EFHW, EF-OCF or End-Fed Long Wire: Which Fits Your Station?

EFHW, EF-OCF and end-fed long wire are often treated as three versions of the same idea. They are better understood as three different ways to choose a standing-wave mode, feed it from one end of the installation and complete the RF return path.

That distinction matters because a low SWR at the radio does not tell us where the current flows, how much power the matching system dissipates, or where the higher-band lobes point. I choose among these antennas by asking what wire fits, which bands matter, what load reaches the feed unit, where the return current should flow, and which measurements are practical at the site.

The Names Describe Different Current Distributions

“End-fed” describes the physical access point. It does not mean that current leaves one transformer terminal and disappears at the far end of the wire. Every one of these systems needs a return path through another conductor, the outside of the coax, capacitance to the environment, or a combination of them.

The more useful definitions are therefore:

• EFHW-like system: the wire is operated near an integer half-wave mode on an intended band. Feeding near a current minimum usually creates a high impedance and a high-voltage region, but the actual complex value depends on the complete installation.
• One-end EF-OCF system: two unequal current-carrying branches meet at an accessible end of the installation. One branch may be the long wire and the other a deliberately defined section of coax exterior that ends at a choke.
• Non-resonant end-fed wire system: the wire length is chosen to produce usable—not magically constant—loads across a set of bands, with a transformation network and tuner completing the matching job.

A conventional two-wire off-centre-fed dipole is different from the one-end EF-OCF discussed here. The conventional OCF feedpoint sits between two physical wire legs. In the one-end arrangement, the second branch can be the outside surface of a declared coax section. That section is part of the antenna until the choke establishes its far boundary.

The EFHW Uses Half-Wave and Higher-Order Modes

An EFHW can be an elegant answer when its target bands and installation suit those modes. The lowest intended band may place a current minimum near the feed end, while higher bands can use higher-order standing-wave modes on the same conductor.

The feedpoint impedance is not one fixed resistor. It changes with wire diameter, electrical length, height, slope, bends, nearby objects, soil coupling and the chosen return path. A nominal transformer ratio is therefore a design starting point, not a statement of the actual load on every band.

Higher-order modes also place additional current maxima and minima along the wire. That produces more azimuth and elevation lobes and deeper nulls. The antenna can still be useful on those bands, but “multiband” does not mean that the fundamental-band pattern, feed impedance or transformer stress repeats unchanged.

A compensation capacitor can be a legitimate part of the matching network. It may counter high-frequency transformer reactance or move a desired SWR minimum. The changed SWR trace describes the revised network; it does not prove efficiency by itself. Its value, loss, voltage/current stress and effect on the complete network must be checked.

The One-End EF-OCF Defines Unequal Branches

Moving away from an extreme current minimum can produce an installed impedance that needs a more moderate transformation. That is the attraction of the EF-OCF architecture. It is not enough to declare a percentage split or assume a resistance from a drawing; the impedance remains complex and installation-dependent.

In the version Mark discusses, the long wire forms one branch and a selected section of coax exterior forms the other. The UNUN handles the impedance transformation. A separate choke raises common-mode impedance at the chosen boundary and prevents the rest of the feedline from freely extending that branch.

Transformation and common-mode suppression are separate functions. A nominal 4:1 UNUN can be a sensible starting hypothesis when measurements support it, while the choke is chosen and placed from the installed exterior-current path. Fewer turns or a lower ratio do not automatically prove lower loss, lower voltage stress or greater efficiency.

The coax manufacturer’s internal velocity factor does not set the length of the exterior branch. The wanted differential signal travels between the centre conductor and the inside of the shield. The common-mode current on the shield exterior sees the routing, jacket, supports, soil and nearby conductors. Measure that installed path instead of cutting it from an internal-coax data-sheet value.

The End-Fed Long Wire Trades Resonant Convenience for Tuner Range

A non-resonant end-fed wire is valuable when the site rewards flexibility: a slope, inverted L, bend or temporary support may be easier than a precisely placed resonant wire. The length should still be chosen deliberately. “Random wire” is shorthand, not an instruction to use any length.

The objective is to avoid loads that are excessively difficult for the UNUN, feedline or tuner on the intended bands. A nominal 9:1 transformation is common, but it does not guarantee a comfortable impedance. The tuner may encounter very high or very low resistance, large reactance, high voltage or high current after the feedline has transformed the load.

A successful tuner match proves that the radio-facing port was transformed to an acceptable impedance. It does not quantify transformer loss, tuner loss, feedline attenuation, ground or return-path loss, common-mode current, radiation efficiency or realized gain.

This antenna family is not inherently noisier. Received noise depends on the antenna pattern, polarization, local electric and magnetic fields, feedline common mode, receiver bandwidth and the noise sources around the site. A different wire route can alter all of those at once.

All Three Antennas Need a Declared Return Path

The transformer has two RF terminals. If the installation does not supply a deliberate second conductor, current finds distributed capacitance, the coax exterior, a mast, station bonding and nearby structures. That may make contacts, but it leaves the antenna boundary dependent on cable routing and connected equipment.

My preferred workflow is to decide the return path before choosing choke placement:

• If radials or a counterpoise complete the intended return, place the choke where it stops the feedline exterior from becoming another uncontrolled branch.
• If a selected coax-exterior section is an intentional EF-OCF branch, place the choke at the far end of that section.
• If a second station boundary still carries current, an additional choke may be useful there, but it does not replace the antenna-side current definition.
• Verify the result by measuring current around the complete coax at several marked positions on every intended band.

A fixed fractional-wavelength choke distance cannot do this job universally. Moving the cable, mast, ground connection or choke changes the external conductor geometry and can change both the feed impedance and the far field.

Transformer Ratio Is Not a Performance Ranking

The square of the turns ratio gives the ideal impedance transformation. A real broadband transformer also has magnetising inductance, leakage inductance, winding capacitance, conductor loss, core loss and temperature dependence. The antenna supplies a frequency-dependent complex load rather than the resistor printed beside a schematic.

That is why a lower nominal ratio cannot, by itself, prove lower loss or higher power handling. A fair comparison identifies:

• topology, turns ratio, winding and exact magnetic material;
• frequency and complex source/load impedances at declared reference planes;
• power, waveform, duty cycle, test duration and ambient temperature;
• common-mode impedance and the actual return conductor;
• insertion loss or accepted-power balance under representative loads; and
• temperature rise and voltage/current margin across the intended bands.

Feedline Loss and Pattern Belong in the Decision

At high SWR, voltage and current vary along the line and conductor/dielectric loss can rise. A source-end SWR may even look better after a longer lossy cable because the reflected wave is attenuated twice. Record insertion loss and move the calibrated reference plane to the antenna system when possible.

Pattern is just as important as accepted power. A longer wire used on higher modes may create useful lobes in some directions and deep nulls in others. Height above real ground changes the elevation pattern; bends, slope and nearby structures alter the current distribution. No acronym guarantees DX, local coverage or a low take-off angle.

When a particular path matters, compare realized field in that direction at equal accepted power. Keep the site, feedline route, receiver linearity and propagation window controlled, and repeat the sequence. One strong contact or one quiet moment is a useful observation, not a complete pattern or efficiency measurement.

A Practical Comparison

For a fixed station, start from the bands and directions that matter, then decide whether you can define the second branch and reach the feedpoint for measurement. For portable work, support options, setup time and tuner range may dominate. For high duty cycle or higher power, matching-network voltage, current and thermal evidence become decisive.

Measure the Installed Antenna, Not the Label

1. Draw every conductor. Include the radiator, counterpoise or exterior-coax branch, feedline, choke locations, mast, bonding and tuner.
2. Measure complex impedance. State the reference plane and record R + jX across every intended band, not only the minimum SWR.
3. Characterise the feedline. Record type, physical length, insertion loss and the manufacturer’s internal velocity factor.
4. Map common-mode current. Clamp around the complete coax at several repeatable positions before and after moving a choke or rerouting the line.
5. Check matching-system stress. Use representative power and duty cycle while monitoring transformer, tuner, choke and connector temperature safely.
6. Evaluate the pattern. Model the installed geometry over real ground and confirm the directions that matter with controlled field comparisons.
7. Repeat one variable at a time. An A/B/A sequence helps separate the antenna change from propagation and local-noise variation.

Primary Engineering References

• Roy Lewallen, W7EL, “Baluns: What They Do and How They Do It”—feedline imbalance, outside-shield current and current-balun behaviour.
• ARRL, Antenna Design and Construction—current distribution, end and off-centre feeding, transmission-line transformation and pattern.
• Lawrence Livermore National Laboratory, Numerical Electromagnetics Code—Method of Moments—wire-current and radiation-pattern modelling for arbitrary geometries and environments.
• Keysight, Techniques for Precise Cable and Antenna Measurements in the Field—calibration planes, insertion loss, return loss and VSWR.
• NIST, “A Two-Port Model for Antennas in an Arbitrary Environment”—separating antenna loss, imperfections and radiation-efficiency evidence.
• IEEE 145, Standard for Definitions of Terms for Antennas—consistent terminology for impedance, efficiency, gain and pattern.

My Bottom Line

I do not buy an antenna from its acronym. I buy the current path I can define and verify.

An EFHW can be the cleanest fit for selected half-wave and higher-mode bands. A one-end EF-OCF can be the better multiband starting point when a moderate installed load and a deliberate coax-exterior branch suit the site. A non-resonant end-fed wire can be the practical answer when geometry is constrained and a tuner is part of the plan.

Whichever one you choose, keep the impedance transformation and common-mode boundary as separate engineering questions. Then measure the complete installed system—because that is the antenna that goes on the air.

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