Monoband EFHW L-Matching on 17 m, 15 m, 12 m and 10 m

My design rule is simple: match the antenna that exists in the air. On 17, 15, 12 or 10 metres, a compact monoband L-network can avoid some broadband-transformer compromises. It is not automatically lower-loss, wider-band or safer. Those results depend on the installed load, the chosen network branch, component Q, layout, return current and operating conditions.

Measure the Installed Complex Load

An end-fed half-wave has a current minimum and voltage maximum near its open end, so its feed resistance is normally high. The actual input is ZL = RL + jXL, not a fixed resistor. Wire diameter and insulation, exact electrical length, bends, end hardware, height, ground, nearby objects, enclosure capacitance and the return structure all move it.

The ARRL EFHW kit guide, for example, uses about 2.5 kΩ as a test load for one particular transformer. That is useful evidence for that implementation, not a universal end impedance. The value for this design must come from the installed antenna or from a complete model that is then checked on the installed antenna.

Define the measurement plane before trusting the number. Calibrate the VNA as close as practical to the antenna-side terminals of the matcher, or de-embed the known feed fixture. The Keysight calibration guidance explains that calibration establishes a fixed reference plane; moving that plane through coax transforms the displayed impedance. At a high-impedance node, connector, lead and enclosure capacitance can be large compared with the component being measured, so keep the fixture stable and check repeatability.

Freeze the Current-Return Geometry

The matching network needs two RF terminals. At the high side, one terminal feeds the radiator. The other is the return node. Return current may use a deliberate counterpoise, a controlled section of coax exterior, a conductive support or distributed capacitance to the surroundings. It cannot be left undefined without making the coax route and station wiring part of the experiment.

A counterpoise is not automatically 0.05λ, and a choke is not automatically correct at one prescribed distance. Both alter the installed impedance and current distribution. Set the intended geometry first, then map exterior-coax current and place or adjust the choke to create the boundary the design requires. ITU-T K.136 defines converted common-mode current as asymmetrical current produced from differential current by cable or network imbalance; that is the mode to control on the coax exterior.

Choose the L-Network Branch From the Load

For a purely resistive load above 50 Ω, one valid low-pass branch places a series inductor on the 50 Ω side and a shunt capacitor across the high-resistance load:

50 Ω side ── series L ──●── radiator
                        |
                      shunt C
                        |
return node ────────────●

The shunt element belongs on the higher-resistance side. A high-pass branch using a series capacitor and shunt inductor is also possible. Once the antenna has appreciable reactance, the existing XL can be absorbed into the network, and the required signs and values can change. Use the full complex load rather than forcing every installation into the drawing above. The Analog Devices complex-load matching calculator returns alternative networks for this reason.

The Real-Load Equations Are a Check, Not a Parts List

For the limited ideal case RL > RS, with both terminations purely resistive and lossless parts, the low-pass branch above uses:

The derivation follows the series/parallel resistance conversion in the Analog Devices L-network treatment. As an arithmetic check only, a hypothetical 50 Ω-to-3.2 kΩ resistive match at 21.2 MHz gives Qmatch = 7.94, L = 2.98 µH and C = 18.6 pF. Those values do not belong to a real 15 m antenna until that antenna has measured 3.2 kΩ with negligible reactance at the declared plane.

For a measured R + jX load, solve the complex network directly or use a Smith chart or simulator, then verify the result. Do not cancel the antenna reactance in one unstated step and apply the real-load equations as though nothing changed.

Matching Q and Component Q Are Different

Qmatch above is set by the resistance transformation in this two-element ideal network. It is not the Q of the inductor or capacitor. A large impedance ratio forces a relatively high matching Q and usually a narrow match, but the useful bandwidth must be measured against a declared reflection or SWR limit. Antenna Q, component loss and load variation also shape the curve. Loss can make an SWR curve look wider while converting more power into heat.

Component Q is approximately |X|/Rloss under the relevant equivalent model. Measure it at the operating frequency and with representative mounting geometry. The Keysight impedance-measurement handbook shows how winding resistance, stray capacitance and self-resonance change an inductor across frequency. An air-core coil avoids ferrite loss, but it does not remove conductor loss, proximity effect, lead inductance, self-capacitance or enclosure coupling.

Select the capacitor by measured capacitance, RF current, ESR, voltage rating, temperature behavior and self-resonance—not nominal picofarads alone. Keep connections short, mechanically stable and separated from the enclosure. After assembly, remeasure because the box is part of the circuit.

Calculate Voltage, Current and Heat

For a purely resistive load accepting power P, the load-terminal values are VRMS = √(PRL) and IRMS = √(P/RL). In the low-pass branch, the shunt capacitor sees the high-node voltage and carries approximately V/|Xshunt|; the series inductor carries the source-side network current. Reactive mismatch and stored energy can create local stress beyond a simple load-terminal calculation.

There is therefore no universal safe transmitter power for an L/C pair. Calculate the worst intended load and mismatch, waveform crest, duty cycle, average heating, enclosure temperature and component tolerances. Apply a justified voltage, current and thermal margin, then increase power in stages while monitoring temperature and match drift. Stop on unexpected heating, arcing, odor or movement of the match.

An Open Coax Tail Is a Stub

A short open-ended coax section can supply shunt capacitive susceptance, but it is a distributed open stub rather than a frequency-independent capacitor. For an ideal lossless open stub of characteristic impedance Z0 and electrical length βℓ:

Only when the stub is electrically short does it resemble a lumped capacitor over a small frequency range. Velocity factor, loss, open-end fringing, bend radius, nearby metal and the final box all matter. Trim and characterize the stub in its installed position at low power, keep the open end insulated and inaccessible, and recalculate it if the operating frequency changes. Analog Devices' CN0507 transmission-line treatment shows why open standards and their line delay need a frequency-dependent model.

Tune in an Order That Preserves Meaning

1. Choose the exact operating segment. Design around the intended frequencies and the operator's current authorization, not the band label alone.
2. Define the complete geometry. Record radiator wire and insulation, height, orientation, bends, supports, matcher enclosure, return conductor, choke and coax route.
3. Start the radiator adjustable. Use c/(2f) only as the free-space half-wave reference; a real installed wire is affected by end geometry, insulation and surroundings.
4. Establish the reference plane. Calibrate or de-embed to the declared matcher/antenna interface and keep the fixture unchanged.
5. Place the intended current mode near the target. Adjust radiator geometry with the final return structure present, then measure R + jX.
6. Select and calculate the network. Choose the low-pass or high-pass branch from the measured complex load, required harmonic behavior, available parts and stress limits.
7. Prototype with adjustment range. Use characterized components and tune at low power while watching both resistance and reactance at the declared plane.
8. Iterate deliberately. Adjust L and C for the match. If the antenna reactance or required values are impractical, change the radiator or return geometry, remeasure the load and recalculate; the wire and network are coupled.
9. Lock the construction and verify again. Close the enclosure, route the coax as installed, and repeat the sweep before applying normal power.

Verify More Than SWR

A finished monoband matcher needs a record for each of 17, 15, 12 and 10 metres on which it will be used. Do not copy one band's parts into another box by frequency scaling unless the measured normalized load and physical parasitics support that step.

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