Ferrite Cores for Phase-Shifted Full-Bridge Converters

The phase-shifted full-bridge converter achieves zero-voltage switching by using the transformer leakage inductance and the output filter inductor as part of its transition, rather than adding a discrete resonant tank. That makes the transformer the central magnetic component, and the core is usually ungapped, because leakage inductance is a design parameter to be controlled rather than removed.
The phase-shifted full-bridge is a workhorse topology for isolated DC-DC conversion at medium to high power. It achieves soft switching without the resonant tank that an LLC converter needs, which keeps the design familiar to engineers who want a fixed-frequency converter with predictable behavior. The price is paid in the magnetics.
Where an LLC converter tunes a resonant network, the phase-shifted full bridge relies on the transformer leakage inductance and the reflected output inductor to shape the switching transition. That places the transformer at the centre of the design rather than merely passing power through it.
How the topology works
Four switches form the bridge, driving the transformer primary. Instead of switching two diagonal pairs on and off simultaneously, the phase-shifted scheme overlaps the conduction intervals so that one leg leads and the other lags. The lagging leg turns on while its switch voltage has already been driven to zero by the energy circulating in the leakage inductance, which produces zero-voltage switching on that leg.
The resonant transition therefore depends on having a specific amount of inductance in the primary path. Too little and the zero-voltage transition cannot complete before the switch turns on; too much and the circulating current costs conduction loss and reduces the effective duty cycle at the output.
Why the transformer core is usually ungapped
Because the transition uses the natural leakage inductance of the transformer, there is no need for a deliberate resonant inductor in most designs — and no need to reduce the magnetizing inductance with a gap. The core is typically used ungapped or lightly gapped, depending on the design, because the transformer is transferring energy rather than storing it.
- Energy transfer, not energy storage, is the primary function
- Leakage inductance is a design parameter, deliberately set and controlled
- High permeability keeps magnetizing current low
- Low core loss suits the elevated switching frequency
- An ungapped core avoids fringing flux near the windings
Leakage inductance as a design parameter
Leakage inductance is normally an unwanted parasitic. In this topology it becomes the element that enables the soft-switching transition, so it has to be set within a window: high enough to achieve zero-voltage switching across the intended load range, low enough to keep the circulating current and the duty-cycle loss acceptable.
Because leakage inductance depends on winding arrangement as much as on core geometry, this becomes a winding design problem. Primary and secondary spacing, interleaving and the shape of the winding window all influence the value. It also means the tolerance on leakage inductance matters as much as the tolerance on the turns ratio.
Duty-cycle loss
During the resonant transition, the primary current circulates without transferring power to the secondary. That interval consumes a portion of each switching period, so the effective duty cycle applied to the output is less than the duty cycle commanded by the controller. The effect worsens at heavy load and low input voltage, where the required duty cycle is highest.
Designers compensate by selecting the turns ratio with some margin, which in turn affects the voltage stress on the secondary rectifiers and the copper loss in the windings. The transformer design and the overall converter design are therefore settled together rather than sequentially.
Flux balance and saturation
In a well-designed full bridge the transformer sees symmetrical volt-seconds, so flux returns to the same point each cycle and the core does not drift toward saturation. Practical designs still account for imbalance from component tolerance and transient conditions, which is one reason the core is sized with some margin below its saturation limit rather than operating at the edge.
Where the drive is not perfectly balanced — and during startup or load transients — a flux offset can appear. Margin below saturation is what keeps the transformer out of trouble under those conditions.
Engineering requirements to define
- Input voltage range and output voltage and current
- Switching frequency and the intended load range
- Target leakage inductance window for zero-voltage switching
- Magnetizing inductance and acceptable magnetizing current
- Turns ratio with margin for duty-cycle loss
- Core loss at the operating frequency and flux swing
- Temperature rise at worst-case ambient and load
- Isolation and creepage requirements
Design considerations
Winding arrangement determines both leakage inductance and the proximity-effect loss in the copper, and the two pull in opposite directions. Interleaving primary and secondary reduces leakage inductance and improves coupling, but it can increase interwinding capacitance and alter the switching behavior. Separating the windings increases leakage inductance, which helps the zero-voltage transition but raises copper loss.
Core shape follows from the same trade. A shape with a large winding window allows the multi-turn windings and insulation the design needs, while a shape with good thermal coupling to the board helps remove the heat that the elevated frequency generates.
Application examples
- Telecom and server power supplies at medium to high power
- Industrial DC-DC converters requiring isolation
- Battery chargers with wide output voltage range
- Welding and plating supplies where high current is required
- Auxiliary converters inside larger power systems
Selection checklist
- Establish the leakage inductance window before selecting a core
- Confirm the core is sized with margin below saturation for transients
- Evaluate core loss at the real frequency and flux swing
- Balance leakage inductance against winding loss when arranging layers
- Check duty-cycle loss across the full input voltage range
- Verify temperature rise at worst-case load and ambient
Custom development for full-bridge transformers
A phase-shifted full-bridge transformer is defined by its leakage inductance and its loss profile as much as by its turns ratio. That makes it an application-specific component: development covers material grade, core shape, winding arrangement to set leakage, insulation design, sampling and qualification.
See Ferrite Cores for LLC Resonant Converters for the resonant alternative, Ferrite Cores for High-Frequency Transformers for material grade selection, and the Ferrite Core Guide for available shapes.
Frequently Asked Questions
What is the difference between an LLC converter and a phase-shifted full bridge?
An LLC converter regulates by varying switching frequency around a resonant tank, while a phase-shifted full bridge runs at fixed frequency and shifts the timing between the two bridge legs. The full bridge uses transformer leakage inductance and the output inductor to achieve soft switching instead of a dedicated resonant network.
Why is the transformer in a phase-shifted full bridge usually ungapped?
The transformer transfers energy rather than storing it, so there is no need to reduce magnetizing inductance with a gap. Leakage inductance, which is a winding property, provides the inductance needed for the zero-voltage switching transition.
What causes duty-cycle loss in a phase-shifted full bridge?
During the resonant transition the primary current circulates without delivering power to the secondary. That interval consumes part of each switching period, so the effective output duty cycle is lower than the commanded value, with the effect worsening at heavy load and low input voltage.
How important is leakage inductance tolerance?
It matters as much as turns ratio tolerance, because the zero-voltage switching condition depends on it. Too little inductance prevents the transition from completing; too much raises circulating current and duty-cycle loss, so the value has to be held within a defined window.
What should be specified when sourcing a phase-shifted full-bridge transformer?
Input and output voltage range, power level, switching frequency, target leakage and magnetizing inductance with tolerances, turns ratio with duty-cycle margin, core loss at the operating point, temperature rise, and isolation and creepage requirements.
Related pages
Need help choosing the right core?
Send the operating point and the mechanical envelope — we will come back with a core and a winding proposal.