A flyback converter is a type of DC–DC converter that stores energy in a magnetic component while a switch conducts and delivers that energy to an output after the switch turns off. Commonly used in switched-mode power supplies, it provides voltage conversion and, when its windings are appropriately insulated, electrical isolation between input and output. Its relatively small component count makes it widely used in adapters, chargers, auxiliary supplies, and low- to moderate-power industrial equipment.
Circuit and operating principle
The basic flyback circuit contains a controlled semiconductor switch, a magnetic component with primary and secondary windings, an output rectifier, and an output capacitor. The switch is usually a metal–oxide–semiconductor field-effect transistor (MOSFET). An input capacitor, control circuit, protection components, and circuitry to limit switching-voltage spikes are normally also required.
Although commonly called a flyback transformer, the magnetic component operates as a coupled inductor: it intentionally stores substantial energy in its magnetizing inductance. This distinguishes it from an ideal transformer, which transfers energy between windings without acting as a net energy-storage element. The flyback topology is closely related to the buck–boost converter, with the energy-storage inductor divided into coupled windings.
Switch-on interval
When the primary switch turns on, the input voltage is applied across the primary winding. The primary current increases approximately linearly:
where is the input voltage and is the magnetizing inductance referred to the primary.
The winding polarity makes the secondary rectifier reverse-biased during this interval. Consequently, the input supplies energy to the magnetic component rather than directly to the output. The output capacitor supplies the load.
The stored magnetic energy is
where is the primary-referred magnetizing current.
Switch-off interval
When the switch turns off, the winding voltages reverse. The secondary rectifier becomes forward-biased, and the stored energy flows through the secondary winding into the output capacitor and load.
The magnetizing current cannot change instantaneously. Instead, current commutates from the primary circuit to the secondary circuit, with its physical magnitude transformed by the turns ratio. During energy delivery, the primary-referred magnetizing current decreases approximately as
where and are the primary and secondary turn counts, is the output voltage, and is the rectifier’s forward-voltage drop.
Primary and secondary conduction therefore occur in separate intervals in the basic circuit. This separation, rather than the mere presence of a transformer, defines flyback operation.
Conduction modes and voltage conversion
A flyback converter can operate in several conduction modes. These describe whether the stored magnetizing energy is fully discharged before the next switching cycle.
Continuous conduction mode
In continuous conduction mode (CCM), the magnetizing current remains above zero throughout the switching cycle. Although current in each individual winding is interrupted, the equivalent magnetizing current remains continuous as conduction alternates between windings.
For an ideal converter in steady-state CCM, the net change in magnetic flux over a cycle is zero. Volt-second balance gives
and hence
where is the fraction of the switching period for which the primary switch is on.
The turns ratio and duty cycle together determine the ideal conversion ratio. Losses and device voltage drops modify this relationship in practical circuits.
CCM generally reduces peak current for a given power level compared with discontinuous operation under comparable conditions. However, it introduces more demanding rectifier commutation and control-loop behavior.
Discontinuous conduction mode
In discontinuous conduction mode (DCM), the magnetizing current falls to zero before the next switch-on event. The cycle contains three intervals: primary charging, secondary energy delivery, and an idle interval.
Because each cycle begins with zero magnetizing current, the peak primary current is approximately
where is the switching frequency. For complete discharge each cycle, output power can be approximated by
where represents conversion efficiency.
Unlike the ideal CCM relationship, DCM voltage conversion also depends on load, inductance, and switching frequency. DCM simplifies some aspects of control and reduces secondary rectifier reverse-recovery effects, but usually requires higher peak currents.
Boundary and quasi-resonant operation
At the boundary between CCM and DCM, the next charging interval begins just as magnetizing current reaches zero. Controllers using this approach typically vary the switching frequency with input voltage and load.
A quasi-resonant flyback uses the oscillation between the magnetic component and circuit capacitances after secondary conduction ends. Turning on near a valley in the switch voltage can reduce switching loss and electromagnetic emissions. Valley switching does not necessarily mean that the switch voltage reaches zero.
Magnetic component and electrical isolation
The magnetic component is central to flyback design. Its inductance, turns ratio, leakage inductance, insulation, winding resistance, and parasitic capacitance all affect performance.
Energy storage and core design
Flyback magnetics commonly use a ferrite core with a deliberate air gap. The gap allows substantial magnetic energy storage while limiting flux density for a given magnetizing current. Much of the energy resides in the gap’s magnetic field.
The approximate flux-density excursion during the switch-on interval is
where is the effective core cross-sectional area. In CCM, peak flux density also depends on the nonzero initial magnetizing current.
Excessive flux density causes core saturation, sharply reducing effective inductance and producing a rapid current rise. Core geometry, gap length, turn count, current limits, and operating frequency must therefore be considered together. Higher frequency can reduce magnetic component size, but increases switching losses and can increase core and winding losses.
Leakage inductance and parasitic coupling
Not all magnetic flux links both windings. The resulting leakage inductance stores energy that does not transfer through the ideal transformer action. When the primary switch opens, this energy can produce a substantial voltage spike.
Winding arrangements that reduce leakage inductance often increase interwinding capacitance. This creates a trade-off: lower leakage can reduce voltage spikes, while higher capacitance increases high-frequency common-mode current across the isolation barrier.
Isolation and multiple outputs
Separate windings can provide galvanic isolation, but safety depends on the insulation system, winding construction, physical spacing, and applicable requirements—not simply on the existence of separate windings.
Additional secondary windings allow multiple outputs with comparatively few extra components. Their voltages are approximately related by turn ratios, but unequal rectifier drops, leakage inductance, and load variations produce cross-regulation errors. A feedback loop sensing one output does not necessarily regulate the others accurately.
Switching stress, losses, and interference
In the basic single-switch topology, the MOSFET must withstand both the input voltage and the output voltage reflected to the primary. Ignoring leakage-induced overshoot,
The secondary rectifier’s reverse-voltage stress is approximately
before allowance for ringing and other transients.
These relationships make the turns ratio a compromise among switch voltage rating, rectifier voltage rating, duty cycle, and winding currents.
Practical losses include:
- Conduction losses in the switch, rectifier, and windings.
- Switching losses during voltage and current transitions and during charging and discharging of device capacitances.
- Magnetic losses, including core loss and frequency-dependent winding loss.
- Clamp and snubber losses associated with leakage energy and ringing.
- Capacitor losses caused by ripple current and equivalent series resistance.
An RC or RCD snubber can damp ringing or limit peak voltage. RCD clamps generally dissipate captured energy as heat. Active-clamp arrangements can recover some energy and enable lower-loss switching under suitable operating conditions.
Both input and output currents are pulsating. Fast voltage transitions and compact, high-current switching loops also generate electromagnetic interference (EMI). Filtering, shielding, winding construction, and circuit-board layout therefore strongly influence practical performance.
Regulation and control
Flyback controllers regulate output voltage or current by changing duty cycle, switching frequency, peak current, or the number of active switching cycles.
Common methods include fixed-frequency pulse-width modulation, peak-current-mode control, boundary-mode control, and burst operation at light load. These are applications of control theory and negative feedback.
In an isolated supply, output information can cross the isolation barrier through an optocoupler or another isolated signaling device. Primary-side regulation instead estimates output conditions from primary-side measurements, often using an auxiliary winding. It can reduce component count, but accuracy depends on sampling conditions and compensation for circuit nonidealities.
Two control effects are particularly important:
- Right-half-plane zero: In CCM, increasing duty cycle initially shortens the energy-delivery interval, even though it eventually increases transferred power. This produces a right-half-plane zero in the control-to-output response and limits practical feedback bandwidth.
- Subharmonic oscillation: Fixed-frequency peak-current-mode operation in CCM can become unstable at high duty ratios without sufficient slope compensation. The required compensation depends on the current-waveform slopes and controller design.
Protection functions commonly include cycle-by-cycle current limiting, output-overvoltage protection, thermal shutdown, and controlled startup. Short-circuit behavior depends on both the controller and the power stage; current limiting alone does not establish that every component remains within its ratings.
Variants
Several variants retain the flyback energy-storage principle:
- Synchronous-rectified flyback: Replaces the secondary diode with a controlled MOSFET to reduce conduction loss, especially at low output voltages. Timing must prevent unwanted reverse current.
- Active-clamp flyback: Uses an auxiliary switch and capacitor to manage leakage energy and support soft switching.
- Two-switch flyback: Uses two primary switches and clamp diodes to control device voltage stress and return leakage energy to the input.
- Interleaved flyback: Operates multiple flyback stages with phase-shifted switching to distribute losses and reduce aggregate ripple.
- Bidirectional flyback: Uses controlled switches on both sides to allow energy transfer in either direction.
Gallium nitride switches can support higher-frequency and higher-power-density implementations, although their benefits depend on the magnetic design, switching strategy, parasitic effects, and thermal constraints.
Applications and limitations
Flyback converters are widely used in mains-powered adapters, appliance electronics, lighting drivers, industrial control supplies, and auxiliary supplies within larger power-electronic systems. In an AC-powered product, an input rectifier first creates a DC supply; the flyback stage itself performs DC–DC conversion.
Related energy-storage and high-voltage pulse-generation principles were used in cathode-ray-tube line-output circuits. The name “flyback” refers to the rapid retrace interval in those systems. Photographic flash charging circuits also use flyback conversion to charge a capacitor to a high voltage.
The topology is attractive when low component count, a broad input range, isolation, or multiple outputs are important. Its limitations become more significant as power rises: high peak and RMS currents increase losses, leakage energy increases clamp stress, and substantial ripple requires filtering. Higher-power implementations are possible, but forward, bridge, and resonant converter topologies may offer better utilization of magnetic components and semiconductor devices. There is no universal power boundary; the preferred topology depends on voltage range, efficiency, cooling, isolation, size, and cost requirements.