Why a Lower-Band Delta Loop Can Disappoint on Upper HF
Why a Lower-Band Delta Loop Can Disappoint on Upper HF
A loop that is one wavelength on a lower band becomes several wavelengths around on upper HF. It can remain an efficient radiator, yet put far less power into the direction and elevation angle you need.
The practical warning is real: an 80 or 40 metre delta loop is not automatically a good 20, 17 or 15 metre DX antenna merely because a tuner finds a match. The important failure mode is usually not a mysterious collapse of radiation efficiency. It is the combination of a multiwavelength current distribution, fragmented three-dimensional pattern, frequency-dependent feed impedance, feed-system loss and a physical height that now represents many more wavelengths.
The useful distinction: radiation efficiency asks how much accepted power becomes radiation. Realised gain asks how much available power appears in a stated direction after mismatch and antenna loss. A multiwavelength loop can radiate efficiently while having poor realised gain toward the path you care about.
The Current Does Not Stop—Its Phase Changes Around the Loop
At its fundamental resonance, a closed loop with a perimeter near one wavelength supports a relatively simple standing-wave current distribution. When the same physical conductor is operated at two, three or four times that frequency, its electrical perimeter becomes several wavelengths. More current maxima and minima appear, and the current phase progresses through more cycles around the loop.
Every segment carrying RF current contributes to the far field. The contributions add in some directions and cancel in others. That interference creates additional lobes and nulls. Saying that “most of the wire no longer radiates” misses the mechanism: the conductor still participates, but its fields no longer combine into the same broadside pattern as the fundamental mode.
The Virginia Tech resonant-loop study shows why loop current amplitude and phase must be solved rather than borrowed from the electrically small-loop model. Once circumference is comparable with wavelength, current is non-uniform and both impedance and pattern become frequency dependent.
Pattern Fragmentation Is Usually the First Practical Penalty
A multiwavelength loop normally develops more azimuth and elevation lobes, separated by deeper nulls. Some lobes can have useful directivity. The difficulty is that their headings and elevation angles are fixed by the complete installed geometry, not by the fact that the wire forms a triangle.
The same physical height also changes meaning with frequency. A loop whose top is modestly high on 80 metres can be several wavelengths above ground on 15 metres. Ground-reflected and direct fields then combine differently, so the upper-band elevation pattern can acquire several maxima and nulls. Those lobes are not universally “high angle” or “low angle”; their positions depend on height, shape, feed location, soil, terrain and nearby conductors.
This is why one strong contact or one attractive SWR dip proves very little. A lobe may happen to favour that path while a deep null suppresses another. For repeatable upper-HF coverage, the complete three-dimensional pattern matters more than the best lobe printed by a model.
Efficiency, Gain and Realised Gain Are Different Questions
Use separate power boundaries:
Radiation efficiency: ηrad = Pradiated / Paccepted at the antenna
Gain: G(θ,φ) = ηradD(θ,φ)
Realised gain: Grealised(θ,φ) = ηmismatchηradD(θ,φ)
A lossless model of a multiwavelength loop has 100% radiation efficiency by definition, yet its directional pattern can be inconvenient. A real installation adds conductor, connector, matching-network, feedline, dielectric, common-mode and environmental losses. Those losses reduce gain, while mismatch reduces realised gain at the chosen reference plane.
The Rohde & Schwarz OTA antenna-parameter guide separates directivity, radiation efficiency, gain and realised gain. That separation is essential here: a fragmented pattern is not itself dissipative loss, but it can reduce field strength in the wanted direction just as decisively.
The Feedpoint Impedance Is Deterministic, Not Chaotic
A loop near its fundamental resonance often begins in a moderate impedance region, but there is no universal 100–120 Ω value. Shape, perimeter, conductor diameter, feed gap, feed position, height, ground and nearby objects all matter.
On upper bands the feedpoint samples a different standing-wave mode. Resistance and reactance can move through high and low values as frequency changes. The result may look dramatic on a Smith chart, but it is not random. It follows from the mode, feed location and environment.
A tuner can transform that impedance to something the transmitter accepts. It cannot rearrange the loop’s current phase or remove an unwanted far-field null. Matching and pattern are different design problems.
Where the Tuner Sits Determines the Feed-System Loss
A low-loss balanced line feeding the loop into a suitable matching system can make multiband operation practical. A remote tuner at the antenna can also prevent a long coax run from carrying a severe standing wave. By contrast, a shack tuner can make the transmitter see 50 Ω while the coax between tuner and loop still operates at high SWR.
In a lossy line, the forward and reflected waves both contribute conductor and dielectric loss. The higher the line attenuation and mismatch, the larger the additional feedline penalty. The Keysight cable and antenna measurement guide treats insertion loss, mismatch and antenna return loss as separate quantities and warns that line loss can mask the load seen by the analyser.
That gives two very different outcomes that can share the same transmitter SWR:
| System | What the transmitter sees | What can still be wrong |
|---|---|---|
| Shack tuner, long coax, high loop mismatch | Comfortable 50 Ω match | High coax loss, voltage/current stress and unchanged multi-lobe pattern |
| Low-loss balanced line or feedpoint tuner | Comfortable matched source | Pattern may still miss the wanted azimuth or elevation angle |
| Dedicated upper-band radiator | Usually simpler band-specific match | Still requires correct height, ground, feed balance and common-mode control |
Ground Does Not Automatically Become the Dominant Loss
An elevated loop interacts with real ground through its near field and through reflection of the radiated field. Soil conductivity and permittivity can change impedance, efficiency and pattern. That does not justify a universal claim that an upper-band loop “heats the soil” because its current paths became shorter.
Ground loss depends on field strength in the lossy region, loop height, polarisation, geometry and soil properties. As frequency rises, the same structure is electrically higher, which often changes the interference pattern more dramatically than it changes dissipative ground loss. The only defensible conclusion comes from an installed model with realistic ground or from calibrated measurement.
The Lawrence Livermore NEC paper describes the method-of-moments treatment of thin-wire antennas and their interaction with finite-conductivity ground. It supports modelling the actual current distribution and environment rather than assigning loss from wire length alone.
Wire Versus Tubing Is a Loss Budget, Not a Verdict
Conductor diameter and material affect RF resistance, bandwidth, mechanical strength and the exact resonant length. Skin effect raises AC resistance as frequency increases, while a larger circumference can reduce surface-current density. Connections, corrosion and current maxima may matter as much as the bulk conductor.
Thick aluminium tubing can therefore be a good engineering choice, and thin or weathered wire can add avoidable loss. But a properly built copper-wire loop can still be highly efficient on HF. Neither material proves the realised gain of the complete antenna in a wanted direction. That requires the conductor-loss model, matching loss, feedline loss and installed pattern at the same frequency.
When the Oversized Loop Is Still Useful
A lower-band loop can remain a valuable multiband antenna when its upper-band lobes serve the wanted paths and the feed system stays low loss. Its nulls may reject noise or unwanted directions. On receive, absolute efficiency may matter less once external noise dominates, while pattern and common-mode pickup remain important.
It can also be an acceptable general-purpose antenna when convenience matters more than uniform azimuth coverage. The honest description is “useful in these directions and angles,” not “a clean low-angle loop on every harmonic.”
When a Dedicated Upper-Band Radiator Is the Better Tool
If the goal is predictable DX coverage on 20, 17 or 15 metres, a band-specific delta loop, dipole or vertical is usually easier to control. Its electrical size, feedpoint mode and current distribution are chosen for that band. The resulting pattern is simpler to model, the feed transformation can be designed around a known impedance region, and line loss is easier to bound.
The advantage is predictability, not magic. A dedicated radiator can still be installed too low, excite the feedline exterior, couple to nearby metal or place a null on the desired path. It simply starts from a less complicated modal problem.
Bottom line: a lower-band delta loop can disappoint on upper HF even when it radiates efficiently. The electrical perimeter becomes multiwavelength, the pattern divides into more lobes and nulls, the feed impedance changes, and a convenient shack-side match can hide line loss. Choose it only after its installed pattern and complete feed system suit the paths you want.
What to Check Before Calling It Efficient—or Inefficient
- Current magnitude and phase: solve the entire closed conductor, feed transition and any common-mode path.
- Full three-dimensional pattern: inspect the wanted azimuth and elevation region, not only peak gain.
- Reference plane: separate transmitter power, power delivered by the line and power accepted by the antenna.
- Loss budget: include conductor, connectors, balun or choke, tuner, feedline and ground interaction.
- Installed environment: include actual height, soil, terrain, supports, nearby wires and metalwork.
- Restored-baseline comparison: compare against a band-specific radiator with the same receiver, path timing and power boundary.
Engineering references
- Lawrence Livermore National Laboratory — Present Capabilities and New Developments in Antenna Modeling with NEC
- Virginia Tech — Electrical Design and Testing of a One-Wavelength Resonant Loop Antenna
- Rohde & Schwarz — Demystifying OTA Testing: Antenna Parameters
- Keysight — Techniques for Precise Cable and Antenna Measurements in the Field
Mini-FAQ
- Does an 80 metre delta loop become inefficient on 20 metres? Not automatically. It can retain good radiation efficiency while developing a multi-lobe pattern that provides poor realised gain in the wanted direction.
- Does most of the wire stop radiating on the higher bands? No. Current magnitude and phase vary around the multiwavelength loop, and the fields from different segments add and cancel in different directions.
- Can a tuner make the upper-band pattern good? No. A tuner transforms impedance at its reference plane; it does not rearrange the loop's far-field lobes and nulls.
- Is a shack tuner enough for a badly mismatched loop? It can protect the transmitter, but a long coax run may still have high SWR and additional loss between tuner and antenna.
- Do ground losses always dominate on the upper bands? No. Ground interaction depends on height, geometry, polarisation and soil. The pattern often changes more obviously because the antenna is electrically higher.
- When should I use a dedicated upper-band antenna? Use one when you need a more predictable pattern, impedance region and loss budget for a defined azimuth and elevation range.