A high frequency transformer works on the same electromagnetic principle as any other transformer, transferring energy between circuits through a changing magnetic field, but it operates at frequencies far above the standard 50 or 60 Hz of mains power. Typical operating ranges start around 20 kHz and extend well into the megahertz region, and that jump in frequency changes almost everything about how the device is built, what materials it uses, how it fails, and where it excels. The payoff for all that redesign is dramatic size reduction: a transformer handling the same power at 100 kHz can be a fraction of the volume of its 60 Hz counterpart, which is why high frequency transformers are found inside laptop chargers, EV fast-chargers, solar inverters, and countless other systems where compactness and efficiency matter.
Why Frequency Changes Everything About Size
The fundamental relationship behind this technology is straightforward. The amount of energy a transformer’s core can handle per cycle is limited by the core’s cross-sectional area and the number of turns in the winding. At low frequencies, each cycle is long, so the core has to be large enough to store energy for that entire cycle without saturating. Raise the frequency and each cycle gets shorter, meaning the core can be smaller and still keep up. A transformer designed for 100 kHz goes through roughly 1,700 times as many cycles per second as one designed for 60 Hz, so the core and windings shrink proportionally. In practical terms, a power supply that once needed a fist-sized iron-core transformer can use a high frequency design smaller than a matchbox.
This is the single biggest reason high frequency transformers exist. Every watt of power that needs galvanic isolation, meaning a physical separation between input and output circuits for safety, requires a transformer. Making that transformer smaller, lighter, and more efficient has been one of the central engineering challenges in power electronics for decades. The shift from heavy, low-frequency “wall wart” adapters to the slim USB-C chargers in use today is largely a story about pushing transformer operating frequencies higher.
Core Materials and Their Limits
Conventional power transformers use laminated silicon steel for their cores. Steel works well at 50 or 60 Hz, but at higher frequencies it generates excessive losses from eddy currents, tiny loops of electrical current induced inside the core material itself. Those currents waste energy as heat, and they get worse as frequency rises. High frequency transformers solve this by switching to materials with much higher electrical resistivity, which suppresses eddy currents.
Ferrite is the workhorse material for most high frequency applications. It is a ceramic compound, typically manganese-zinc or nickel-zinc based, with very high resistivity and reasonable magnetic performance up to several megahertz. Ferrite cores are inexpensive, widely available in standardized shapes, and well-characterized. Their main limitation is a relatively low saturation flux density compared to metals, which means the core still needs to be large enough to avoid saturation at high power levels.
For applications demanding higher power density, nanocrystalline alloys have become increasingly popular. These are metallic ribbons with grain structures only a few nanometers across, giving them a combination of high saturation flux density and low core losses at elevated frequencies. Nanocrystalline cores can handle more power in a smaller volume than ferrite, though they cost more and are less forgiving of mechanical stress. Amorphous metal alloys occupy a middle ground, offering better saturation than ferrite at moderate cost, and they show up in medium-frequency applications like solid-state transformers for grid use.
The choice of core material cascades into every other design decision. A nanocrystalline core running at 50 kHz might achieve the same power density as a ferrite core at 200 kHz, but the lower-frequency design will have different winding losses and different thermal behavior. Engineers often iterate between core material, frequency, and winding layout multiple times before settling on an optimized design.
Winding Challenges at High Frequency
At mains frequency, a solid copper wire carries current more or less uniformly across its cross-section. At high frequency, that stops being true. Skin effect pushes current toward the outer surface of a conductor, reducing the effective area carrying the load. At 100 kHz, the skin depth in copper is only about 0.2 mm, meaning the center of any wire thicker than that is essentially wasted. At 1 MHz the skin depth drops to about 0.066 mm. A thick solid wire at these frequencies behaves like a hollow tube, carrying current only in a thin shell near the surface.
Proximity effect compounds the problem. When multiple turns sit close together, the magnetic fields from neighboring conductors push current around inside each wire in ways that increase losses further. In a tightly wound multilayer coil, proximity-effect losses can actually exceed skin-effect losses, sometimes by a wide margin.
The standard countermeasure is litz wire, a cable made of many individually insulated fine strands woven or braided together so that each strand spends roughly equal time on the inside and outside of the bundle. This averages out the skin and proximity effects across all strands. The strand diameter needs to be comparable to or smaller than the skin depth at the operating frequency for litz wire to be effective. Selecting the right strand count and diameter is a nontrivial engineering problem, and researchers continue to refine analytical methods for predicting the ac resistance and leakage inductance of litz-wire windings in high frequency transformers.1IEEE Xplore / CrossRef. Accurate Calculation and Sensitivity Analysis of Leakage Inductance of High-Frequency Transformer With Litz Wire Winding
For very high frequencies or very flat form factors, foil windings or printed-circuit-board (PCB) traces replace round wire entirely. Planar transformers, discussed below, take this approach to its logical endpoint.
Parasitic Effects That Shape Performance
Every transformer has parasitic elements: leakage inductance from magnetic flux that does not link both windings, parasitic capacitance between winding layers and between windings and the core, and winding resistance that varies with frequency. At 60 Hz, these parasitics are usually small enough to ignore. At 100 kHz and above, they dominate behavior.
Leakage inductance stores energy that is not transferred to the secondary winding each cycle. That stored energy has to go somewhere, and in a hard-switched converter it typically causes voltage spikes across the switching transistors. In resonant converter topologies, leakage inductance is actually put to work as part of the resonant tank, which is one reason resonant designs are so popular at high frequency. Some transformer designs deliberately engineer a specific amount of leakage inductance into the structure, using controlled spacing between primary and secondary windings or magnetic shunts, to eliminate the need for a separate inductor in the circuit.2IEEE Xplore / CrossRef. Planar Transformer with High Leakage Inductance Integration for Resonant Converters
Parasitic capacitance creates its own headaches. Capacitance between primary and secondary windings provides a path for common-mode noise to pass through the isolation barrier, defeating one of the transformer’s main purposes. Interwinding capacitance within a single winding can cause self-resonance, where the winding behaves as a capacitor rather than an inductor above a certain frequency. Analytical models for these parasitic elements help designers predict and control them before building a prototype.3Journal of Magnetics. Analytical Calculation of Resistance, Leakage Inductance and Parasitic Capacitance for High Frequency Transformer
Insulation and Long-Term Reliability
High frequency operation stresses insulation in ways that low-frequency designs rarely encounter. The voltage waveforms in modern power converters are not smooth sine waves; they are fast-edged rectangular pulses with rise times measured in nanoseconds, particularly when driven by silicon carbide or gallium nitride transistors. Those steep edges create momentary voltage concentrations across thin insulation layers, and the rapid polarity reversals happen tens or hundreds of thousands of times per second.
Partial discharge is the primary long-term failure mechanism. When the electric field across a tiny air gap or void in the insulation exceeds a threshold, a localized electrical breakdown occurs. Each discharge event erodes a microscopic amount of insulating material. Over millions of cycles, this erosion creates a conductive path and the insulation eventually fails. The fast switching edges of modern converters make partial discharge characterization more complex than in traditional line-frequency applications, and researchers have been working to understand how pulse-width-modulated waveforms affect the onset and severity of partial discharge in high frequency transformers.4IEEE Transactions on Power Electronics. Characterization of Partial Discharges in High-Frequency Transformer Under PWM Pulses
For safety-critical applications like medical equipment and EV chargers, regulatory standards require transformers to meet specific insulation ratings. Designers use triple-insulated wire, Kapton tape layers, bobbin barriers, and potting compounds to achieve the required creepage and clearance distances. Overdesigning insulation adds bulk that partly offsets the size advantage of high frequency operation, so there is a constant tension between compactness and reliability.
Thermal Management in a Small Package
Shrinking a transformer does not shrink the heat it produces. Core losses and winding losses both generate heat, and a smaller component has less surface area to dissipate it. Thermal management is often the factor that limits how far a design can push power density.
Core losses in ferrite increase roughly with the square of the frequency and a power of the flux density (the exact relationship depends on the material grade and is described by manufacturers in their datasheets). Winding losses, as discussed, are amplified by skin and proximity effects. The sum of these losses becomes the thermal budget the designer has to manage.
Common cooling strategies include potting the transformer in thermally conductive epoxy, mounting it on a heatsink or cold plate, or designing the winding structure so that hot spots are close to surfaces where heat can escape. In very high-power designs, forced air or even liquid cooling may be used. PCB-integrated planar transformers benefit from the copper planes of the circuit board itself acting as heat spreaders.
Planar Transformers
A planar transformer replaces conventional wound wire with flat copper traces, either etched onto a PCB or stamped from copper sheet, interleaved with core halves that clamp around the board. The result is a transformer with a very low profile, excellent repeatability between units, and the ability to be assembled by standard surface-mount manufacturing equipment.
Planar designs naturally lend themselves to high frequency operation because the flat, wide conductors have a large surface area relative to their cross-section, which partly mitigates skin-effect losses. The consistent geometry also makes parasitic elements highly predictable from unit to unit, a significant advantage in resonant converter designs where the circuit depends on specific inductance and capacitance values.
The trade-off is that planar transformers tend to have limited window area for conductors, which caps the number of turns and the current-handling ability. They work best in low-to-medium power ranges, roughly a few watts up to a few kilowatts, and at frequencies where a small number of turns provides the needed voltage transformation. High-power applications still tend to use wound transformers, though hybrid designs that combine PCB windings with external conductors are gaining traction.
The Role of Wide-Bandgap Semiconductors
The rise of high frequency transformers is inseparable from advances in switching devices. Traditional silicon MOSFETs and IGBTs have practical switching-frequency ceilings, above which their switching losses become prohibitive. Silicon carbide (SiC) and gallium nitride (GaN) transistors can switch far faster with lower losses, enabling converter designs that operate at frequencies where the transformer becomes remarkably small.
GaN transistors in particular have pushed charger frequencies into the hundreds of kilohertz and even low megahertz range, which is why modern laptop and phone chargers are so much smaller than those from a decade ago. The transformer in a 65-watt GaN charger might be smaller than a sugar cube, something unthinkable with older silicon switching technology at the same power level. Researchers continue to co-optimize transformer design alongside wide-bandgap device selection, since the two are deeply interdependent: the switching device determines the frequency, which determines the transformer’s core material, winding approach, and parasitic requirements.5IEEE Xplore / CrossRef. Design and Optimization of High-frequency Transformer for Isolated Single-Stage Three-Phase AC/DC Converter using Bidirectional Switches
Electric Vehicle Charging and Renewable Energy
One of the most active areas for high frequency transformer development is electric vehicle charging. DC fast chargers need to convert grid-level AC power to the high-voltage DC required by an EV battery, and they need galvanic isolation for safety. Doing this at high power levels (50 kW and above for public chargers, with some exceeding 350 kW) while keeping the charger cabinet a reasonable size demands transformers operating well above mains frequency. The dual active bridge (DAB) topology, which uses a high frequency transformer as its central element, has become a preferred architecture for bidirectional EV chargers because it offers soft switching, high efficiency, and natural support for power flow in both directions.
Three-phase versions of the DAB are particularly relevant for high-power chargers because they distribute the current and magnetic flux more evenly, reducing peak stresses on any single component. The transformer design for a three-phase DAB is considerably more complex than a single-phase version, involving multiple core legs and interleaved windings, and remains an active area of research.6IEEE Xplore / CrossRef. Design Analysis of High-Frequency, High Power 3-Phase DAB Transformer for Electric Vehicle Charger Applications
Solar and wind power systems face a similar set of demands. String inverters, microinverters, and power optimizers all use high frequency transformers or coupled inductors to convert and isolate power. As renewable installations scale up and grid-tied energy storage becomes more common, the push toward higher power density and higher efficiency in these converters keeps driving transformer frequency upward.
Solid-State Transformers and the Grid
The most ambitious application of high frequency transformer technology is the solid-state transformer (SST), sometimes called a power electronic transformer. The idea is to replace the large, heavy distribution transformers used by electric utilities with a power-electronics-based system that converts voltage through a high frequency intermediate stage. A typical SST rectifies the incoming AC to DC, switches that DC through a high frequency transformer at tens of kilohertz, and then converts the output back to AC at the desired voltage and frequency.
The appeal is flexibility. An SST can regulate voltage, compensate for power-quality issues, interface directly with DC sources like solar panels and batteries, and enable bidirectional power flow, all things a passive iron-core distribution transformer cannot do. The challenge is that the conversion stages add losses and complexity, and the high frequency transformer at the heart of the system must handle very high voltages and powers with extreme reliability, since utility equipment is expected to last decades with minimal maintenance.
Nanocrystalline and amorphous cores are often favored for SST applications because their higher saturation flux density allows a smaller core at the medium frequencies (typically 5 to 50 kHz) used in these systems. Insulation design is especially demanding because the transformer may need to withstand voltage levels of 10 kV or more while operating at high frequency. Despite years of research and several demonstration projects, SSTs have not yet displaced conventional distribution transformers on a wide scale, mainly because the passive transformer is extraordinarily reliable and inexpensive. The economics may shift as grids become more complex and the need for active voltage regulation and DC integration grows.
Common Misconceptions
One widespread misunderstanding is that high frequency transformers are fundamentally different devices from conventional transformers. They are not. The underlying physics is identical: a time-varying current in one winding creates a time-varying magnetic field in the core, which induces a voltage in the other winding. What changes is the engineering. The materials, geometries, and loss mechanisms that matter at 100 kHz are very different from those at 60 Hz, but the electromagnetic principle is the same.
Another misconception is that higher frequency is always better. In reality, losses in both the core and the windings increase with frequency, and parasitic effects become harder to manage. There is an optimal frequency for any given application, set by the trade-off between size reduction and loss increase. Pushing beyond that optimum makes the transformer smaller but less efficient, and the thermal management challenge may wipe out any size advantage. Many commercial designs land in the 50 to 500 kHz range not because higher frequencies are impossible, but because that range represents the best practical compromise for their power level and topology.
Finally, people sometimes assume that a high frequency transformer can simply replace a low-frequency one in an existing circuit. It cannot. The transformer needs a matching converter topology that generates the high frequency AC waveform, rectifies the output, and manages the energy storage and control loops. The transformer is just one component in a tightly integrated power-conversion system, and swapping it out in isolation does not work.