Designing Multiband Dipoles with Off-Resonance Traps
Designing Multiband Dipoles with Off-Resonance Traps
A parallel-LC trap does not have to resonate on an operating band. Place its resonance between two bands and its inductive and capacitive reactance can help both wire sections carry current—but only the complete antenna decides whether that trade works.
I like the off-resonance trap because it asks a better design question. Instead of trying to make the outer wire disappear on one band, can we choose an impedance that shapes current usefully on both bands? Sometimes we can. The answer comes from current, loss, stress and pattern—not from calling one topology smarter.
The familiar name non-resonant trap needs one clarification: the LC network still has a resonance. It is deliberately placed away from the operating bands, often somewhere between them. Off-operating-band resonant trap is less elegant, but it describes the circuit more precisely.
What a Parallel-LC Trap Does
A conventional dipole trap is commonly a parallel inductor-capacitor network inserted in series with each leg. In the ideal lossless model its two-terminal impedance is:
f0 = 1 / [2π√(LC)]
Ztrap = jωL / (1 − ω²LC)
Below its parallel-resonant frequency, the ideal network is inductive. Above resonance, it is capacitive. At resonance, its impedance tends toward infinity in the ideal equation; a real trap reaches a finite peak because the coil, capacitor, leads, enclosure and surrounding antenna all have loss and parasitics.
A trap resonant on the upper operating band can make the outer conductor carry relatively little current there while the full structure operates on the lower band. The isolation is approximate, not a perfect switch. An off-resonance trap keeps finite reactance on both operating bands, allowing the outer sections to participate on both.
The useful idea: below the trap resonance, use its inductive reactance as part of the lower-band electrical length; above resonance, use its capacitive reactance as part of the upper-band solution. Then adjust the complete wire lengths and network together.
The Frequency Mean Is a Starting Point, Not the Design
For an 80/40-metre dipole, a trap resonance near the gap between the bands can be a practical starting point. A value around 5.2 MHz places an ideal parallel LC below resonance on 80 metres and above resonance on 40 metres. It does not follow that 5.2 MHz, an arithmetic mean or a geometric mean is automatically optimal.
The final frequency depends on the chosen wire lengths, trap position, conductor diameter, height, ground, component Q, feed impedance and the match target on both bands. The design variable is the complete complex trap impedance R(f) + jX(f) at the installed current, not merely the unloaded resonant frequency.
The two dipole legs also need matched networks. Component tolerances, coil spacing, lead length and weatherproofing can make nominally identical traps differ in resonance, resistance and phase. Pair matching reduces one source of asymmetry; it does not guarantee that a real installation is balanced.
Continuous Current Is Not Automatically Better Current
Keeping the outer segments active can avoid the abrupt current change associated with a high-impedance boundary. That can be useful. It does not guarantee equal leg currents, a single smooth lobe, low takeoff angle or higher efficiency.
Radiation follows current magnitude and phase along the entire conductor. On a higher band, an electrically long wire can develop multiple maxima and lobes. Current in an outer segment may reinforce radiation in one direction and cancel it in another. Height, ground, leg angle, feed-line routing and nearby structures also shape the installed pattern.
A balanced dipole requires equal and opposite feed currents in the intended differential mode. Symmetric traps help preserve geometric symmetry, but the trap does not replace a suitable balanced feed arrangement or common-mode verification. Measure current on both feed conductors and the feed-line exterior.
Resonant Traps Do Not Automatically Fall Short
A well-designed conventional trap can make an effective multiband antenna. Its high impedance near resonance can provide a useful current boundary, while its off-resonance reactance loads the longer-band conductor. It need not create an imbalanced pattern when the pair and installation are symmetric.
The trade is stored energy and component stress. A high-Q parallel trap can carry substantial circulating current and develop high voltage even when the conductor current beyond it is small. Coil resistance, skin and proximity effects, capacitor ESR and dielectric loss convert some accepted RF power into heat. A lower-Q trap has a broader, lower impedance peak but may dissipate more power. Neither Q nor bandwidth is an efficiency verdict by itself.
An off-resonance trap moves those stresses; it does not abolish them. Because current continues into the outer segment, the trap may carry meaningful series current on both bands. The coil and capacitor must be checked for RF current, voltage, temperature, dielectric strength, corona clearance and weather-driven detuning at the intended power and duty cycle.
Bandwidth Belongs to the Complete Antenna
Moving the trap resonance away from an operating band can reduce one sharp impedance transition. That may make tuning less sensitive in a particular design. It does not automatically widen the usable antenna bandwidth.
State what bandwidth means: an SWR limit at a declared reference plane, a tuner load region, a gain or efficiency limit, a pattern constraint, or a component-temperature limit. Feed-line attenuation can make the transmitter-end SWR look broad while power is lost. A lossy trap can flatten a curve for the wrong reason.
The antenna may be adjusted so its feedpoint reactance crosses zero on each target band, but resonance is not mandatory for power transfer when a matching network is used. Keep antenna resonance, trap resonance and acceptable system match as three separate statements.
Model the Physical Network You Intend to Build
NEC-class modelling is valuable because it shows current magnitude and phase along the wire and predicts the pattern for a declared geometry and ground model. Represent each trap with its measured complex impedance versus frequency, not an ideal inductance or perfect open circuit unless that is explicitly the comparison.
A simple lumped LC model may be adequate around two HF bands when the trap is electrically small. A transformer-coupled, coaxial or distributed structure can behave differently and needs its own equivalent circuit or measured network data. Include leads, enclosure capacitance and nearby conductors when they are large enough to matter.
Simulation makes it easier to find promising dimensions. It does not prove component loss, power handling, weather stability or a real site pattern. Those return to measurement.
The ON7WP Builds Are Valuable Field Context
Belgian operator Pedro, ON7WP, has built 160/80- and 80/40-metre dipoles around this off-resonance-trap idea. Those installations show that the topology can be built and tuned in the field.
They do not, by themselves, establish a universal bandwidth, efficiency or low-angle-DX advantage. A transferable result needs the antenna dimensions, trap impedance and Q, calibrated feedpoint data, current distribution, site and ground description, accepted power, thermal behaviour and a pattern or controlled field comparison.
Do Not Transplant the Network by Name
The same broad idea can be explored in loops, verticals, end-fed wires and other multiband structures: insert a frequency-dependent impedance to reshape current. The component values and position cannot simply be copied from a centre-fed dipole.
A vertical has a ground or counterpoise system; a loop has a closed current path; an end-fed wire has a high-impedance feed and a required return branch; a fan dipole already contains mutually coupled resonators. Each topology changes the impedance seen on both sides of the network and therefore its current, voltage, loss and pattern.
A Measurement Programme That Can Decide
- Characterise both traps. Measure complex two-terminal impedance across and beyond every target band with fixture open/short/load correction or de-embedding. Record resonance, peak resistance, Q and pair mismatch.
- Freeze the antenna geometry. Document wire lengths, trap positions, height, leg angle, feed line, balun or choke, ground and nearby conductors.
- Save complex feedpoint data. Calibrate at a declared plane and record R, X and complex S11. Do not judge the design from SWR minima alone.
- Measure current. Compare magnitude and phase at marked positions on both legs and measure exterior feed-line current on each band.
- Separate reflection and loss. Use component S-parameters or a justified loss method, then compare accepted power, gain or efficiency for the complete antenna.
- Verify pattern. Compare modelled and measured azimuth/elevation behaviour at the intended installation height and site.
- Run a stated-power test. Record waveform, duty cycle, duration, ambient temperature, trap temperature, impedance drift and post-test inspection. A low-power VNA sweep is not a power rating.
- Repeat wet and dry. Water on the coil form, capacitor, enclosure, support rope or conductor can change capacitance, loss and flashover margin.
Primary and Authoritative Technical Sources
- ARRL, HF Trap Antennas—parallel-tuned traps, between-band trap design and practical component boundaries.
- David Birnbaum, K2LYV, Design of a Two-Band Loaded Dipole Antenna—parallel-LC reactance below, at and above resonance in a complete radiator design.
- IEEE Std 149-2021, Recommended Practice for Antenna Measurements—impedance, pattern, gain, efficiency, site and uncertainty measurement practice.
- Lawrence Livermore National Laboratory, Numerical Electromagnetic Code v5—full-wire, load, transmission-line, ground, current and pattern modelling.
- Keysight, Impedance Measurement Handbook—complex impedance, real-component models, fixtures, calibration and compensation.
- Coilcraft, Testing Inductors at Application Frequencies—frequency-dependent inductance, Q, loss and self-resonance.
- Murata, Capacitor Impedance and ESR Frequency Characteristics—real capacitor ESR, ESL, dielectric loss and self-resonance.
Joeri’s Bottom Line
An off-resonance trap can be an elegant way to make both the inner and outer wire sections work on two bands. That is the narrative worth keeping. The network is not a gentle current roundabout by definition; below its resonance it is inductive, above it capacitive, and its resistance turns power into heat.
Choose the trap resonance and wire lengths together. Match the component pair, model the complete antenna, and verify current, loss, stress, bandwidth and pattern. When those results are good, the topology has earned its place.
Mini-FAQ
- Is a non-resonant trap actually non-resonant? Not literally. Its LC resonance is placed away from the operating bands, often between them, so the trap presents finite reactance rather than peak impedance on those bands.
- What does an ideal parallel-LC trap do below and above resonance? Below resonance it is inductive; above resonance it is capacitive. Real resistance and parasitics limit the impedance and add loss.
- Does an off-resonance trap guarantee that both wire sections radiate efficiently? No. It permits current to continue, but the magnitude, phase, loss and resulting pattern depend on the complete antenna and installation.
- Is the arithmetic or geometric mean always the best trap frequency? No. A mean is only a starting point; wire lengths, trap position, complex impedance, height, ground and targets on both bands determine the optimum.
- Are conventional resonant traps always narrow and lossy? No. Their bandwidth, loss and stress depend on Q, construction, placement, component ratings and the complete antenna.
- How should two traps for a dipole be accepted? Match their complex impedance across the bands, then verify feedpoint impedance, leg and feed-line currents, loss, thermal stress, pattern and wet/dry repeatability in the installed antenna.