The Doublet Is a Feed-System Decision, Not a Magic Length
The Doublet Is a Feed-System Decision, Not a Magic Length
A doublet does not become useful because it matches one famous dimension. Its adaptability comes from the complete system: the wire you can install, the current it supports on each band, the balanced line, the tuner and the pattern your site actually produces.
My starting point for a doublet is simple: use a sensible, symmetric wire span that fits the site, feed it with a low-loss two-conductor line, and make sure the tuner sees loads it can handle. Then inspect the pattern on every band you intend to use. That is smarter than chasing a celebrated length and assuming the rest will take care of itself.
The useful freedom is bounded. A centre-fed wire can operate away from self-resonance, but not every length, feedline length or tuner arrangement is equally easy, efficient, balanced or safe.
Begin With Current, Not the SWR Display
A doublet is a centre-fed conductor system. On the band where its top wire is near a half wavelength, a simple symmetric horizontal installation can support a familiar current maximum near the feedpoint and a broadside pattern. Change frequency, height, V angle, surroundings or return-current symmetry and both the current distribution and pattern change.
The transmitter does not need the wire to be self-resonant. It needs a load that the matching network can transform without unacceptable loss, voltage, current or heating. A low SWR at the transmitter confirms a match at that plane; it does not by itself prove low feedline loss, current balance, radiation efficiency or a useful pattern.
That is why the doublet is not fundamentally a fixed-length recipe. The length is a design variable that helps set the band-by-band current distribution and the impedances delivered to the feed system.
Electrical Length Changes on Every Band
For frequency f and free-space propagation speed c, wavelength is:
λ = c/f
A physical top-wire length L therefore becomes a different value of L/λ on each band. That normalized length is more informative than metres alone, but it is still not a complete pattern prediction. Installed conductor diameter, insulation, end loading, height, V angle, ground properties, nearby metal and feedline current all alter the result.
Near a half wavelength, a reasonably symmetric straight wire often has its strongest radiation broadside. As the wire becomes several half wavelengths long, extra current regions appear and the azimuth and elevation patterns divide into more lobes. Some lobes can occur at useful low elevations; other directions can develop deep minima. “Longer means more DX” is therefore not a dependable rule.
On bands where the wire is electrically short, radiation resistance can fall and tuner current or component loss can become more important. Matching an electrically small system does not remove its loss mechanisms. The lowest practical band must be judged with conductor loss, ground and nearby-object coupling, feedline loss, matching loss and pattern included.
Height Is Also Measured in Wavelengths
The same physical height becomes a different fraction of a wavelength on every band. A 12 m support is low in wavelengths on 80 m and much higher in wavelengths on 10 m. The ground-reflected field combines with the direct field differently at those two electrical heights, changing elevation lobes and nulls.
Low horizontal wires often have strong high-elevation response, which can support regional NVIS work when the ionosphere supports the path. Raising the wire can redistribute energy toward lower elevations, but there is no universal height at which every installation changes from “NVIS” to “DX.” Ground, terrain, polarization, current distribution and the actual ionospheric path all matter.
Model the installed geometry at representative frequencies, then confirm it with controlled on-air or field measurements. A single SWR sweep cannot reveal the pattern.
The Feedline Is an Impedance Transformer
Open-wire and window line are attractive because their loss can be low, including under substantial mismatch. “Low loss” does not mean “no loss,” and a nominal 600 Ω label is not a measurement of every installed line. Conductor size and spacing, insulation, supports, water, nearby objects and construction tolerance affect characteristic impedance, velocity factor and attenuation.
For a lossless line with characteristic impedance Z0, load ZL, phase constant β and length l, the input impedance is:
Zin = Z0[ZL + jZ0tan(βl)]/[Z0 + jZLtan(βl)]
The equation explains why feedline length matters even when line loss is small. The line transforms the complex antenna-terminal impedance into a different complex load at the tuner. A line length that is comfortable on one band can present high voltage, high current or a load outside the tuner's range on another.
Do not cut line merely to obtain a pretty SWR without recording the measurement plane. Use a calibrated analyzer, measure resistance and reactance across every intended band, and treat the final routed line as part of the design. Bends, proximity to metal and unequal coupling can alter both differential and common-mode behaviour.
Balance Is a Property of the Installed System
A symmetric wire and a two-conductor line encourage equal-and-opposite differential currents. They do not guarantee them. Unequal leg geometry, sloping terrain, a conductive support, asymmetrical line routing, tuner capacitance and attached cables can all provide a third current path.
A balanced-output tuner can be a good interface when its actual load range, loss, voltage and current limits suit the transformed loads. An unbalanced tuner followed by a suitable current balun can also work. The balun must be characterized at the complex impedances and frequencies it will encounter; a catalogue impedance ratio or common-mode dB label does not establish installed loss or heating.
Keep matching and common-mode control as separate questions. The tuner transforms the differential load. The current balun or choke opposes unwanted common-mode current. Depending on topology and placement, one component may influence both, but neither job should be assumed from its name.
Choose Lengths by Load Domain and Pattern
Examples such as 2 × 20 m, 2 × 13 m or 2 × 10 m can be useful starting geometries. They are not guaranteed band packages. The better selection process is:
- Define the bands and directions: decide which operating windows, arrival angles and azimuths matter.
- Fit a mechanically credible span: include support load, wire tension, end clearance, access, wind, ice and nearby services.
- Model current and pattern: use the installed height, V angle, ground and nearby conductors rather than a free-space straight wire alone.
- Calculate or measure tuner-plane loads: include the real feedline length, characteristic impedance, velocity factor and loss.
- Avoid extreme operating points: revise wire or line length when a required band repeatedly drives the tuner, balun or line toward excessive voltage, current, loss or heating.
- Verify balance and accepted power: measure current on both line conductors and the net enclosed current where practical.
The longest wire that fits is not automatically the best multiband answer. Extra electrical length can provide useful lobes, but it can also place nulls in wanted directions or create difficult loads. Likewise, a compact wire may be convenient but inefficient on a band where it is too electrically short. Choose the span that gives the best complete-system trade.
Compare Feedlines at the Same Conditions
Open-wire line, window line, twin-lead and coaxial cable should not be ranked from construction names alone. Their attenuation depends on frequency, conductor and dielectric loss, environment, length and mismatch. The same standing-wave ratio can produce different loss in cables with different matched-line attenuation.
Coax can be perfectly reasonable when the feedpoint impedance remains within a manageable range and common-mode current is controlled. Open-wire line often retains an advantage when the antenna presents a large mismatch over many bands, but routing and tuner interface become more demanding. Wet or contaminated line, close metal and tight bends can erode the expected advantage.
For a fair comparison, define the transmitter-side reference plane, measure accepted power, estimate or measure line and matching loss, monitor temperature and compare radiated field or received signal using the same geometry and time-controlled A/B/A procedure.
Commission the Whole Doublet
A practical commissioning record should include:
- top-wire dimensions, height, shape, orientation, conductor and insulation;
- feedline construction, routed length, spacing from metal and weather condition;
- complex impedance sweeps at the antenna, line input and transmitter where those planes can be measured;
- tuner settings, component voltage/current limits, loss or temperature and duty cycle;
- paired conductor currents and net common-mode current;
- band-specific modelled or measured patterns and their useful/null directions;
- RF-exposure assessment, inaccessible high-voltage regions and electrical/overhead-line clearances.
Retest after rain, routing changes and seasonal ground changes. A doublet is adaptable precisely because the feed system can transform many loads; that same adaptability makes the final installation worth measuring.
Engineering references
- IEEE 145-2025 — Standard for Definitions of Terms for Antennas
- IEEE 149-2021 — Recommended Practice for Antenna Measurements
- Lawrence Livermore National Laboratory — Numerical Electromagnetic Code, NEC v5.0
- Keysight — Impedance Measurement Handbook
- Roy Lewallen, W7EL — Baluns: What They Do and How They Do It
- ICNIRP 2020 — Radiofrequency Exposure Guidelines
The doublet's strength is not a magic number. It is the ability to choose a useful wire, transform its changing loads with a low-loss balanced path, and verify the current and pattern on every band that matters.
Mini-FAQ
- Can a doublet be almost any length? It can operate away from self-resonance, but wire and feedline lengths must still keep patterns, loss, voltage, current and tuner load acceptable on the intended bands.
- Does a tuner make the antenna resonant? No. It transforms the impedance presented at its output into a load the transmitter can accept; it does not erase feedline loss or change the antenna's current distribution into resonance.
- Is 600 Ω open-wire line always the best choice? No. It can be very low loss, especially under mismatch, but its actual construction, routing, environment, length and tuner interface determine the result.
- Does a symmetric doublet guarantee balance? No. Unequal surroundings, feedline routing, supports, tuner capacitance and attached cables can create common-mode current.
- Will a longer doublet always improve DX? No. More electrical length creates additional lobes and nulls. Some may help a wanted path and others may remove it.
- What should I measure before transmitting at full duty cycle? Measure complex loads at declared planes, current balance, accepted power, line and matching loss or temperature, and confirm RF-exposure and electrical clearances.