The Guanella Transformer: Current Balun and UNUN Explained

The Guanella transformer is often introduced with three shortcuts: it is a “current transformer,” it forces equal currents, and two or three lines automatically produce a broadband 4:1 or 9:1 device. Those phrases point in the right direction, but they conceal the engineering. We need to follow the intended transmission-line mode, the common-mode path and the external port references separately.

Guanella's Idea Was a Transmission-Line Network

Gustav Guanella's high-frequency matching-transformer patent claimed priority in Switzerland in 1944, was filed in the United States in 1945 and was granted as US 2,470,307 in 1949. The important idea is not simply “wire on ferrite.” It is a transformer system with transmission-line character whose sections are interconnected to change the port voltage and current relationship.

A transmission-line section carries an intended differential mode: current travels out on one conductor and returns on the other. If that line is wound through magnetic material, the equal-and-opposite differential currents ideally cancel their magnetizing effect while an in-phase or common-mode current encounters inductive and lossy impedance. Real cancellation, transmission and choking are finite because conductors, cores, leads and fields are not ideal.

This is not a conventional transformer with a primary winding magnetically inducing a secondary voltage through mutual flux. Calling the input “primary” and the output “secondary” can hide the fact that the signal propagates along connected transmission-line conductors. Some common-mode flux may still be stored and dissipated in the core; it is simply not the intended mechanism for differential power transfer.

The 1:1 Connection Is About Mode Control

A practical 1:1 Guanella current balun is one transmission-line section arranged so that intended differential current passes with little insertion loss while unwanted in-phase current sees a larger impedance. With coax, the intended mode is current on the centre conductor and the inner surface of the shield. Additional net current can flow on the shield exterior relative to the antenna environment.

The choke does not “push” current into equality by magic. Its finite complex common-mode impedance changes the external-current loop. If that impedance is large relative to the rest of the loop, mode conversion and exterior current can be reduced. If another low-impedance path exists through a mast, enclosure, counterpoise, control cable or station wiring, equal antenna-branch currents are not guaranteed.

At 1:1 there is no intended differential impedance transformation. The useful results are differential insertion and return loss, amplitude and phase balance at the balanced port, mode conversion, common-mode impedance and installed current. One low SWR trace proves none of those on its own.

Series and Parallel Connections Create Nominal Ratios

In the classical Guanella series-parallel arrangement, n equivalent transmission-line sections are paralleled at one port and connected in series at the other. The ideal voltage ratio is n, so the ideal resistive impedance ratio is n²:

That gives the familiar square ratios: one section for 1:1, two sections for a nominal 4:1 impedance transformation, and three for nominal 9:1. Reversing the ports reverses the transformation direction. Other transmission-line-transformer networks and combinations exist, so a ratio label must still be tied to an actual circuit diagram.

“4:1” means an ideal impedance relationship, not a promise that every 200-ohm antenna becomes 50 ohms. The load may be reactive, the line sections have finite characteristic impedance and electrical length, and the interconnections add leakage and parasitic capacitance. Transformation accuracy, phase and loss therefore vary with frequency and termination.

The same boundary applies to a nominal 9:1 network. It is not automatically the correct interface for a random wire, nor does it certify a 450-to-50-ohm match over an amateur-band span. Three line sections create a possible ideal ratio; the installed complex load decides whether that ratio is useful.

BALUN and UNUN Describe the External Ports

A BALUN connects an unbalanced port to a balanced port; an UNUN connects ports that are both intentionally unbalanced. The Guanella series-parallel principle can appear in networks used for either purpose. The external terminal references, interconnection and common-mode paths decide the port type—not Guanella's name by itself.

A balanced port supports the intended differential mode with equal-and-opposite branch currents relative to the surrounding reference. That does not require equal branch voltages when the load arms have different impedances. Conversely, two equal-and-opposite voltages into unequal load impedances do not prove equal currents.

Port balance is also different from common-mode impedance. A transformer can show good differential amplitude and phase balance in a controlled load while presenting insufficient impedance to an installed exterior-current loop. It can also measure a high common-mode impedance in one fixture while its differential transformation is poor into the intended reactive load.

Guanella and Ruthroff Solve Different Circuit Problems

Ruthroff transmission-line transformers use a different interconnection in which a delayed or bootstrapped transmission-line voltage is combined with another port voltage. This can produce useful broadband voltage and impedance transformation. A Guanella network instead uses transmission-line sections and their series-parallel port relationship, with common-mode impedance developed by the line sections and magnetic loading.

That distinction does not make one family universally more efficient, quieter or better. Differential insertion loss depends on line impedance, conductor and dielectric loss, core loading, parasitics, termination and frequency. Common-mode behaviour depends on the external path and the impedance presented by the completed assembly. A Ruthroff-derived transformer may be the right impedance transformer when a separate measured choke defines the common-mode boundary; a well-designed Guanella network may combine both functions successfully.

Transformation and common-mode control should remain separate entries in the test record even when one enclosure performs both. If the required ratio and choke bandwidth pull the design in different directions, using an impedance transformer plus a separately specified choke can make the two jobs easier to verify. That is an engineering choice, not a universal topology ranking.

The Core and Winding Set Real Limits

The intended differential signal sees a transmission line, so its characteristic impedance, propagation delay, conductor size, dielectric and termination matter. The common-mode signal sees the winding as a coupled inductive structure loaded by complex permeability, core loss and parasitic capacitance. These are two different equivalent circuits sharing the same physical conductors.

At the low-frequency edge, insufficient common-mode inductive impedance and magnetic excitation can limit balance or increase core loss. At the high-frequency edge, winding electrical length, leakage inductance, inter-turn capacitance, section mismatch and resonances can degrade transformation and create mode conversion. The useful range may end for one function before it ends for another.

More cores or turns are not automatically better. Turns can raise low-frequency common-mode impedance while adding capacitance, delay and differential loss. Core material data are measured under stated geometry, frequency, field and temperature conditions; they are design inputs, not a completed transformer rating.

Power handling likewise cannot be inferred from topology or core size alone. Differential current heats conductors and connections; common-mode current can dissipate energy in the core; mismatch can raise terminal voltage or current; waveform and duty cycle set average heating. Qualify the finished assembly at declared frequency, complex load, accepted power, duty cycle, ambient and permitted temperature rise.

Measure Each Function at Its Own Reference Plane

• Record the circuit and port references. Draw every terminal, shield, chassis, jumper and intended return branch. State which logical port is single-ended, balanced, differential or common mode.
• Calibrate to the transformer terminals. Use a calibration or defensible de-embedding method and bound fixture leakage, capacitance, inductance and dynamic range.
• Measure differential transfer. Save complex return and insertion data, transformation ratio and phase over representative resistive and reactive loads.
• Measure balance and conversion. A multiport mixed-mode measurement can separate differential transmission, common transmission and differential-to-common conversion.
• Measure common-mode impedance. Save resistance and reactance, not only magnitude, with a series-through or other validated fixture suited to the impedance range.
• Challenge the load range. A dummy resistor verifies one point. Repeat across the complex loads and frequencies that the antenna or tuner can present.
• Map the installed currents. Measure both antenna branches and every accessible exterior return path. A bench result is not the final installed current distribution.
• Run electrical and thermal tests separately. Record accepted power, waveform, duty cycle, voltage/current stress, temperature and drift; stop for rapid heating, arcing or unstable behaviour.

Keysight's balanced-measurement framework is useful here because it treats differential and common signals as distinct modal quantities derived from calibrated complex port data. That is the discipline a balun test needs. A two-port insertion trace, a common-mode impedance curve or an output-balance test can each answer one question; none answers all of them together.

A Guanella drawing tells me how the lines are connected. The measurements tell me whether the finished device transforms the intended load, preserves the intended differential current and blocks the unwanted common-mode path.

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