IGBT vs MOSFET: Side-by-Side Comparison

Read the table from the loss equations: a 40A MOSFET at 80mΩ dissipates 40² × 0.08 = 128W conduction; a 40A IGBT at 1.45V dissipates 40 × 1.45 = 58W. Same current, half the heat.

But at 5A the same MOSFET burns 2W while the IGBT still burns 7.25W. The choice is a current question first, a frequency question second.

What about the frequency side of the equation?

Key Differences

1. Conduction loss: quadratic vs linear

MOSFET loss is I²R — double the current, quadruple the heat. IGBT loss is V_CE(sat) × I — double the current, double the heat. The crossover for 600V silicon devices sits around 20–30A at 25°C, and R_DS(on) doubling at 150°C drags it lower in real boards.

2. The tail current: why IGBTs are slower

The IGBT's low drop comes from minority-carrier injection, but those stored carriers must recombine at turn-off — producing a tail current that delays switch-off and burns extra energy every cycle.

A MOSFET, unipolar, turns off clean in tens of nanoseconds. Above ~20–50kHz, IGBT switching loss dominates and the frequency ceiling decides.

3. Reverse conduction: the missing body diode

A MOSFET's body diode conducts reverse current for free. An IGBT has no built-in reverse path — in bridge circuits you must add an antiparallel diode per device, or the freewheeling current has nowhere to go. That's extra parts and extra layout care in every IGBT bridge design.

4. Robustness: short-circuit and linear region

IGBTs typically withstand short circuits for 5–10µs, giving protection circuits time to act; MOSFETs need faster response.

But the reverse is true in the linear region: an IGBT in a dummy-load application runs away thermally, while linear-rated MOSFETs survive. Know which abuse your design will throw at the switch.

When to Choose IGBT / When to Choose MOSFET

Choose the IGBT when:

• Your average current sits above ~20–30A at 650V. The linear V_CE(sat) beats the quadratic I²R. Motor drives, industrial inverters, UPS, and welding supplies live here.
• Switching frequency stays under ~20kHz. Below that, the tail-current penalty is small and conduction dominates. A 10kHz motor drive is textbook IGBT territory.
• You need short-circuit robustness. The 5–10µs withstand window gives protection circuits room to act before the die dies.
• Cost per amp matters at high power. The die area needed for low R_DS(on) at 650V gets expensive; an IGBT delivers the same current from a smaller, cheaper die.

Choose the MOSFET when:

• Your current is low to moderate. Below the crossover, the millivolt-level drop beats any V_CE(sat). A 5A, 650V PFC or flyback stage is MOSFET country.
• Frequency is high. Above ~20–50kHz the tail current makes IGBTs inefficient; superjunction MOSFETs switch clean to hundreds of kHz and beyond.
• You want the body diode for free. Half-bridge and synchronous rectifier stages get reverse conduction without adding a diode per switch.
• You need linear or SOA-controlled operation. Electronic loads and linear stages demand the MOSFET's stable linear region, not the IGBT's thermal-runaway behavior.

Frequently Asked Questions

Q1: At 650V, which one wins?

A: It depends on current and frequency — below ~20–30A or above ~50kHz the MOSFET wins; above that current at low frequency the IGBT wins. 650V is the contested overlap zone between the two technologies. Evaluate total loss (conduction + switching + drive) at your real load current, frequency, and junction temperature — not at the datasheet's 25°C.

Q2: Why does the IGBT get cheaper to use at high current?

A: Because its conduction loss grows linearly with current while a MOSFET's grows quadratically. A 40A IGBT at 1.45V burns 58W; a 40A MOSFET at 80mΩ burns 128W. Double the current and the IGBT doubles its heat while the MOSFET quadruples it. That's the math behind every motor-drive inverter.

Q3: Why are IGBTs slower?

A: The tail current — stored minority carriers must recombine at turn-off. The same carrier injection that gives the low V_CE(sat) leaves charge behind, so the collector current doesn't stop cleanly when the gate discharges. A MOSFET is unipolar and turns off in tens of nanoseconds; IGBTs are practical to roughly 100kHz and efficient only below ~20–50kHz.

Q4: Does an IGBT need an external diode?

A: Yes — IGBTs have no body diode, so bridge circuits need an antiparallel diode per device. A MOSFET's integral body diode conducts reverse current for free. In any inductive load or bridge design with IGBTs, budget for the extra diodes and their layout.

Q5: Can I use an IGBT in an electronic load?

A: Generally no — IGBTs have poor linear-region behavior and thermal-runaway risk. Forum discussions on dummy loads consistently steer to linear-rated MOSFETs instead. The IGBT's small die and lack of guaranteed DC SOA make it the wrong tool for continuous linear operation.

Q6: Does temperature change the crossover?

A: Yes — R_DS(on) roughly doubles from 25°C to 150°C, which shifts the crossover current lower. A design that crosses at 25A on paper may cross at 15A at a 100°C junction. IGBT V_CE(sat) also rises with temperature but weakly. Always evaluate at the worst-case hot junction, not the 25°C datasheet number.

Q7: What about short-circuit behavior?

A: IGBTs typically withstand shorts for 5–10µs; MOSFETs need faster protection. The IGBT's current-saturation behavior gives a wider protection window, which is why industrial drives favor them in fault-prone environments. If your protection loop is slow, that extra microseconds can be the difference between a trip and a crater.

Q8: Are there 650V MOSFETs at all?

A: Yes — superjunction MOSFETs (CoolMOS, MDmesh class) work well at 650V. They're the reason the crossover zone is contested: they switch fast and conduct well at moderate current. Above ~1kV silicon MOSFETs run out of practical R_DS(on), which is why 1200V-class designs are IGBT (or SiC) territory.