Elevated Radials: What N6LF Measured—and What He Did Not
Elevated Radials: What N6LF Measured—and What He Did Not
Four carefully balanced elevated radials produced almost the same received signal as 64 surface radials on Rudy Severns’s 40 m test range. That is an important measurement—but it is not proof that every pair of elevated radials is 90% efficient.
Greg Mihran’s July 2026 primer reproduces Severns’s four-versus-64 comparison on slide 19. Elsewhere it attributes about 4 Ω loss and 90% efficiency to two elevated radials, calculates a 4.8 dB advantage over two surface radials and claims equivalence to 120 surface radials. Those are separate propositions. The four-radial field experiment does not establish them.
The result, stated accurately: at 7.2 MHz, four nearly balanced 33 ft radials elevated 48 inches produced normalised |S21| about 0.1 dB higher than 64 similar radials on the surface in one carefully controlled installation. Severns called the two systems equivalent for practical purposes in that experiment. He did not measure a universal 4 Ω loss or 90% radiation efficiency for two elevated radials.
1. The Experiment Everyone Quotes
The detailed source is Severns’s 2009 “Ground Systems, Part 3”. This was not a casual on-air comparison. Severns changed the radial system while holding the rest of the range as constant as practical.
| Test condition | What Severns used | Why it matters |
|---|---|---|
| Frequency and radiator | 7.2 MHz; 33.5 ft tubular aluminium vertical | One defined 40 m antenna, not a frequency-independent rule |
| Surface reference | 64 insulated 33 ft #18 AWG radials on the ground surface | The result belongs to this radial length, construction, soil and site |
| Elevated system | Four 33 ft radials and the antenna base raised through 0, 6, 12 and 48 inches | The cited measurement was four radials, not two |
| Feedline control | A common-mode choke in the feedline; antenna insulated from ground | This limited the coax shield acting as an uncontrolled radial |
| Current balance at 48 in | 0.235, 0.271, 0.247 and 0.247 A, normalised to 1 A base current | The four currents were unusually close to equal |
| Repeatability | The entire sequence repeated three times on different days | Severns reported no significant change between runs |
| Soil | Good-to-very-good soil at the test location | Severns explicitly called replication over other soils desirable |
All |S21| values were normalised to four quarter-wave radials lying on the surface:
- 64 surface radials: approximately +5.8 dB;
- four radials elevated 48 inches: approximately +5.9 dB.
The 0.1 dB difference was not practically meaningful for the experiment, and Severns described the two ground systems as equivalent for practical purposes.
That is a strong A/B result. It is also a bounded result.
2. What “Equivalent” Meant
It meant that the two configurations produced almost the same transmission magnitude on the controlled measurement path. It did not mean that their feedpoint impedance, current distribution and complete three-dimensional pattern were identical.
| Radial system | Measured feedpoint impedance | Normalised |S21| |
|---|---|---|
| 64 surface radials | 39.7 − j1.2 Ω | Approximately +5.8 dB |
| Four radials elevated 48 inches | 34.8 − j9.7 Ω | Approximately +5.9 dB |
The impedances are plainly different. Feedpoint resistance is therefore not a direct efficiency meter. A one-direction |S21| comparison also does not integrate the full radiation pattern or independently determine absolute radiation efficiency.
The defensible statement is narrower: with the documented radiator, choke, current balance, site and receive path, four elevated radials performed about as well as the 64-radial surface reference.
Relative field is not absolute efficiency. A configuration that matches a reference within 0.1 dB has nearly the same field on that path. Absolute efficiency still requires either an efficiency measurement or a validated power-and-pattern model of both systems.
3. How the Primer Extends the Claim
| Primer statement | What the source supports | What remains unproved |
|---|---|---|
| Slide 19: four radials at 4 ft are equivalent to 64 surface radials | A fair shorthand for Severns’s normalised |S21| result at 7.2 MHz | The slide omits current balance, choke, soil, impedance difference and one-path limitation |
| Slide 17: two radials at 36 in have about 4 Ω loss and 90% efficiency | Not the system Severns measured | Complete model, soil, currents, losses, feedline treatment and physical validation |
| Slide 20: two elevated radials provide a “lossless path” | The wire loss itself may be small; using a choke is sound advice | The radials still couple to soil, can be unequal and can drive other conductors |
| Slide 24: 90% versus 30% creates a 4.8 dB advantage | The logarithm is correct if the two efficiencies are already known | The arithmetic does not establish the assumed efficiencies |
| Slide 25: 120 surface radials equal two elevated radials | Severns compared four elevated with 64 surface radials | Neither the two-radial nor 120-radial configuration is the measured pair |
Four does not algebraically become two. Removing two radials changes mutual coupling, resonance, current division, azimuth symmetry, soil field and the opportunity for the coax exterior to carry return current.
4. Why Severns Later Advised 10–12 Radials
Severns kept investigating the limits of the successful four-radial result. In Elevated Ground Systems, Part 1, he showed how radial length, height, count, soil and vertical length interact in a resonant structure. In Part 2, he examined current asymmetry and ended with direct practical advice: use at least 10 to 12 radials, keep the radial system symmetric and away from nearby conductors, and use a base choke with more than about 2 kΩ shunt impedance.
This does not retract the four-radial experiment. It separates what four carefully controlled radials can do from what a copied field installation will do reliably.
5. Current Imbalance Is the Central Weakness
A nominally symmetric four-radial fan should divide current roughly four ways. The 2009 experiment came close because Severns deliberately minimised asymmetry.
Real installations are less cooperative. One radial may cross wetter soil; another may approach a fence, vehicle, mast, buried cable or tree. Small differences in length, height, insulation and end loading shift the resonance of these quarter-wave conductors.
Part 2 reproduces field data that make the sensitivity vivid. One two-radial system divided current 1.00:0.80 at 1.805 MHz but 1.00:0.10 at 1.885 MHz—only 80 kHz higher. Other systems placed almost all current in one radial.
Consequences can include:
- shifted feedpoint impedance and resonance;
- greater field in lossy soil;
- azimuth-pattern distortion;
- large and unequal RF voltage on radial ends;
- higher common-mode current on feedline or mast; and
- narrow, installation-sensitive useful bandwidth.
More radials do not make the environment disappear. They make each imperfect radial a smaller part of the total return structure and reduce sensitivity to one local asymmetry.
6. Soil Still Matters When Radials Are Elevated
Elevation can greatly reduce loss compared with putting only a few wires on or in the soil. It does not electromagnetically remove the earth.
The elevated conductors remain coupled to the soil beneath them. Conductivity, permittivity, moisture and terrain affect resonance, current division and dissipation. If the soil differs from one side of the fan to another, the radial currents can differ.
This is why a parallel-plate shortcut cannot calculate the result. Thin resonant wires over a lossy half-space have non-uniform fringing fields and distributed currents. Height is important, but it is not the only independent variable.
7. Nearby Conductors Become Part of the Problem
A vertical with perfect radials in an empty model may be beautifully symmetric. A garden or portable station contains coax, supports, guy hardware, fences, vehicles, gutters, masts, control wiring and buried services.
Any conductor can couple unequally to the radial system. With only two or four intended radials, one omitted conductor can be electromagnetically important. With more intended return conductors, the result is usually less dependent on that one path.
This is engineering margin, not magic. The complete installation remains the antenna.
8. The Coax Exterior Can Become an Extra Radial
The centre conductor and inside of the shield carry the wanted differential transmission-line mode. Current on the shield exterior belongs to an external common mode. If the radial currents and other return paths do not balance the feedpoint current, the coax exterior, mast or station earth can participate.
Then a “two-radial antenna” may actually be two radials plus a third uncontrolled conductor. That can change match, signal and pattern, sometimes making the sparse radial system appear better while reducing repeatability.
Severns used a common-mode choke in the four-versus-64 experiment. His later >2 kΩ recommendation is a useful starting point, not a universal guarantee. Measure common-mode current below the choke at the operating frequency and with the final feedline route.
9. Two Radials Deserve More Evidence, Not Less
Two opposite elevated radials can form a good antenna. Under favourable symmetry, they may be efficient. The criticism is not that two radials cannot work; it is that the four-radial measurement does not quantify a different two-radial system.
Two is the sparsest nominally balanced radial fan. Any change in one wire affects half the intended return structure. If one current falls, the other radial, the soil and external conductors must participate differently.
The primer’s efficiency equation is:
η = 37/(37 + 4) ≈ 90%
The arithmetic is fine; the 4 Ω input is the unverified part. Equivalent resistance is not measured merely by writing it in the denominator. A reproducible claim needs the complete model, ground, conductor loss, radial-current phasors, feedline/choke treatment, accepted power, radiated power and soil dissipation—or a recognised physical efficiency measurement.
10. Pattern Redistribution Can Mimic Gain
Unequal radial currents can tilt or skew the pattern. A receiving antenna in a favoured direction can report a stronger field even if total radiated power did not rise by the same amount.
That is why a one-path field comparison validates the field on that path, not automatically absolute efficiency. To distinguish efficiency from redistribution, measure enough of the pattern or use a validated radiated-power method.
The primer’s +4.8 dB is simply 10 log10(90/30). It is not independent evidence. If the 90% and 30% assumptions are unverified, their ratio is unverified too.
11. How to Test a Two-Radial Claim Properly
- Document frequency, radiator, radial geometry, wire, height, soil and nearby conductors.
- Characterise the feedpoint choke and measure coax-exterior current below it.
- Measure magnitude and phase of each radial current across the operating bandwidth.
- Compare two, four, eight and 12 radials at equal accepted feedpoint power.
- Retune every configuration without hiding matching-network loss.
- Measure several azimuths so pattern change is not relabelled efficiency.
- Repeat the sequence and publish uncertainty, not only the best run.
- Release the complete model, raw output and power budget.
A careful test may show that a particular two-radial system performs very well. That is useful installation-specific evidence. It still does not create a universal 4 Ω specification.
12. What 10–12 Radials Buys You
| Four or fewer tuned radials | At least 10–12 radials |
|---|---|
| Can perform very well under favourable symmetry | Retains the elevated-system concept with more tolerance of imperfect conditions |
| Each radial strongly affects current division | Local differences are averaged across more conductors |
| More sensitive to soil, nearby metal and exact length | Less sensitive to any single local disturbance |
| Greater opportunity for the feedline to become an unintended radial | More complete intended return structure, though a choke remains important |
| Often narrower and more tuning-sensitive | Generally more stable impedance and usable bandwidth |
Do not forget RF contact safety. Elevated radial ends can reach substantial RF voltage. Keep them out of reach of people and animals, use suitable insulation and support, and assess touch/contact risk at the intended transmitter power.
13. Takeaways You Can Trust
- Severns measured four elevated radials at 48 inches within about 0.1 dB of 64 surface radials on one 7.2 MHz receive path.
- The elevated radial currents were close to equal, and a feedline choke was part of the test.
- The result was specific to the geometry, frequency, soil, current balance, site and measurement path.
- It was not an absolute efficiency measurement of two radials, 4 Ω loss or equivalence to 120 surface radials.
- Sparse elevated systems are coupled resonant structures, not isolated lossless return wires.
- Current imbalance, soil asymmetry, nearby conductors and feedline current can change impedance, loss and pattern.
- Severns advised at least 10–12 radials mainly for robustness, symmetry and repeatability.
- Two radials can work well, but their performance must be demonstrated for the actual installation.
In Summary
The honest reading of Rudy Severns is more useful than the slogan. Four elevated radials can be remarkably effective: in a carefully balanced 40 m installation, they matched the measured path signal from 64 surface radials.
The same research showed why the result is easy to misuse. Sparse elevated systems are sensitive to radial resonance, current imbalance, soil, nearby conductors, frequency and feedline coupling. Those are part of the electromagnetic structure, not minor construction details.
Severns therefore moved from demonstrating what four radials can do to recommending at least 10–12 for what different installations can do more reliably.
Final point: four controlled radials versus 64 surface radials is measured evidence. Two radials, 4 Ω loss, 90% efficiency and equivalence to 120 surface radials are separate claims that still need disclosed, reproducible evidence.
Mini-FAQ
- Did Severns prove four elevated radials equal 64 surface radials? He found their normalised |S21| differed by about 0.1 dB in his controlled 7.2 MHz installation. He called them equivalent for practical purposes in that experiment.
- Did he prove two elevated radials are 90% efficient? No. The cited experiment used four elevated radials and did not directly measure absolute radiation efficiency.
- Why did he later advise 10–12 radials? More radials reduce sensitivity to unequal currents, soil variation, nearby conductors, exact tuning and feedline coupling.
- Can two elevated radials still work well? Yes. Their performance must be established for the actual installation rather than inherited from a four-radial experiment.
- Does a choke solve every imbalance? No. It can impede shield-exterior current but does not equalise radial resonances or remove soil and object coupling.
- What should be measured? Individual radial-current phasors, coax common-mode current, accepted power, impedance versus frequency and field in enough directions to reveal pattern change.