bends, loading, feed geometry and nearby matter make the structure distributed.
A resonant wire normally carries forward and backward current components whose phase relationship produces maxima and minima. At a declared feed plane, a resonance is commonly identified where the input reactance crosses zero. A matching network can cancel reactance at another plane without proving radiator resonance. Neither result proves low conductor loss, low ground loss, a particular pattern or good matching to 50 Ω.
A monopole adds a radial, ground-screen, vehicle-body, mast or environmental return system. A Yagi adds mutual coupling among driven and parasitic elements. Those structures may have open conductor ends, but their behaviour follows the complete coupled system.
Keep two statements separate: an ideal open end imposes a local current boundary; a resonant input is a property measured at a specified port and frequency. One does not guarantee the other.
A Closed Loop Does Not Contain Its Fields
A loop closes the conducting path around its perimeter, apart from the feed or loading network. It does not close electromagnetic space. Time-varying loop current produces electric and magnetic fields, stores near-field energy and radiates into the far field.
A full-wave loop is commonly resonant, with current magnitude and phase varying around the perimeter. An electrically small loop is a different limit in which current may be approximately uniform, radiation resistance is small and conductor or matching loss can dominate. “Loop” therefore covers structures with very different impedance, efficiency and pattern behaviour.
Shape, circumference, conductor diameter, feed point, polarization, height, ground and nearby conductors all matter. A vertically oriented loop can be fed for different polarizations and azimuth patterns. A horizontal loop changes elevation pattern as its electrical height and circumference change. No closed outline guarantees smooth lobes, broad bandwidth or efficiency equal to—or better than—a dipole.
The same applies to receive noise. A loop may sound quieter because its pattern rejects a dominant local source, its feed system carries less unwanted common-mode current, its polarization differs, or its total gain is lower. None of those outcomes follows from closure alone.
A Termination Reduces Reflection; It Does Not Abolish It
For a uniform transmission-line approximation with characteristic impedance Zc and load ZL, the load reflection coefficient is:
ΓL = (ZL − Zc) / (ZL + Zc)
Only ZL = Zc makes that ideal local reflection zero. A real radiating wire is not a uniform, shielded transmission line: its effective propagation and impedance depend on height, ground, taper, bends, coupling and continuous radiation. Feed transitions, supports, the termination network and environmental discontinuities can all launch a backward component.
The current along a finite terminated antenna is therefore generally the sum of attenuating forward and backward components. “Traveling-wave antenna” describes the design goal of making one component dominant over the intended band. It does not mean that a standing-wave component is mathematically or physically absent.
The Beverage, Rice and Kellogg wave-antenna work is useful because it ties directivity to wave build-up along a long wire, its earth relationship, propagation and termination—not to a resistor in isolation. Countryman’s terminated folded-dipole measurements likewise describe an experimental antenna with a stated geometry and frequency span, not a universal property of every loaded folded wire.
Termination Power Belongs in the Efficiency Ledger
At one declared antenna input boundary:
Paccepted = Pradiated + Ptermination + Pconductor + Pground + Pmatching/feed + Pother loss
ηradiation = Pradiated / Paccepted
The resistor’s first-order sinusoidal dissipation is IR,rms2R at its own terminals. That current changes with frequency, propagation loss, loading and reflection. A termination rating must therefore cover actual voltage, current, average power, waveform, duty cycle, enclosure and ambient temperature—not transmitter output power alone.
A resonant antenna can be inefficient when conductor, loading-coil, matching, common-mode or ground losses are high. A terminated antenna can radiate a useful fraction of accepted power when it is electrically long and deliberately designed, but the termination still consumes real power. Geometry labels cannot replace this ledger.
Bandwidth Has More Than One Meaning
| Bandwidth claim | Evidence required | Common trap |
|---|---|---|
| Input-impedance bandwidth | Complex impedance or reflection coefficient at a calibrated reference plane | A resistor can make SWR look broad while dissipating power. |
| Efficiency bandwidth | Accepted and radiated power, including termination, conductor, ground and matching losses | Assuming a smooth input match means most power is radiated. |
| Pattern bandwidth | Gain or normalized patterns over frequency, polarization and the relevant elevation/azimuth region | Assuming a stable feed impedance means stable lobes, nulls or beam direction. |
| Receive-SNR bandwidth | Wanted-signal and noise measurements with receiver state, propagation, site and time controlled | Calling a lower noise-floor reading better reception without checking signal level. |
| Power bandwidth | Voltage, current, termination temperature, matching loss and spectral cleanliness at declared drive | Applying small-signal impedance data as a universal transmit rating. |
A Beverage can produce a valuable low-band receive pattern while sacrificing absolute gain and radiation efficiency. A T2FD can offer broad input behaviour while its pattern changes as its dimensions span different fractions of a wavelength. A terminated rhombic can be directive, yet beam angle and sidelobes still vary with electrical length, height and ground. “Wideband” needs the noun after it.
Noise Is a Pattern-and-System Result
The receive port integrates environmental brightness through the antenna’s gain and polarization pattern. It also sees loss-generated thermal noise, feed-system pickup, common-mode coupling and receiver noise. A loop does not reject noise as a class, and a terminated antenna is not quiet merely because it is non-resonant.
ITU-R’s radio-noise measurement guidance explicitly treats antenna pattern, ground, obstructions, earthing, antenna factor, receiver overload and uncertainty as measurement conditions. That is the right framework for an antenna comparison:
- Use the same receiver bandwidth, attenuation, preamplifier, AGC state and detector.
- Record wanted signal and noise separately; compare SNR, not only the displayed noise floor.
- Control time and propagation, preferably with simultaneous receivers or rapid switching.
- Record polarization, azimuth, elevation response, height, ground and feedline route.
- Map feedline-exterior current so a change in common-mode pickup is not mislabelled an intrinsic antenna property.
- Repeat A/B/A and include drift and measurement uncertainty.
Prove the Installed Antenna
LLNL’s Numerical Electromagnetics Code can model wire structures, sources, loads, ground and current distribution. Its validation examples include loaded long-wire structures over ground—the kind of full boundary needed here. A model is most useful when its impedance and current predictions first agree with measurement.
| Question | Measurement or model output |
|---|---|
| Is it resonant at the intended boundary? | Complex input impedance with feedline and fixture effects declared. |
| Is one traveling component dominant? | Complex current magnitude and phase along the structure, with the termination changed or removed for comparison. |
| Where does accepted power go? | Radiated-power estimate plus measured or modeled termination, conductor, matching and ground loss. |
| Is the pattern useful? | Gain and polarization patterns over the intended band in the installed geometry, not one free-space trace. |
| Is reception better? | Repeatable wanted-signal, noise and SNR records under controlled receiver and propagation conditions. |
| Is it safe at power? | Peak voltage/current, resistor and matching temperature, RF exposure and access boundaries under declared waveform and duty cycle. |
Primary and Authoritative References
- Beverage, Rice and Kellogg — The Wave Antenna: A New Type of Highly Directive Antenna
- G. L. Countryman, W3HH — Performance of the Terminated Folded Dipole
- IEEE Open Journal of Antennas and Propagation — Loop current distribution, impedance and pattern analysis
- MIT Lincoln Laboratory — Forward/reflected waves, load reflection coefficient and standing-wave boundary
- Lawrence Livermore National Laboratory — NEC-5 validation for wires, loads and ground
- NIST and IEEE TAP — Radiation efficiency versus total efficiency and measurement uncertainty
- ITU-R SM.1753-2 — Radio-noise measurement, antenna pattern, site and uncertainty
Joeri’s Bottom Line
Open-ended resonant wires, resonant loops and terminated traveling-wave designs remain useful ways to organize antenna thinking. Just do not let the outline become the explanation.
Trace the boundary conditions and complex current. Measure the termination power. Separate impedance, efficiency, pattern and SNR bandwidth. Include polarization, ground, feedline common mode and the installed environment. Then the three families become engineering tools instead of three new antenna myths.
Mini-FAQ
- Does an open-ended wire have to be resonant? No. An open end imposes a local current boundary, but resonance depends on the complete geometry, feed, loading, coupling, return system and measurement plane.
- Are loops traveling-wave antennas? Not inherently. A full-wave loop commonly supports a resonant standing-current distribution. A loop can also be loaded or terminated, so its actual current must be solved or measured.
- Do loops contain their fields or always receive less noise? No. Loops radiate electric and magnetic fields. Received noise depends on their installed gain and polarization pattern, loss, common-mode coupling, site and receiver.
- Does a terminating resistor remove every reflection? No. It minimizes one local reflection only when it matches the effective wave impedance. Radiation, ground, geometry and other discontinuities can still produce backward current.
- Does broad SWR prove a good traveling-wave antenna? No. Input-impedance bandwidth is separate from radiation efficiency, pattern, receive-SNR and power bandwidth. The termination may be creating the broad match by dissipating power.
- How should I compare these antenna families? Declare the reference plane and installation, then compare complex current, accepted and radiated power, termination loss, gain/polarization pattern and controlled wanted-signal-to-noise measurements.