Modern power electronics depend on fast switching to convert and control electrical energy, yet almost every converter stage still contains a magnetic component that transforms voltage, stores energy or isolates the control side from the power side. In these circuits the transformer is judged less on raw kVA and more on core loss, leakage inductance and thermal behaviour at the switching frequency.
Where Transformers Fit in Power Electronics
A converter chain normally runs through several conversion steps: rectification on the input, a switched stage that creates a high-frequency square wave, a magnetic stage that scales and isolates that waveform, and a final rectifier and filter on the output. The transformer in the middle is what makes compact, efficient conversion possible, because raising the operating frequency shrinks the core for a given power rating.
Power Conversion in Switched-Mode Stages
In a switched-mode supply the primary winding is driven by a high-frequency bridge or half bridge, and the turns ratio sets the secondary voltage. Flyback, forward, push-pull, half-bridge and full-bridge topologies all use the same underlying principle, but they load the magnetic component differently. Flyback designs use the winding as an energy store and need an air gap to control inductance, while forward and bridge designs transfer energy directly and aim for low leakage inductance.
Flyback: gapped core, energy storage, moderate power, low component count.
Forward and push-pull: direct energy transfer, needs reset of the core, higher output power.
Half-bridge and full-bridge: highest power levels, core driven in both quadrants.
Resonant converters: sinusoidal current shapes, so winding losses and core excitation need separate treatment.
Galvanic Isolation, Noise Rejection and Operator Safety
Isolation is often the reason a transformer is specified at all. By separating primary and secondary circuits it blocks common-mode current, keeps fault energy inside the supply, and allows a safe extra-low-voltage secondary to feed control electronics and communication interfaces. Creepage, clearance and insulation coordination are set by the applicable safety standards for the end product, and they constrain winding layout and the physical gap between windings.
Choosing Core Material by Frequency
Material selection follows the operating frequency. Silicon steel remains the economic choice at line and low switching frequencies; ferrite dominates from roughly 20 kHz upward; amorphous and nanocrystalline ribbons fill the band where both loss and saturation flux density matter.
| Material | Typical frequency band | Saturation flux density | Typical role |
|---|---|---|---|
| Grain-oriented silicon steel | 50-400 Hz | About 2.0 T | Line-frequency power and isolation transformers, reactors |
| Thin silicon steel laminations | 400 Hz-10 kHz | About 1.8-2.0 T | Aerospace and traction magnetic components |
| Manganese-zinc ferrite | 20 kHz-several hundred kHz | About 0.4 T | Switched-mode supplies, planar transformers |
| Amorphous or nanocrystalline ribbon | 1 kHz-100 kHz | About 1.2-1.6 T | Low-loss common-mode chokes, compact converters |
The trade-off is clear: ferrite accepts little flux swing but loses very little at high frequency, while silicon steel accepts a large flux swing but loses much more as frequency rises. Because hysteresis loss scales roughly with frequency and eddy current loss with the square of frequency, lamination thickness has to fall as the switching frequency climbs, which is why 0.20-0.35 mm grades are used for 400 Hz magnetic components.
Voltage Regulation, Thermal Design and Testing
Regulation in a power electronic converter is achieved by controlling duty cycle or switching frequency rather than by tapping the winding, but the transformer still shapes the result through its leakage inductance and winding resistance. Thermal design is equally important: core loss and copper loss both appear as heat in a component that is usually the smallest item in the assembly, so loss data measured at the actual excitation level and frequency should be requested. Routine checks include turns ratio, polarity, insulation resistance, and core loss measurement at the specified flux density and frequency.
Solar inverters, wind converters, battery chargers, uninterruptible supplies, rail traction and on-board chargers for electric vehicles all depend on the same magnetic building blocks. Trends toward higher switching frequencies and wider band-gap switching devices push designs toward thinner laminations, ferrite and amorphous materials, and better thermal management, while efficiency regulations continue to squeeze no-load and load losses in every linked component.
Frequently Asked Questions
Q: Why are transformers used in power electronics instead of a simple resistor divider?
A divider cannot change voltage without dissipating the surplus as heat or provide isolation; a transformer converts voltage efficiently and separates the two circuits at the same time.
Q: What determines the size of a switching transformer?
Core size follows the power rating, the flux swing and the switching frequency. Increasing frequency reduces the required flux swing per cycle, so the core can be made smaller for the same power.
Q: When should ferrite replace silicon steel?
Ferrite becomes the better choice above roughly 20 kHz, where its low high-frequency loss outweighs its much lower saturation flux density compared with silicon steel.
Q: Why do flyback transformers have an air gap?
The gap raises the reluctance of the magnetic path, which sets the inductance needed to store energy during the on-time and prevents the core from saturating under peak current.
Q: How is isolation measured in a magnetic component?
The design must satisfy the creepage, clearance and dielectric strength requirements of the end-product safety standard, and production units are hi-pot tested at the specified voltage.
Q: What loss data should be requested from a core supplier?
Specific core loss at the actual flux density and frequency of the application, plus stacking factor, coating type and thickness tolerance, so the design margin is based on real operating conditions.