The EFLW37: A Practical 37-Metre End-Fed Wire

I call the EFLW37 quick and dirty with affection. Thirty-seven metres is long enough to be interesting on 80 m yet short enough to fit installations that cannot carry a full-size low-band array. It can be a very useful antenna. What it cannot be is reduced to “37 m plus a 9:1 box equals easy local 80 m.”

Thirty-Seven Metres Is a Case, Not a Sweet Spot

A wavelength in free space is approximately 300 / f metres when frequency f is in megahertz. A 37 m conductor is therefore a substantial fraction of a wavelength across the 80 m amateur band and can approach a half-wave current mode toward part of that range. That observation does not fix its feedpoint impedance.

Insulation, conductor diameter, height, bends, slope, end loading, nearby trees and buildings, ground and the return conductor all change the installed electrical length. At an end feed, moving toward a half-wave mode can produce a high resistance and considerable voltage; away from that mode, resistance and reactance can change rapidly. The same nominal wire can present an ordinary tuner load at one site and a difficult load at another.

This is why published non-resonant-wire lengths are search seeds, not electrical constants. Thirty-seven metres may fit the garden and avoid an extreme impedance on several chosen bands in one installation. It does not earn the word optimum until the complete geometry has been measured against a stated goal.

The 9:1 UNUN and Tuner Have Separate Jobs

An ideal 9:1 impedance transformer refers a load impedance by a factor of nine. It transforms both resistance and reactance; it does not cancel the reactance. A real UNUN also has magnetising impedance, leakage inductance, winding capacitance, conductor and core loss, voltage and current limits, and frequency-dependent behaviour under a complex load.

The tuner then transforms the impedance at its own reference plane to the value wanted by the transmitter. A completed tune cycle tells us that the radio-side match is acceptable. It does not reveal transformer loss, tuner loss, feedline dissipation, common-mode current, radiation efficiency or the direction of the radiated field.

A fixed 9:1 choice is therefore a starting hypothesis. Sweep resistance and reactance at the antenna-side plane across every intended band, then verify that the selected transformer and tuner can handle those loads with acceptable loss, temperature and voltage/current stress.

An End-Fed Wire Cannot Work Without a Return Path

Current leaving one UNUN terminal must return to the other. The return may include a deliberate counterpoise wire, an elevated branch, the outside of the coax, capacitance to nearby objects, soil, bonding conductors and station wiring. If the installation drawing shows only one wire, the missing branch still exists electromagnetically.

I prefer to make that branch deliberate. A defined length of coax exterior can be part of the antenna between the transformer and a separately specified common-mode choke. The choke marks the intended end of that branch only when it provides suitable impedance at the measured current and frequencies. There is no universal quarter-wave counterpoise or fixed choke distance that survives every combination of feedline, ground and surroundings.

Moving the choke changes the antenna. So can rerouting the coax, connecting station equipment or adding a ground lead. Measure exterior current on accessible conductors band by band and recheck impedance after the routing is final. Protective bonding and lightning requirements remain safety functions; do not create an isolated electrode or defeat required bonding to improve an RF reading.

Matching Loss Must Be Counted at Declared Planes

Mismatch does not consume power in an ideal lossless line. A real line has conductor and dielectric loss, and its standing-wave voltage and current distribution changes how much power is dissipated. The added loss depends on the cable type, length, frequency, load and matched attenuation.

If the tuner is in the shack, a low SWR between radio and tuner does not remove the standing wave on the cable between tuner and UNUN. A tuner near the feedpoint can reduce that particular feedline penalty, but its own losses and the UNUN and return-network losses remain. Neither location is automatically correct; compare the two arrangements at the same accepted-power boundary.

A useful efficiency budget includes tuner, feedline, UNUN, connector and conductor loss, plus loss in soil or other return paths. Measure temperature after representative duty rather than after a brief tune carrier. High resistance can raise voltage, low resistance can raise current, and a large reactive component can raise circulating energy without producing a dramatic transmitter-side SWR once the tuner has done its work.

Low Height Does Not Guarantee “Local Only”

A low horizontal or gently sloping wire often produces substantial high-angle radiation on 80 m because its height in wavelengths and the ground-reflected field reshape the pattern. That can support regional near-vertical-incidence skywave when the ionosphere returns the operating frequency with enough link margin.

But “low” is not a complete antenna model. Height, ground conductivity and permittivity, slope, azimuth, bends, nearby conductors and feedline-exterior current all affect the three-dimensional pattern. A sloping end-fed wire with a vertical feedline section may have material vertical as well as horizontal current. Ground and conductor loss do not automatically dominate merely because the wire is close to the ground.

Nor does strong high-angle radiation guarantee NVIS. The path still depends on ionospheric critical frequency, absorption, time, season, location, bandwidth, noise and required reliability. Very short contacts may involve ground wave; regional contacts may involve ionospheric return; neither can be identified from distance alone.

The same installation may also radiate useful lower-angle energy and work DX when the pattern, azimuth and propagation align. Calling it a cloud burner or a poor-DX antenna hides the evidence we actually need: realised gain by elevation and azimuth, accepted power at a declared plane and controlled observations over the paths of interest.

Higher Bands Bring More Lobes, Not Automatic Inefficiency

As frequency rises, 37 m becomes several wavelengths long. The current distribution then develops more maxima and minima, and the far field usually divides into more lobes and nulls. Bends, height changes and return-current geometry can rotate, split or fill those lobes.

That does not make the wire inherently inefficient on 15 m or 10 m. It can radiate strongly in some directions and poorly in others. Its feedpoint impedance can be easy for the chosen network on one higher band and awkward on the next. Transformer and tuner loss can also change sharply with frequency and load.

For multiband use, “the tuner found it” is only the first gate. Map the installed impedance, current, loss and three-dimensional pattern on every band that matters. A multiband wire can be efficient yet place a null toward the desired station, or less efficient yet deliver more field on one path because a lobe points the right way.

Turn the Quick Antenna Into a Measured System

• Record the geometry: wire coordinates, height, slope, bends, conductor and insulation, intended return branch, feedline route, choke, ground and nearby structures.
• Choose the operating jobs: bands, target paths and bearings, local or regional coverage, DX interests, mode, bandwidth, power and duty cycle.
• Sweep at useful planes: measure complex impedance at the UNUN antenna port and at the tuner plane, not only SWR at the transmitter.
• Characterise the network: test transformer and tuner loss, voltage, current and temperature with representative complex loads on each band.
• Map the return current: scan coax exterior and accessible conductors before and after the intended choke boundary.
• Model the complete structure: include wire, return branch, feedline exterior, ground and nearby conductors when evaluating elevation and azimuth pattern.
• Compare fairly: use the same accepted power, rapid A/B/B/A switching or restored baselines, several receiving sites and contemporaneous propagation data.

Primary and Authoritative Sources

• NIST, A Two-Port Model for Antennas in an Arbitrary Environment—a measured network framework separating antenna efficiency, loss and environmental influence.
• Roy W. Lewallen, W7EL, Baluns: What They Do and How They Do It—original analysis and experiments on imbalance current and feedline participation.
• ARRL Laboratory, Tuner Matching and Loss Measurements—load- and frequency-specific bench evidence showing that a completed match and tuner loss are separate results.
• Lawrence Livermore National Laboratory, Modelling with the Numerical Electromagnetics Code—method-of-moments modelling of wire antennas and their environment.
• ITU-R P.341-7, The Concept of Transmission Loss for Radio Links—standard definitions separating antenna, feeder, mismatch, polarization and propagation contributions.
• ITU-R P.533-14, Method for Predicting HF-Circuit Performance—the in-force method covering HF frequency availability, field strength, SNR and circuit reliability.

Joeri’s Bottom Line

The EFLW37 earns its place by being practical. If 37 m is what the site can carry, put it in the air, give it a deliberate return path and measure it. A 9:1 UNUN and tuner may make that installation very usable, but their success is not a certificate for low loss or a promise that the pattern will stay local.

I would rather call it a flexible compromise than a local-only antenna. On 80 m it may favour high angles, lower angles or both. On the higher bands it may produce useful lobes and inconvenient nulls. Trace the current, count the losses and let the installed pattern and propagation tell us what the antenna actually does.

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