Why Not Use a Galvanically Isolated Transformer for HF?

Galvanically isolated transformers look like an obvious solution whenever two circuits need to exchange power or signals without sharing a direct electrical connection. They block DC, break low-frequency ground loops, allow voltage conversion, and can improve safety.

So it is natural to ask: why not just use a galvanically isolated transformer for HF?

The answer is not “never use one.” The better answer is this: at high frequency, a transformer is no longer close to ideal, and galvanic isolation does not mean RF isolation.

In amateur radio, HF normally means the 3 to 30 MHz radio spectrum. In power electronics, people may use “high frequency” for switching systems in the tens of kilohertz to several megahertz. In both worlds, the same practical problem appears: parasitic capacitance, leakage inductance, winding loss, core loss, and self-resonance start to dominate the design.

The Misunderstanding: Galvanic Isolation Is Not Infinite Isolation

A galvanically isolated transformer has no direct conductive path between the primary and secondary windings. That is useful for blocking DC and low-frequency ground-loop currents.

But the windings still sit physically close to each other. Two conductors separated by insulation form a capacitor. Therefore every real transformer has primary-to-secondary capacitance, often called interwinding capacitance.

At low frequency, that capacitance may look insignificant. At HF, it becomes a real coupling path.

The capacitive reactance is:

XC = 1 / (2πfC)

For example, with only 10 pF of interwinding capacitance:

That means a transformer can be perfectly isolated in the DC sense while still passing common-mode HF energy through its parasitic capacitance. This is why common-mode rejection generally becomes harder as frequency rises.

At HF, the Transformer Becomes a Network of Parasitics

An ideal transformer has only magnetic coupling and a turns ratio. A real HF transformer behaves more like a small RF network made from many unwanted but unavoidable elements:

• Primary winding resistance
• Primary leakage inductance
• Magnetizing inductance
• Core loss
• Interwinding capacitance
• Winding-to-core capacitance
• Secondary leakage inductance
• Secondary winding resistance
• Package and PCB parasitics

At HF, these are not small errors in the margin. They can define the whole behavior of the circuit.

This is why an RF transformer, an audio transformer, a mains isolation transformer, and a high-frequency DC-DC converter transformer are not interchangeable parts. They may all be called transformers, but they are optimized for completely different operating conditions.

Leakage Inductance Limits Bandwidth

A transformer transfers energy through mutual magnetic flux. In a real transformer, some of the magnetic flux does not link both windings. That uncoupled part appears as leakage inductance.

At low frequency, leakage inductance may be tolerable. At HF, its reactance rises quickly:

XL = 2πfL

A leakage inductance of only 100 nH has a reactance of about 18.8 Ω at 30 MHz. In a 50 Ω RF system, that is already a large error. It can cause mismatch, insertion loss, poor return loss, ringing, unwanted phase shift, and waveform distortion.

This is one reason ordinary mains isolation transformers are not used for HF signal transfer. They are designed for 50 or 60 Hz power isolation, not for controlled impedance, low leakage, low capacitance, and predictable behavior in the MHz range.

Interwinding Capacitance Creates Common-Mode Noise

In high-frequency power electronics, especially with fast SiC or GaN switching, the voltage on one side of the transformer can move extremely quickly. The interwinding capacitance then injects common-mode current into the isolated side:

i = C × dv/dt

This is why isolated power supplies can still be noisy. The DC path is broken, but displacement current still flows through capacitance.

The same principle matters in RF systems. A transformer may stop DC current, but the primary and secondary can still be coupled together by capacitance at HF. That capacitive path can carry RF noise, common-mode current, or unwanted feedback.

Reducing One Parasitic Often Worsens Another

A major difficulty is that transformer parasitics are interconnected. You can reduce leakage inductance by placing windings closer together or by interleaving them. But that usually increases interwinding capacitance.

You can reduce capacitance by separating windings, adding insulation, or using split-bobbin construction. But that usually increases leakage inductance.

This tradeoff is central to HF transformer design. A good RF transformer is not just an isolated transformer made smaller. It is a carefully controlled compromise between coupling, capacitance, leakage inductance, loss, impedance, voltage rating, and bandwidth.

Core Loss and Copper Loss Rise at HF

At HF, the transformer core is not just a passive magnetic guide. Magnetic materials have frequency-dependent losses. If the wrong core material is used, the transformer may heat, lose efficiency, distort the waveform, or stop behaving like a useful transformer.

The windings also suffer. Due to skin effect and proximity effect, current crowds into smaller regions of the conductor as frequency increases. That raises the effective AC resistance.

More turns may increase inductance, but they also add capacitance and copper loss. Fewer turns may reduce capacitance and resistance, but they may not provide enough magnetizing inductance and may push the core closer to saturation.

This is why RF transformers often use ferrite cores, powdered-iron cores, transmission-line structures, bifilar or trifilar windings, or even air-core construction depending on the frequency, power level, impedance, and bandwidth.

Self-Resonance Can Make the Transformer Unpredictable

Leakage inductance and parasitic capacitance form resonant circuits. Near or above self-resonance, a transformer may stop behaving like a transformer. It may attenuate the signal, peak unexpectedly, shift phase, radiate, or couple noise into places where it should not.

A transformer intended for 100 kHz may be unusable at 10 MHz. A transformer that works well at 14 MHz may not work well at 30 MHz. A transformer that works for a narrowband matched RF application may fail badly with broadband pulses or fast digital edges.

This is why useful HF transformer datasheets specify more than isolation voltage and turns ratio. They should also specify frequency range, impedance, insertion loss, return loss, interwinding capacitance, leakage inductance, power rating, and sometimes phase balance or common-mode rejection.

Safety Isolation and HF Performance Fight Each Other

Safety isolation requires physical spacing, insulation thickness, creepage, clearance, and dielectric strength. HF performance often wants compact geometry, tight coupling, controlled impedance, and low parasitic inductance.

Those goals conflict.

• Increasing the distance between windings improves insulation and often lowers capacitance, but it increases leakage inductance.
• Interleaving windings improves magnetic coupling, but it puts conductors closer together and raises capacitance.
• Adding an electrostatic shield can reduce direct primary-to-secondary capacitive coupling, but the shield itself adds capacitance and must be grounded correctly.

So a transformer can be safe but poor at HF, or excellent at HF but not suitable for safety isolation. A proper design must satisfy both sets of requirements.

Why This Matters in RF Systems

In HF radio circuits, transformers are used everywhere: baluns, ununs, impedance transformers, receive transformers, mixers, splitters, couplers, and DC-blocking signal interfaces.

But these are RF transformers, not generic galvanic isolation transformers.

Using the wrong isolated transformer in an HF RF path can cause:

• Poor impedance matching
• Loss of transmit power
• Unexpected heating
• Distorted modulation
• Common-mode current on coax shields
• RF feedback
• Worse noise pickup
• Receiver desensitization
• Failure at higher power

Also, the transformer may isolate DC while still allowing RF common-mode current to pass through capacitance. That means it may not solve the ground-noise, receive-noise, or RF-in-the-shack problem the designer hoped to fix.

Why This Matters in Switching Power Supplies

In high-frequency isolated DC-DC converters, transformers are common and often necessary. But they are carefully designed parts of the converter, not generic isolation blocks.

The switching frequency, duty cycle, topology, insulation system, leakage inductance, interwinding capacitance, core material, winding arrangement, PCB layout, and EMI path are all part of the design.

Fast switching makes the problem harder because common-mode current is proportional to capacitance and voltage slew rate. A small capacitance can still inject a lot of noise when dv/dt is high.

So Why Not Use One?

You may avoid a generic galvanically isolated transformer for HF when these effects matter:

When a Galvanically Isolated Transformer Is Still the Right Choice

A transformer can absolutely be the right HF solution when it is designed for the job.

Good examples include:

• RF baluns
• RF impedance transformers
• Receive transformers
• Pulse transformers
• Wideband RF transformers
• Isolated gate-driver bias transformers
• Flyback converter transformers
• LLC converter transformers
• Ethernet magnetics
• Measurement isolation transformers

The important condition is that the transformer must be chosen or designed for the actual job, not selected only because it is “isolated.”

For HF work, the relevant questions are:

• What frequency range must it cover?
• What impedance level must it work into?
• What power level must it handle?
• How much insertion loss is acceptable?
• How much phase shift is acceptable?
• What isolation voltage is required?
• How much interwinding capacitance is acceptable?
• What leakage inductance can the circuit tolerate?
• What common-mode current or EMI limit must be met?

Conclusion

A galvanically isolated transformer is not automatically unsuitable for HF. In fact, transformers are used constantly in high-frequency RF and power-electronics systems.

The mistake is assuming that galvanic isolation equals perfect high-frequency isolation.

At HF, a transformer is no longer just a magnetic device. It is also a capacitor, an inductor, a lossy core, a resonator, an antenna-like structure, and a layout-sensitive component. If those effects are ignored, the transformer can pass noise, distort signals, radiate EMI, overheat, or fail to provide the isolation performance expected.

Interested in more technical content? Subscribe to our updates for deep-dive RF articles and lab notes.

Questions or experiences to share? Feel free to contact RF.Guru.