EFHW Efficiency: Any Antenna Works. That Was Never the Issue.
Any antenna works.
A wet string works. A dummy load with a short wire works. A badly installed end-fed half-wave works. That is not the discussion.
The real question is not: does it make QSOs? The real question is: where do the watts go? Into radiated field, in the wanted current distribution and pattern? Or into ferrite, copper, common-mode feedline current, lossy return paths, tuner loss, ground loss, nearby objects and heat?
This is not about liking or disliking the EFHW. It is not antenna religion. It is just what is and what is not.
The Video That Triggered This Article
Peter G3OJV made a video about EFHW efficiency. The video contains useful practical reminders, especially about not panicking over small coax losses. But then the argument slides from a narrow coax-loss point into a much broader EFHW-efficiency conclusion. That is where the engineering starts to go off the rails.
Click any timestamp in the table below to jump to that part of the video.
Where the Argument Goes Off the Rails
| Timestamp | Claim or implication | Why this needs correction |
|---|---|---|
| An EFHW is described as a half-wave wire on the lowest band that operates on harmonics above it. | That is the simplified brochure version. Real open-ended wire resonances do not land perfectly where clean octave arithmetic says they should. End effect, wire diameter, installation height, transformer parasitics, nearby objects and return-path behavior all move the result. Multiband is not the same as broadband octave behavior. | |
| A 20 m wire is said to be resonant on 40 m and also operate on 20, 15 and 10 m. | A 20 m wire can support useful higher-order modes, but saying it is simply “resonant” across those bands hides the real behavior. The current distribution changes radically on the higher bands, feed impedance is not stable, and 15 m is not an octave step. | |
| The discussion moves to coax loss and a typical 2:1 VSWR. | This is not wrong by itself, but it answers a different question. Coax mismatch loss is not EFHW transformer loss, not ferrite loss, not common-mode loss, and not radiation efficiency. SWR and efficiency are not the same measurement. | |
| The extra coax loss from a 2:1 VSWR is described as very small. | Usually true on a short HF coax run, but still incomplete. Mismatch loss is not transmission loss. A system can accept power and still dissipate a large part of it elsewhere. A pretty SWR curve can be caused by a good antenna, a lossy antenna, or a cleverly hidden problem. | |
| After discussing coax, the video moves toward “realistically” estimating total EFHW efficiency. | That jump is too large. You cannot move from modest coax-loss numbers to overall EFHW efficiency without measuring the transformer under realistic complex loads, the return path, common-mode current, heating and actual radiated field. | |
| A generic insertion-loss response based on FT240-43 material is used as part of the argument. | Insertion loss into a friendly resistive load is not the same as loss in a real EFHW transformer connected to a high-voltage, high-impedance, frequency-dependent and reactive antenna. A real EFHW is not a fixed resistor. | |
| The FT240-43 response is shown as having a sweet spot around the mid-HF range, with more loss lower and higher. | That observation is important, but the conclusion should be stronger: one ferrite material and one winding geometry cannot be optimal from 80 to 10 m. Low bands need magnetizing inductance and core area. Upper HF punishes excess turns with leakage inductance, distributed capacitance and self-resonance. | |
| The curve is stated to be based on a well-matched antenna. | That is exactly the hidden assumption. EFHW feed impedance is rarely a stable, well-behaved, purely resistive load across all bands. Once the load becomes reactive, transformer stress and loss can change substantially. | |
| The figures are said to be based on a good installation, set up and resonant. | That condition does most of the work. A good installation is not proof that a generic EFHW design is efficient. It also does not define what “resonant” means: wire resonance, box-compensated resonance, or a system resonance involving coax and environment? | |
| Coax losses occur in all antenna systems, and the EFHW may have shorter coax. | True, but incomplete. In an EFHW, the coax may not be just feedline. Without a defined return path and proper choking strategy, the outside of the coax can become part of the antenna system. Then the coax is also a radiator, noise pickup path and loss contributor. | |
| A larger or higher-rated transformer is recommended to reduce losses. | That advice is directionally useful. More core volume and better winding geometry can help. But a bigger box does not automatically solve the 80–10 m broadband compromise, ferrite-material choice, winding capacitance, reactive load behavior or common-mode return path. | |
| All multiband antennas have increased losses, so it becomes a bit of a level playing field. | That is false equivalence. Trap loss, tuner loss, feedline loss, high-ratio ferrite loss, common-mode loss and ground loss are different mechanisms. Different systems lose power in different places and by different amounts. | |
| A line isolator at the shack end is recommended. | A shack-end isolator can help, but it does not define the EFHW return path at the feedpoint. If the coax has already acted as the counterpoise before it reaches the shack, the pattern, noise pickup and loss behavior have already been affected. |
Mismatch Loss Is Not Transmission Loss
The video starts with a mostly correct small point and then builds the wrong comfort story on top of it.
Yes, a 2:1 SWR on a short run of decent coax at HF is usually not catastrophic. Nobody serious says otherwise. But mismatch loss is not transmission loss. SWR and return loss are about reflected power. Transmission loss and insertion loss are about real power being dissipated as heat.
Those are different mechanisms. Fixing one does not automatically fix the other.
A low SWR reading only tells you that the transmitter sees a load it can work into. It does not tell you how much of the accepted power becomes useful radiation. A dummy load has excellent SWR. That does not make it an efficient antenna.
The Coax Argument Is Not the EFHW Efficiency Argument
The coax-loss part is the easy part. A short length of RG213 or similar cable on 40 m or 20 m will not usually eat several dB just because the SWR is around 2:1. That is a useful correction against SWR panic.
But that does not answer the actual EFHW question.
The EFHW efficiency question is not simply: how much extra loss does the coax have at 2:1?
The real questions are:
- How much power is dissipated in the transformer core?
- How much is lost in winding resistance and circulating reactive current?
- How much current flows on the outside of the coax shield?
- How much of the coax has become part of the antenna?
- How much of the system current is in the intended radiator?
- What happens on the upper bands where the current pattern is no longer simple?
- What does the system do under real reactive loads, not only resistive test loads?
Until those are answered, the coax calculation is only a side note.
A 49:1 Transformer Is Not Being Used Into a Nice Resistor
Insertion-loss measurements can be useful. Back-to-back transformer tests can be useful. A resistively terminated transformer test can be useful. But none of those automatically proves the efficiency of a complete EFHW antenna.
A real EFHW feedpoint is not a stable 2450 Ω resistor.
It is high impedance, frequency dependent, installation dependent and often strongly reactive. The wire routing, end effect, height, slope, nearby objects, soil, coax routing, counterpoise length and choke position all affect the real load seen by the transformer.
When the transformer is tested into a clean resistive load, much of that ugly reality is removed. The test may tell you something about the transformer under controlled conditions, but it does not tell you what happens when the box is connected to a real end-fed wire on every HF band.
“Resonant on 80, 40, 15 and 10” Is Too Clean
The harmonic story is often sold as if the EFHW were a perfect octave machine: half-wave on the lowest band, then everything neatly repeats above it.
That is not how real open-ended wires behave.
A wire can support higher-order standing-wave modes. That part is true. But higher-order operation does not mean the feedpoint impedance remains convenient, the current distribution remains useful, or the transformer sees a friendly load.
Also, 15 m is not an octave step from 40 m. It can be a useful harmonic-mode band for certain lengths, but it is not evidence that the antenna is a clean broadband octave structure.
On the higher bands, the current distribution becomes more complex. Lobes appear. Nulls appear. Feedpoint impedance changes. Transformer parasitics matter more. The shunt capacitance, winding capacitance and coax return path can all become part of the result.
One Ferrite Material Cannot Be Perfect Everywhere
A single ferrite material and one winding layout cannot magically be optimal from 80 to 10 m.
Low HF wants enough magnetizing inductance, enough core area, enough turns and enough thermal margin. Upper HF punishes too many turns with leakage inductance, distributed capacitance, self-resonance and winding loss.
Mix 43 can be good at certain jobs. Mix 52 can be good at other jobs. But “good” is not the same as “universal.” The correct material depends on ratio, power, duty cycle, lowest band, highest band, core stack, winding geometry and actual load impedance.
This is why serious transformer design is not just “grab a ferrite ring, wind the internet recipe, add a capacitor and call it 80–10 m.”
The Missing Capacitor Discussion
The common EFHW capacitor deserves its own warning.
That capacitor is not a magic efficiency component. In many EFHW boxes it is a shunt element used to tame the upper-band match. It can make the SWR curve look prettier. It can pull a dip into a band. It can make the radio happy.
But that does not prove the wire is resonant. It does not prove the transformer is efficient. It does not prove the antenna is radiating better.
A shunt capacitor can mask transformer behavior by moving the apparent match. It can also create circulating reactive current inside the transformer system. That current does not radiate. It heats copper, ferrite, capacitor ESR and connections.
The Return Path Cannot Be Ignored
An end-fed antenna always needs return current. Calling it “end-fed” does not delete the other side of the RF circuit.
If the design does not provide a controlled return path, the system will invent one. That return path may be the outside of the coax shield, the mast, shack wiring, the operator, nearby structures, ground capacitance or a mixture of all of them.
A line isolator at the shack end can help reduce RF in the shack, but it does not automatically fix the antenna system. If several meters of coax have already acted as the counterpoise before the choke, then that coax section has already become part of the radiator and pickup system.
This is why EFHW installations can be quiet in one location and noisy in another. It is also why two people can use the “same” antenna and get very different results. They are not using the same antenna system. They are using different radiator, transformer, coax, return-path and environment combinations.
The “All Multiband Antennas Lose Something” Argument Is Not Engineering
It is true that multiband antennas involve compromises. But that does not make all compromises equal.
A trap dipole does not lose power in the same way as a high-ratio EFHW transformer. A 4:1 off-center-fed dipole transformer is not stressed like a 49:1 EFHW transformer. A doublet with open-wire feed and an external tuner is not the same system as an EFHW box with a hidden shunt capacitor and undefined common-mode return path.
Different antennas have different loss mechanisms. Different loss mechanisms have different magnitudes. You do not average them into “everything loses a bit” and call the argument finished.
What Would a Better EFHW Efficiency Test Look Like?
A better test would not stop at SWR and coax-loss calculators. It would look at the whole system.
- Measure transformer temperature rise at real power and duty cycle.
- Test transformer loss under realistic complex loads, not only resistive loads.
- Measure common-mode current on the coax shield at multiple points.
- Define and document the counterpoise or return path.
- Compare field strength against a known reference antenna.
- Repeat the test across bands, not only at the easiest operating point.
- Separate SWR, transformer loss, feedline loss, common-mode behavior and radiation performance.
Only then are we talking about antenna efficiency instead of comfort numbers.
The Correct Conclusion
The fair conclusion is narrow:
A short HF coax run with moderate SWR is usually not the main loss disaster. Fine.
But that does not prove that a typical 49:1 EFHW is efficient from 80 to 10 m.
To claim that, you need to show transformer loss under realistic complex loads, temperature rise at real power and duty cycle, feedline and common-mode current, counterpoise behavior, actual current distribution and ideally field-strength comparison against a known reference.
Until then, “it works” remains only “it works.”
It does not become an efficiency measurement.
Mini-FAQ
- Does RF.Guru say EFHW antennas do not work? No. Any antenna can work. The question is not whether it makes contacts, but how efficiently and predictably the system converts transmitter power into useful radiation.
- Is a 2:1 SWR on short coax a disaster? Usually not. The problem is using that true statement to imply that the complete EFHW system is efficient.
- Can SWR prove antenna efficiency? No. SWR shows how well power is accepted at a measurement point. It does not separate radiation from loss.
- Is a shunt capacitor always bad? No. A capacitor can be part of a valid design. The problem is using it to make the SWR curve look better while hiding transformer limitations or reactive circulating current.
- Where should an EFHW choke or isolator go? That depends on the intended return path. A shack-end isolator may help, but it does not automatically define the antenna system at the feedpoint.
Interested in more technical content? Subscribe to our updates for deep-dive RF articles and lab notes.
Questions or experiences to share? Feel free to contact RF.Guru for practical antenna-system support.