Balun Functions: Voltage, Current, Transformation and Choking

The argument becomes confused when one word is expected to answer five questions at once. Does the network transform impedance? Does it connect an unbalanced port to a balanced one? Does it impede common-mode current? Does it preserve equal branch currents into the installed load? Can it do all of that at operating power? Start by separating those jobs.

Balun, UNUN and Choke Describe Different Boundaries

A balun connects a balanced port to an unbalanced port. An UNUN connects two ports intended to be unbalanced. Either may have a 1:1 ratio or transform impedance. A common-mode choke adds impedance to a common-mode current path while passing the intended differential current.

Those descriptions concern ports and modes, not the shape of a toroid or the label on a box. A network can combine more than one function, but each function still needs its own evidence. A 4:1 marking does not prove that a balanced port remains balanced, and the word “balun” does not state the common-mode impedance.

Voltage and Current Are Not Competing Marketing Camps

A voltage-balun connection is intended to establish a voltage relationship at its output terminals. If the two load branches are unequal, equal-and-opposite terminal voltages do not guarantee equal-and-opposite branch currents. The unbalanced current must return somewhere, potentially through the feed line, station bonding, nearby conductors or stray capacitance.

A current-balun or choke connection raises the impedance of the unwanted common-mode path. It does not reach into arbitrary antenna branches and command identical currents regardless of the rest of the circuit. The achieved current balance depends on the choke impedance compared with the complete common-mode loop and on differential-to-common-mode conversion at the load and transition.

Guanella transmission-line arrangements can combine balance with an impedance transformation through series/parallel transmission-line sections. Ruthroff arrangements use conductive and magnetic coupling, often with series voltage addition. Either family can be connected in different ways. The inventor’s name alone does not establish the ports, ratio, common-mode impedance or load range of a finished device.

Impedance Ratio Is a Differential-Port Claim

For an ideal transformer with voltage ratio n, the impedance ratio is n². A nominal 4:1 impedance transformer therefore corresponds to a 2:1 voltage ratio under its defined port conditions. The real network departs from that ideal through finite magnetising inductance, leakage, transmission-line impedance, winding capacitance, conductor loss and core loss.

The ratio also does not create a 50-ohm load. It transforms the complex impedance presented at its output. If the antenna presents 320 − j120 ohms at the declared plane, a nominal ratio cannot erase the reactance or promise exactly 80 − j30 ohms across a multiband installation. Frequency, topology, termination and parasitics determine the measured result.

Port balance is separate again. A differential impedance measurement can look satisfactory while one branch current differs from the other or while current flows on the coax exterior. Measure the intended differential transformation and the unwanted mode rather than inferring one from the other.

Common Mode Exists on Transmit and Receive

Differential and common mode describe conductor-current relationships with respect to a declared reference. They do not change definition when the station switches from receive to transmit.

In the intended coaxial differential mode, centre-conductor current returns on the shield’s inner surface. A separate current on the shield exterior belongs to a different path involving the antenna, station and environment. On transmit it can radiate, alter the antenna pattern and produce RF voltage in unintended places. On receive it can carry wanted signal or locally coupled noise into a conversion point. In both directions, its magnitude and phase follow the complete installed loop.

The Dutch mantelstroomfilter and German Mantelwellensperre are helpful physical descriptions of a sheath-current filter. They do not make “common mode” technically wrong. The mode remains a valid circuit description whenever the conductors and reference are stated.

Height Does Not Select the Transformer

Ground and nearby structures can unbalance an antenna, but there is no universal half-wavelength height at which a current balun suddenly becomes valid. A low symmetric dipole can remain sufficiently balanced for one purpose; a high antenna can be disturbed by feed-line routing, a support, another wire or unequal capacitance.

Choose the network from the installed electrical boundary:

• Load symmetry: compare the complex impedance and current in each branch under the intended geometry.
• Return path: identify whether the coax exterior, counterpoise, radial system or another conductor is intentional.
• Required ratio: derive it from the measured complex load over the working band, not from an antenna name.
• Common-mode boundary: decide where an intentional return section should end and what impedance is needed there.
• Operating conditions: include waveform, power, mismatch, duty cycle, environment and failure margin.

Height matters because it changes coupling, impedance and pattern. It is an input to the installed measurement, not a selector switch between “voltage” and “current.”

Why Joeri Separates Transformation from Choking

After roughly two decades of studying baluns and UNUNs in HF installations, Joeri’s practical conclusion is that the textbook balanced case is uncommon in amateur stations. Feed-line routing, unequal capacitance to the surroundings, supports, nearby conductors and the station return path often disturb it.

RF.Guru’s practical default is therefore explicit: when the measured complex load calls for a 4:1 transformation and the transformer ports are intentionally unbalanced, use a 4:1 UNUN for that transformation and specify a separate, measured 1:1 choke for common-mode control. The choke defines the chosen end of the intended return structure; it is not automatically placed directly beside the UNUN.

This is a design method, not a universal winner. A genuinely balanced installed load can be well served by a suitable measured current balun. A deliberately arranged UNUN-plus-choke structure can also serve a balanced boundary when both functions and the branch currents are verified. An integrated transformer/choke design remains valid when its differential ratio, common-mode impedance, balance, loss and powered behaviour are demonstrated over the required loads.

Isolation Is Not Implied by the Name

Many Ruthroff and Guanella transmission-line transformer connections contain a conductive path between ports. They are not safety-isolation transformers. A balanced RF port may also sit at a common-mode voltage relative to earth even when its differential voltage is correct.

Check continuity, DC grounding, lightning and protective bonding as separate requirements. An RF choke can alter high-frequency current without satisfying electrical-safety isolation or lightning-current bonding. Conversely, a DC bond does not prove a low-impedance RF boundary across the working band.

Core Flux and Heat Follow the Actual Circuit

For a voltage-excited winding, the flux swing follows the time integral of winding voltage divided by turns and effective core area. The familiar V/(fNA) relationship is a sinusoidal approximation whose numerical factor depends on whether voltage is expressed as peak or RMS and on waveform.

Load still matters. It sets winding current and copper loss, changes the impedance reflected to the source and can alter waveform, terminal voltage and imbalance. Leakage flux, common-mode excitation and parasitic current can add core loss. It is therefore incomplete to say that a voltage transformer’s heating depends only on applied voltage.

In an ideal current-compensated choke, equal-and-opposite differential currents cancel their core excitation. A real assembly also sees residual imbalance, common-mode current, leakage, parasitic capacitance and mode conversion. The choke can heat without saturating, and nonlinear core excitation can occur before an external temperature reading looks alarming.

No universal temperature threshold proves permanent ferrite damage. Curie temperature, recommended operating temperature, permeability drift, coating, adhesive, winding insulation, connector and enclosure limits are material- and assembly-specific. A powered rating needs core material and geometry, winding construction, frequency, waveform, common-mode current, differential current, load, SWR, duty cycle, ambient temperature, cooling and permitted drift.

Measure Every Function You Plan to Claim

Small-signal VNA data is an essential start. Mixed-mode parameters distinguish differential transmission, common-mode transmission and conversion between them. A powered test is separate because ferrite, dielectric, insulation and conductor behaviour can change with flux, current, voltage and temperature.

Primary and Authoritative Technical Sources

• G. Guanella, “New Method of Impedance Matching in Radio-Frequency Circuits,” Brown Boveri Review, 1944—the original series/parallel transmission-line transformer treatment.
• C. L. Ruthroff, “Some Broad-Band Transformers,” Proceedings of the IRE, 1959—the primary broadband transmission-line transformer paper.
• Keysight, Balanced Measurements—differential, common-mode, mode-conversion and imbalance measurement definitions.
• Fair-Rite, Ferrite Cores for Low-Frequency EMI Cable Suppression—manufacturer treatment of a ferrite around a complete cable as a common-mode choke and the frequency dependence of its impedance.
• TDK, Ferrites and Accessories: Application Notes—broadband-transformer flux, frequency, load, distortion and core-selection boundaries.
• TDK, Ferrites and Accessories Data Book—material-, frequency-, flux- and temperature-dependent core and winding loss.

Joeri’s Bottom Line

Do not ask whether voltage baluns or current baluns are “better” in isolation. Ask which ports are meant to be balanced, what complex impedance must be transformed, where the return current is intended to flow, where it must stop and what the assembly does under the real load.

My default for the common, deliberately unbalanced HF installation is a measured-load UNUN plus a separately specified choke because it keeps those two jobs visible. That does not disqualify a current balun on a genuinely balanced load or a properly validated integrated design. The circuit and measurements decide; the label does not.

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