Roy Lewallen’s Balun Paper: What the Experiments Actually Showed

A balun is not proved by its name, turns ratio or SWR. Define the port conditions and unwanted mode, then measure whether the device supplies the transformation, balance and isolation the installation needs.

The Historical Claim We Can Verify

The American Radio Relay League lists Roy W. Lewallen, W7EL, and Baluns: What They Do and How They Do It in its History of QST. The article appeared in The ARRL Antenna Compendium, Volume 1, in 1985.

The paper matters because it asks what the balun is meant to accomplish, builds explicit test circuits and measures current balance. It does not establish that every product called a current balun is superior to every product called a voltage balun.

Start With Modes and Ports, Not Labels

For coax, the wanted transmission-line mode has current on the centre conductor and an equal opposing current on the shield's inner surface. A separate external mode can place net longitudinal current on the shield's outside surface, with its return through the antenna, mast, earth, station wiring and surrounding field.

At a balanced two-terminal port, “balance” also needs a reference. It may mean equal and opposite terminal currents, equal and opposite terminal voltages relative to a defined reference, or specified common-mode rejection. Those conditions coincide only in ideal symmetric cases.

A device can therefore perform several distinct jobs:

• transform differential impedance;
• provide a balanced-to-unbalanced port relationship;
• present common-mode impedance;
• provide or avoid a DC path;
• withstand differential and common-mode voltage and current; and
• limit differential insertion loss and heating.

“1:1 balun” or “4:1 balun” does not specify all of those properties.

The Coax Exterior Is an Available Path—not an Automatic Current

When the coax shield connects to one feedpoint terminal, its outside surface is physically continuous with that terminal. It is therefore an available conductor in the external electromagnetic structure.

But availability is not the same as a predetermined current. The current depends on the common-mode source created by asymmetry and coupling, the impedance of the antenna and surroundings, the coax route and length, and the termination at the station end.

A feedpoint choke is often good preventive engineering on a coax-fed dipole, especially when repeatable pattern and low station coupling matter. It is not a theorem that every symmetric dipole must have one. If measured external current is already negligible over the required bands, a second box may add little.

What Lewallen’s Bench Tests Actually Demonstrated

The official ARRL copy of the paper includes construction and test details in Appendix 2:

• the voltage balun used a modified RG-178/U and wire winding on an FT82-61 core;
• the current balun used 15 turns of RG-178/U on an FT82-61 core;
• the output test used unequal 27 Ω and 54 Ω resistors to a reference plane;
• the voltage-balun result depended on which load was connected to which output lead; and
• the current balun produced the predicted two-to-one load-voltage relationship within 0.2 dB with either lead assignment.

That is strong evidence for those circuits, components, loads, fixture and test frequency. It is not a timeless rating for every ferrite material, winding, frequency or installed antenna.

The paper's broader antenna experiments also showed the same ordering in the configurations tested: the current balun gave the best balance, the voltage balun was second and no balun was worst. Lewallen also identified results that he had not had time to explain. That scientific restraint should be preserved.

Voltage Balance Is Not Current Balance

If two branches have impedances Z₁ and Z₂, equal-magnitude terminal voltages do not generally create equal-magnitude currents when Z₁ ≠ Z₂. This is the useful core of the voltage-versus-current distinction.

However, “antennas radiate from current, not voltage” is too crude. Time-varying charge and current are linked by Maxwell's equations and the continuity equation. High voltage can also matter for insulation, arcing and electric-field exposure. The engineering question is which modal voltage and current the device controls.

A current balun does not make the physical antenna symmetric. It raises impedance in the unwanted common-mode path so the differential current relationship is less disturbed by that path. Residual asymmetry, displacement current to nearby objects and finite choke impedance remain.

Ruthroff and Guanella Are Families, Not Quality Grades

Ruthroff and Guanella describe circuit families and transmission-line connections. They are not synonyms for “bad voltage device” and “good current device.” Performance depends on the exact schematic, transmission-line impedance, winding arrangement, coupling, parasitics, core material and termination.

A Ruthroff-style transformer may provide an efficient impedance transformation in its intended circuit. A Guanella arrangement can provide transformation and useful common-mode isolation. Neither name supplies a frequency response, power rating or installed common-mode impedance.

For a real design, publish the circuit and measure the ports. Topology is the beginning of the specification, not the end.

Balanced Line Can Radiate Even With Opposing Currents

Open-wire or ladder line is not made non-radiating by its name. With equal opposing currents and small conductor spacing relative to wavelength, far fields largely cancel. Cancellation is not mathematically perfect for finite spacing, bends, transitions and nearby objects.

Unequal currents generally make radiation and pickup worse because a net common-mode component remains. But “equal currents means zero radiation” is also an idealisation. Keep the line symmetric to its environment, avoid abrupt routing changes, and measure when line radiation matters.

Choke Placement Defines a Boundary

A choke works in the common-mode circuit at its installed position. Moving it changes the lengths and impedances on both sides, so the resulting current can change even when the choke itself does not.

This is why the best location is derived from the intended antenna boundary and measured current path, not from a universal one-, two- or three-choke rule.

What to Measure in 2026

1. Publish the schematic. State which conductors form each port and whether transformation and choking share a structure.
2. Measure complex common-mode impedance. Report resistance and reactance versus frequency, not only an unexplained dB number.
3. Measure differential insertion loss and match. A strong common-mode choke can still be a poor differential transmission device.
4. Test unequal loads. Swap the load branches as Lewallen did so hidden reference asymmetry becomes visible.
5. Probe the installed feedline. Measure at several positions and on every operating band; one standing-wave minimum can mislead.
6. Check thermal behaviour. Use actual duty cycle, mismatch and common-mode current. Core count or transmitter watts alone is not a rating.
7. Recheck antenna behaviour. A changed SWR after adding the choke may reveal a changed antenna boundary, not a defective choke.

What Lewallen Got Right—and What the Re-Read Changes

The Practical Verdict

Lewallen did not make baluns stop being difficult. He made them testable. That is the paper's lasting achievement.

Use “current balun,” “voltage balun,” “Ruthroff,” “Guanella,” “1:1” and “4:1” as circuit descriptions—not performance guarantees. Define the unwanted mode, measure the actual device, install it at a deliberate boundary and verify current in the complete antenna system.

The memorable line is not “every dipole needs a choke.” It is: the current path decides whether the balun solved the problem.

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