Short Radials and Rudy Severns: Read the Test Before the Rule
Short Radials and Rudy Severns: Read the Test Before the Rule
“Rudy Severns proved short radials are just as good” is not an engineering conclusion. Neither is “long radials always win.” N6LF measured specific verticals, radial fields, soils and frequencies. His work shows why radial count, length, height, symmetry and the complete current path must remain attached to the result.
The misreading I object to is not the existence of a useful short radial. It is taking one conditional comparison and turning it into a shopping-list rule. Rudy’s own reports repeatedly name the antenna, frequency, soil, radial geometry, common-mode isolation and measured quantity. Remove those conditions and the result no longer says what people claim it says.
My practical position: use as much useful-length conductor and radial coverage as the site permits, then verify the installed system. More short radials can beat fewer long ones in a fixed-wire comparison; longer radials can win in another geometry; and a sparse radial screen can develop loss-producing resonant current. The correct choice comes from the complete antenna, not one length or count slogan.
What the Popular Shortcut Leaves Out
A ground-mounted vertical drives displacement and conduction current through the soil as well as current in the radial wires. Radials change the current density close to the base, the loss in the surrounding earth, the feedpoint impedance and sometimes the radiation pattern. They are not interchangeable pieces of copper whose value can be predicted from total length alone.
Five variables travel together:
- Count and angular spacing: these determine how well conductor coverage is distributed around the high-current region near the feedpoint.
- Individual length: this changes coverage, radial current distribution and the possibility of resonant behaviour close to the soil.
- Soil: conductivity and complex permittivity vary with composition, moisture, temperature, depth and frequency.
- Radiator: height, loading and base current change the electric and magnetic fields that interact with the ground system.
- Unintended conductors: the coax exterior, mast, ground stakes, fences and station bonding can become parallel return paths.
A low SWR does not identify how much accepted power became radiation or loss. Feedpoint resistance is also not “ideal radiation resistance plus ground loss” unless the radiation resistance and all other losses are known at the same current reference. Severns warns explicitly that radiation resistance itself changes with the ground system.
What N6LF Measured With Sparse Surface Radials
In Experimental Determination of Ground System Performance for HF Verticals, Part 2, Rudy investigated excessive loss in sparse radial screens.
One test used a 34-foot aluminium vertical at 7.2 MHz with four insulated radials lying on the surface. Their equal length was varied from 33 feet down to 18 feet. The antenna and radials were isolated from ground and feedline common mode with a choke, while feedpoint impedance, relative transmission and radial-current division were recorded. In that installation, shortening the four radials towards roughly 23 feet improved the normalised transmission result because the 33-foot wires were soil-loaded and exhibited a loss-producing radial resonance.
That is not proof that 23 feet is a universal 40-metre radial length. It is evidence that a free-space quarter wavelength can be the wrong design assumption for a sparse conductor lying on real soil.
Rudy then used a remotely adjustable vertical, re-resonated it as the ground system changed, and compared 4, 8, 16 and 32 radials at both 33 feet and 21 feet. The shortening advantage became much smaller as count increased. With 32 radials, the 33-foot set was only 0.12 dB ahead in that particular comparison—small enough that Rudy did not pursue the intermediate optimum.
| What the experiment supports | What it does not support |
|---|---|
| A sparse on-ground radial field can have a length-dependent resonance and loss. | Every short radial is efficient or every quarter-wave surface radial is harmful. |
| Increasing count reduced sensitivity to radial length in this 7.2 MHz installation. | One universal count at every frequency, soil and radiator height. |
| Complex feedpoint impedance changed as the radial field changed. | That SWR or resistance alone measured radiation efficiency. |
| Choke isolation was part of the experiment. | A field result that can be copied while allowing the coax exterior to become an uncontrolled extra radial. |
The Equal-Wire Comparison Is Often Quoted Without Its Boundary
In Ground System Performance for HF Verticals, Part 4, Rudy reported a fixed-wire comparison at 7.2 MHz using a top-loaded one-eighth-wave vertical. Sixteen 33-foot radials and thirty-two 17-foot radials produced almost the same measured relative signal. The feedpoint impedances were substantially different, and the radiator had to be lengthened to restore resonance with the shorter set.
This is the source of a useful design idea: when the amount of wire is fixed, distributing it into more shorter conductors can improve coverage near the base and can equal or outperform fewer longer conductors in a particular installation. It is not a proof that radial length never matters. Rudy’s own conclusion was to model or calculate the particular vertical and soil when space forces that trade.
The same paper tested several different 7.2 MHz verticals over measured soil, adding 33-foot radials in the sequence 0, 4, 8, 16, 32 and 64. Most of the measured improvement in those cases arrived with the first 16 radials, while additional conductors still gave smaller gains. That is a result for those antennas and that soil—not a law that sixteen is always sufficient.
Elevated Radials Are a Different Antenna
Surface and elevated radial systems should not be mixed into one rule. In Part 3 of the 2009 series, a controlled 7.2 MHz experiment compared a 64-radial surface system with a carefully balanced four-radial system elevated 48 inches. The normalised transmission results differed by only 0.1 dB under those test conditions.
The result is real; the shortcut built from it is not. In his 2012 follow-up on elevated ground systems, Rudy stressed that a four-radial fan is sensitive to asymmetry, nearby conductors and soil variation. Those changes can alter radial-current division, resonance, feedpoint impedance, pattern and efficiency. He therefore recommended 10–12 or more elevated radials for a more robust installation.
That later recommendation is still not a universal minimum. An elevated system must be treated as a coupled antenna:
- measure current in each radial rather than assuming equal division;
- declare radial height, slope, length, count and angular spacing;
- control or measure current on the feedline exterior;
- keep high RF voltage and accessible conductors within an installation-specific safety assessment; and
- validate the pattern when asymmetry or nearby metal matters.
Why “Long Always Wins” Is Also Wrong
Length is valuable only when the additional conductor produces a useful current distribution. With too few surface radials, extending each wire can move a radial resonance into an operating band and increase earth loss. In another installation, longer wires may reduce loss or improve the desired field. With enough conductors, the difference between two lengths may become small.
The sentence I am willing to defend is narrower: do not discard useful conductor length merely because someone extracted “many short radials” from one graph. Compare candidate fields using the same wire budget, the same radiator, the same feedline boundary and measured or defensible soil data. Then check current, accepted power and a repeatable field result.
ITU-R P.527-6 describes the Earth’s surface with frequency-dependent complex permittivity and conductivity, including moisture, temperature and layered-ground effects. “Average ground” is a model input, not a substitute for the site.
Current VertX Radial Options
For current VertX installations, RF.Guru offers two equal-length radial kits:
| Current option | Supplied conductor | Engineering boundary |
|---|---|---|
| 24 × 5 m radials | Twenty-four equal 5 m conductors; 120 m total supplied wire | A current physical configuration, not a guaranteed efficiency, gain or soil-independent optimum |
| 32 × 5 m radials | Thirty-two equal 5 m conductors; 160 m total supplied wire | More equal-length coverage than the 24-wire option, but still installation-dependent |
The equal 5 m format keeps the physical choice clear and avoids pretending that a mixed bundle contains a hidden universal optimum. Select between the current options by available area, angular coverage, soil and installation constraints. Neither kit description replaces an installed current map, accepted-power comparison or repeatable field measurement.
A Fair Radial Test
- State the objective. Decide whether the comparison concerns accepted power, radiated field in a direction, feedpoint impedance, bandwidth, pattern stability or material cost.
- Hold the radiator fixed. Record its height, loading, conductor, feedpoint and nearby structures.
- Declare the radial field. Record count, length, spacing, height, soil contact, conductor and total wire.
- Control the feedline path. Use a characterised choke and map current around the complete coax at repeatable positions.
- Record the vector impedance. Save R + jX at the same calibrated plane; do not compare only the lowest SWR.
- Measure individual current where practical. Unequal radial current can reveal asymmetry that the input match hides.
- Compare accepted power. Correct for feedline and matching loss before comparing remote field or received signal.
- Use A/B/A restoration. Reinstall the first field and confirm that weather, soil moisture and geometry did not create the apparent result.
- Repeat by band. A multiband radial field can redistribute current and develop different resonances at every frequency.
Primary Evidence
- Rudy Severns, N6LF — Experimental Determination of Ground System Performance for HF Verticals, Part 2: sparse surface screens, radial resonance, length, count and measured S21.
- Rudy Severns, N6LF — Ground System Performance for HF Verticals, Part 4: radial-count experiments, fixed-wire comparison and feedpoint-resistance limits.
- Rudy Severns, N6LF — Ground System Performance for HF Verticals, Part 3: controlled surface-versus-elevated comparison.
- Rudy Severns, N6LF — A Closer Look at Vertical Antennas With Elevated Ground Systems: radial height, coupling, asymmetry and robustness.
- Rudy Severns, N6LF — Radial System Design and Efficiency in HF Verticals: NEC study of radial length, count, soil and total-wire tradeoffs.
- ITU-R P.527-6 — Electrical Characteristics of the Surface of the Earth: frequency-, moisture-, temperature- and layer-dependent ground properties.
Joeri's Bottom Line
Rudy Severns did not give us permission to stop thinking. He showed that a sparse radial screen can behave badly, that shortening can sometimes improve it, that equal wire can be distributed more effectively, and that a few elevated radials require unusually careful symmetry and isolation.
So I will not replace “short is always enough” with “long always wins.” I will keep the test conditions attached to the result. Put useful copper where the installed current needs it, control the other return paths, and prove the decision with the same disciplined measurements that made Rudy’s work valuable in the first place.
Mini-FAQ
- Did N6LF prove that short radials are always as good as long radials? No. His results belong to declared frequencies, soils, radiators, counts, lengths and feedline boundaries. Some shorter sets performed better; other length differences became small or reversed.
- Why can shortening a surface radial improve a sparse system? A wire close to soil can be electrically loaded. In a sparse field, a radial resonance can raise current and ground loss; changing length can move that condition.
- Does more total radial wire always produce more signal? No. Distribution, count, length, soil and radiator current all matter. Compare the same accepted power and restored installation rather than total wire alone.
- Are four elevated radials equivalent to 64 surface radials? They were within 0.1 dB in one carefully controlled 7.2 MHz experiment. N6LF later stressed that a four-radial elevated system is sensitive to asymmetry, nearby conductors and soil variation.
- Do the 24 × 5 m and 32 × 5 m VertX kits have guaranteed efficiency? No. They are current equal-length hardware configurations. Installed efficiency and pattern still depend on soil, layout, return currents and the complete antenna.
- What should I measure when comparing radial fields? Record R + jX at one plane, individual and feedline-exterior currents, accepted power, geometry, soil condition and an A/B/A remote-field result on every band.