The 96% SWR Myth
When “Hard Numbers” Don’t Measure Antenna Efficiency
The KJ6ER “Challenger Halfwave Antenna” (Rev: Feb 2025) describes a portable, end-fed half-wave style concept that can absolutely be made to work. The issue is not the concept. The issue is the confidence level of the claims versus the metrics actually shown. (This is an educational critique of measurement logic, not a verdict on whether someone can make contacts in the field.)
KJ6ER uses a 4:1 unun at the end-feed and highlights “higher efficiency” compared to high-ratio EFHW transformers. That direction can be sensible: a 4:1 transformer can indeed be lower loss than a high-turns-ratio transformer that is being pushed outside its comfort zone. The derail starts when the paper jumps from component insertion loss and SWR math into hard statements like “>94% efficient” for the antenna system, without demonstrating the dominant real-world loss mechanisms (ground loss, conductor loss, matching losses, and common-mode/feedline participation).
The recurring failure mode
This is the same pattern RF.Guru called out in “NECtasy in the Park”: take a tidy model, add a couple tidy numbers, then blur gain, directivity, radiation efficiency, component loss, and match into one story. You end up with “hard numbers” that are real… but they are measuring the wrong thing.
Mismatch efficiency → how much power is reflected due to mismatch (SWR story).
Insertion loss → how much power survives a transformer/choke (hardware story).
Radiation efficiency → how much accepted power is radiated vs lost as heat (soil, conductors, matching, common-mode).
Gain combines directivity and radiation efficiency (G = η · D). Mixing these is where “efficiency claims” go wrong.
Where the Challenger paper goes off the rails
Midpoint current is not an efficiency proof
A half-wave-ish radiator normally shows a current maximum near the middle. That is standing-wave physics, not a performance certificate. Radiation efficiency is about the ratio of radiated power to accepted power, and it requires accounting for losses that a current plot does not quantify.
Transformer loss is not antenna efficiency
Converting insertion loss in dB to a power ratio is valid math. And yes, a 4:1 can be lower loss than some high-ratio EFHW transformers. But that only answers: “how much power survives this component.” It does not answer: “how much accepted power becomes radiated RF.”
Once you start stacking components, the illusion gets worse: you can have “good” parts and still have a mediocre system if the dominant losses live elsewhere. Portable setups especially amplify the losses you did not model: ground coupling, near-field objects, operator proximity, feedline routing, and finite choke impedance.
The SWR trap: “1.5:1 means 96% efficient”
“96% at 1.5:1 SWR” is mismatch efficiency. It means roughly 96% of the forward power is not reflected due to mismatch. It does not mean 96% of accepted power is radiated.
A dummy load can show 1:1 SWR while radiating essentially nothing. So SWR cannot be used as “antenna efficiency.”
“No radials” becomes word games when a tuned counterpoise is required
If the system requires a short counterpoise wire to provide return current, then that wire is part of the antenna system in the functional sense. You can call it “counterpoise,” but electrically it is a ground-return conductor that behaves like a radial element in monopole-ish systems.
If a counterpoise is essential for resonance and stability, it cannot be dismissed as “minor impact” without showing current distribution (including the coax shield), and without showing stability under counterpoise-length and routing changes.
“The coax is not part of the antenna” needs current data
Advising a choke is correct. Treating the choke as a binary “problem solved” device is not. Chokes have finite common-mode impedance that varies by frequency, coax type, layout, height, nearby objects, and how the operator is positioned.
If you want to claim “predictable behavior” and “coax not participating,” you need at least one hard validation path: clamp-on current measurements, controlled A/B feedline routing tests, or a model that includes feedline plus choke as a common-mode impedance.
The “aluminum makes waves slower” explanation is misdirected
Length shifts in portable vertical-ish systems are far more plausibly explained by geometry and capacitance: end effects, effective diameter/taper, proximity to the tripod/mount, ground coupling, and nearby conductors. If the build requires careful isolation washers/gaskets and mechanical details to control coupling, that is already a clue that capacitance is the real knob.
NEC plots are useful, but not portable guarantees
NEC modeling is valuable for understanding trends and sensitivity. The failure mode is using NEC results as an oracle while skipping field validation. If you don’t validate assumptions (soil, clutter, coax routing, operator coupling, choke effectiveness), pretty plots can be very “true” in the model and very wrong in deployment.
Comparing antennas by peak angle and dBi while declaring “efficiency” is a category error
Gain already bundles directivity and radiation efficiency. If you want to compare “efficiency” across different antennas, you must separate directivity (pattern shape) from radiation losses, plus include feed/matching losses and common-mode/feedline participation. If that separation is not shown, an “efficiency conclusion” is not supported.
Claim → what it would take to prove it
| Claim | What would actually validate it (portable-real-world) |
|---|---|
| “>94% efficient antenna system” | Separate and report: mismatch, transformer loss, choke loss, and a radiation-efficiency estimate. Then validate with repeatable field-strength A/B tests against a known reference under controlled geometry. |
| “SWR ≤ 1.5:1 proves efficiency” | State it correctly as mismatch efficiency only, then provide separate evidence for radiation efficiency. |
| “Coax is not part of the antenna (because choke)” | Measure common-mode current vs frequency (clamp-on RF current probe), plus A/B tests with different feedline lengths and routings. |
| “Counterpoise is negligible” | Show current distribution on counterpoise and coax shield, and show stability when counterpoise length and routing are varied. |
| “NEC pattern proves field behavior” | Validate with controlled comparisons: same site, same time window, documented height/soil/coax route/operator position, and a repeatable reporting method (beaconing, fixed-distance field strength, or disciplined A/B on-the-air protocol). |
The takeaway
A 4:1 end-feed approach can be a practical, lower-loss choice. The problem is when the paper treats transformer insertion loss as antenna efficiency, treats SWR math as antenna efficiency, and treats NEC plots as field guarantees. That’s how “hard numbers” become hard claims with soft foundations.
Mini-FAQ
- Does low SWR mean high antenna efficiency? — No. Low SWR mainly indicates good impedance matching (low reflected power). It does not tell you how much accepted power is radiated versus lost as heat.
- Is transformer insertion loss the same as system efficiency? — No. It only measures losses in that component. Total efficiency also includes ground loss, conductor loss, matching losses, and common-mode/feedline losses.
- How do I check if the feedline is radiating? — Measure common-mode current on the coax with a clamp-on RF current probe, and do controlled A/B tests by changing feedline length and routing while observing repeatable signal metrics.
- Is a counterpoise basically a radial? — Functionally, yes: if it carries return current, it is part of the radiating/return system even if you choose a different name.
- Is NEC useless for portable antennas? — Not at all. NEC is excellent for trends and sensitivity. The mistake is skipping field validation and treating model outputs as guaranteed deployment results.
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 with your portable antenna measurements and test results.