Access Official Technical PDF

Share Summary:LinkedInEmail

Power switch selection shapes a converter’s efficiency, thermal behavior, size and cost. Silicon IGBTs and MOSFETs remain important options, while silicon carbide (SiC) MOSFETs and gallium nitride (GaN) high-electron-mobility transistors (HEMTs) extend the choices available to design engineers. No technology is the best fit for every converter: the right choice depends on operating voltage and current, switching frequency, topology and the system’s lifecycle requirements.

For industrial automation teams, this decision affects more than the semiconductor itself. Switching losses influence cooling requirements, while device behavior during commutation can affect performance in motor drives, inverters and other power-conversion stages. Evaluating the application as a whole is more useful than selecting a device based on a technology label alone.

Power switch selection starts with the application

Power converters use four main switch types: silicon IGBTs, silicon MOSFETs, SiC MOSFETs and GaN HEMTs. Hybrid arrangements are also possible. For example, an IGBT may be combined with a parallel SiC Schottky barrier diode, or IGBT and SiC switches may be used at different points in a multilevel topology such as a neutral-point-clamped (NPC) or active neutral-point-clamped (ANPC) converter.

Voltage and current ratings help narrow the options, but they do not settle the decision. GaN is associated with applications around a 650 V blocking voltage, while SiC is often considered for higher-voltage designs, particularly above 1,000 V. These are broad application tendencies rather than substitutes for reviewing actual device ratings and circuit conditions. IGBTs and silicon MOSFETs can remain more economical or perform better in particular operating ranges.

Topology matters, too. A new design can be optimized around a wide-bandgap device, but directly replacing an IGBT or silicon MOSFET with GaN and adjusting the gate drive may fail to deliver benefits and could increase electromagnetic interference (EMI). In some existing topologies, replacing silicon IGBTs with SiC can reduce losses without changing switching frequency or turn-on and turn-off times.

Commutation and third-quadrant behavior

In half-bridge circuits, current may flow in the opposite direction from the normal blocking and conduction direction during commutation. This is often described as third-quadrant operation. It is relevant in circuits including inverters, motor drives, totem-pole power-factor-correction stages, LLC converters, phase-shifted bridges and active-clamp flyback converters.

Device behavior in this mode affects losses. Silicon and SiC MOSFETs have intrinsic body diodes. A silicon MOSFET’s body diode can have significant reverse-recovery charge, which may increase losses, especially at higher frequencies. A SiC MOSFET has negligible reverse-recovery charge, but its diode’s forward-voltage drop can still dissipate energy. IGBTs do not have an intrinsic diode; their turn-off current tail can contribute to switching loss.

GaN conducts in the third quadrant through its channel, with a voltage drop related to channel resistance and no reverse-recovery charge. This can help reduce losses in suitable circuits. Designers may also use synchronous rectification—actively controlling switches in place of relying on diodes—to reduce static losses. The most suitable approach depends on the device, control strategy and operating conditions.

Balance efficiency, switching frequency and power density

Higher switching frequency can reduce the size of passive components, making it one route to greater power density. GaN can support high-frequency designs, while Si or SiC MOSFETs operated at lower frequencies may be preferable when somewhat larger passive components are acceptable. In soft or resonant switching, where device voltage or current is zero at the switching instant, superjunction silicon MOSFETs can remain a strong option up to about 250 kHz. In hard-switched half-bridge applications, SiC and GaN can be attractive for efficiency optimization.

Efficiency also affects thermal design. More efficient conversion can reduce heat generation and may allow smaller heatsinks, but surrounding components and the complete thermal path still require careful design. Control strategy and topology influence the result: hard and soft switching impose different conditions on the power devices and their losses.

Power density is not the same as semiconductor performance in isolation. It reflects how the converter’s switches, passive components, cooling and layout work together. Engineers should compare candidate designs against the application’s volume, temperature-rise and operating requirements rather than assuming that the highest switching frequency will produce the best overall result.

Use figures of merit and total cost to compare options

Device figures of merit can help organize a comparison. The product of on-resistance and die area, RDS(on) × A, indicates how much die area is needed for a given on-resistance and voltage rating. A lower value can support a smaller die, with associated implications for capacitance, switching speed and potential device cost.

Another comparison combines on-resistance with output-capacitance energy, often written as RDS(on) × EOSS. EOSS represents energy stored in the device’s output capacitance; that energy is dissipated during switching. Gate charge also matters because gate-drive power depends on charge per cycle, switching frequency and gate-voltage swing. A simplified relationship is PGSW = QG(TOT) × F × VGSW. Thus, increasing frequency can raise gate-drive power, and gate resistance and capacitance can affect switching speed and EMI.

Finally, compare total cost of ownership (TCO), not just the purchase price of a switch. The broader calculation can include energy lost through inefficiency, design and qualification effort, and the cost of replacing obsolete equipment. In applications with substantial energy use, efficiency improvements may influence payback; the outcome depends on factors such as electricity costs and system lifetime.

Make the choice at system level

There is no universal winner among silicon IGBTs, silicon MOSFETs, SiC and GaN. Start with the converter’s voltage and current requirements, topology, switching mode and commutation behavior. Then compare conduction, switching and gate-drive losses alongside thermal constraints, power-density goals and lifecycle cost.

Wide-bandgap devices can deliver important advantages when a design is built around their characteristics. Silicon technologies, meanwhile, remain appropriate in many applications and operating ranges. A system-level comparison—rather than a direct device-for-device substitution—gives engineering and operations teams a more reliable basis for choosing a power switch.

Access Official Technical PDF

About The Author