Choosing a Doublet Feedline Length by Measurement

A reader emailed me asking for the “most optimal” feedline length for a doublet using ladder line or window line. It is a good practical question, but there is no universal answer. Antenna geometry, frequency, line impedance, propagation constant, routing and tuner topology make one installed system different from another. The answer is a reproducible measurement record, not a magic length.

The Feedline Is an Impedance Transformer

The impedance at the tuner end is generally not the impedance at the doublet feed point. For a uniform line of length l, characteristic impedance Z₀, propagation constant γ = α + jβ and load Z_L, the input impedance is:

The familiar half-wave repetition and quarter-wave transformation are special lossless cases, not length recommendations. On a multiband doublet, Z_L, βl, attenuation and the line’s surroundings all change with frequency. One physical line can be near a benign transformation on one band and present an extreme resistance or reactance on another.

That is why “90 feet,” “27 metres” or an odd multiple of one-eighth wavelength can only seed a trial list. None is a universal optimum. A candidate survives only when the complete band-by-band result fits the actual tuner, line, balun and operating envelope with margin.

Declare the Reference Plane Before Reading the Display

An analyzer reports reflection and derives R + jX at its calibrated reference plane and reference impedance. Anything between that plane and the balanced-line terminals—coax jumper, adapters, 1:1 balun, choke, binding posts and leads—is part of the measured network unless it has been removed by a valid calibration, port extension or de-embedding model.

Record the analyzer model, firmware, calibration method and standards, reference impedance, sweep power, frequency points, jumper, adapters, fixture, line routing and environmental state. Changing a display from 50 Ω to 450 Ω may renormalize SWR or Smith-chart presentation; it does not physically move the reference plane or remove the fixture.

A 1:1 Current Balun Is a Fixture, Not a Perfect Window

A practical one-port analyzer can be connected to balanced line through a 1:1 current-balun fixture, but the result is the input impedance of the fixture plus line and antenna. Accuracy depends on the fixture’s differential insertion behaviour, transformation ratio, loss, phase balance, common-mode impedance, parasitic capacitance, lead geometry and behaviour over the measured impedance range.

High common-mode impedance helps keep the measurement from recruiting the analyzer, operator or coax exterior as another antenna conductor. It does not prove that the differential path is transparent. A choke designed only from a scalar impedance curve can add enough differential error or resonance to change the apparent load, especially near extreme impedances.

Check the fixture first with known balanced loads spanning the intended frequency and impedance range. Keep leads short, symmetric and repeatable. If the operating station will use that exact balun at that exact plane, retaining it in the measurement can be the most relevant tuner-input test. If the tuner has a different internal balanced network, use an appropriate fixture and account for the difference. A calibrated multiport VNA with mixed-mode analysis is the stronger method when differential-to-common-mode conversion itself must be separated.

Electrical Length Starts With Measured Propagation

Electrical length is phase, not a tape-measure number:

Published velocity factor is a starting value. Conductor spacing, insulation, construction, frequency, nearby material, moisture and installation geometry can change propagation and loss. When length choice is sensitive, measure a known sample or the installed line using the analyzer’s supported time-domain, phase-delay or open/short method, with the correct reference plane and termination.

For time-domain reflectometry, a round-trip delay Δt from a known physical length L gives the first-order relation vp = 2L/Δt and VF = vp/c. End effects, fixture delay and the instrument’s spatial resolution belong in the uncertainty budget.

This lossless phase example is not a recommended length. It shows why a rule derived on the lowest band cannot predict the other bands. The antenna terminal impedance must also be known or measured at each frequency before the transformed input impedance can be calculated.

“Tuner-Friendly” Means Inside Documented Limits

Do not reduce tuner suitability to SWR. Obtain the tuner manufacturer’s matching range versus frequency, power and duty cycle, plus the voltage, current, capacitor, inductor, switch, connector and balun limits that apply to the chosen configuration. Published limits vary substantially among tuners and can narrow at band edges or higher power. If the required boundary is not documented, a successful low-power match is not evidence of safe full-power operation.

For a measured terminal impedance Z = R + jX carrying accepted sinusoidal real power P, a first-order terminal estimate is:

These equations describe that declared plane. They do not reveal the larger circulating current or voltage that may occur inside a particular matching network, across a balun winding, or at a standing-wave maximum along the line. Low resistance tends to demand current; large impedance magnitude tends to demand voltage; large reactance can require substantial stored energy in tuner components. Network topology decides the internal peaks and loss.

A low-power analyzer sweep cannot qualify a line, tuner or balun for transmit power. Ferrite permeability, conductor loss, contact resistance, dielectric loss, arcing distance and temperature can change under power. The selected length needs margin, and the finished installation must follow the tuner and station manufacturers’ commissioning and protection procedures.

Build a Reproducible Length Comparison

1. Freeze the antenna geometry. Install the doublet at its intended height and shape. Route the balanced line with its final spacing from metal, earth, wiring and wet surfaces.
2. Define the required operating set. List exact frequencies, modes, expected power and duty cycle—not only amateur-band names or centres.
3. Record tuner limits. Use the manual for the exact tuner and connection mode. Note frequency-dependent matching, power, duty-cycle and component limits.
4. Make the system safe. Disconnect and inhibit transmitters and amplifiers, avoid storms and nearby RF, discharge the line by the instrument-approved method and connect the analyzer only after the system is de-energized.
5. Establish the reference plane. Calibrate at the accessible connector, document the fixture and decide whether it is intentionally included or validly de-embedded.
6. Validate the balanced fixture. Check known loads and common-mode isolation over the bands and impedance range of interest.
7. Measure the installed candidate. Save R, X, |Z|, reflection coefficient or SWR and uncertainty at every required frequency. Repeat enough points across each band to catch rapid transitions.
8. Model voltage and current. Use a lossy-line model and the actual tuner topology where possible. Check terminal stress, line maxima, tuner internal limits, balun loss and expected thermal conditions.
9. Compare practical alternatives. Simulate several mechanically feasible lengths, then verify promising candidates physically with the same routing and fixture. Start long and use known removable sections; do not coil or fold surplus line in a way that changes coupling and then call it equivalent.
10. Select margin, not the prettiest trace. Choose the length that stays inside documented limits across the complete operating set and remains repeatable after weather, routing and connection checks.

What a Good Record Looks Like

For each candidate length, keep one row per frequency with physical length, measured or declared velocity factor, electrical length, R, X, |Z|, reference plane, fixture identity, analyzer uncertainty, tuner configuration, power/duty assumption and estimated voltage/current margin. Add notes for routing, weather and any nearby-conductor change.

SWR can remain in the table, but it is not the selection metric by itself. Two loads with the same SWR can have different resistance, reactance, voltage/current distribution and tuner loss. The winning length is the defensible multiband compromise for the named station—not the one that produces a photogenic dip.

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