MOSFETs & FET Circuits

gate charge

The MOSFET gate draws no steady current, but it is a capacitor, and turning the device on means filling that capacitor with charge. Gate charge, written Qg, is the total amount of charge you must pour into the gate to take the device from off to fully on at a given drive voltage. It is measured in nanocoulombs, and it is the truest measure of how hard a MOSFET is to switch quickly, because time equals charge divided by current.

If your driver can source a current I, the gate takes a time of about Qg divided by I to charge. For example, a part with Qg = 20 nanocoulombs driven by a 1 amp gate driver switches in roughly 20 nC / 1 A = 20 nanoseconds. Want to switch in 5 ns? You need 4 amps of gate-drive current, briefly. The gate-charge curve also reveals the Miller plateau, a flat stretch where the gate voltage stalls while the drain voltage swings: this plateau is where the actual switching transition happens and where most switching loss is born.

Gate charge sets the cost of switching at frequency. The average power your driver spends just moving the gate is roughly Qg times the gate-drive voltage times the switching frequency: for instance 20 nC times 10 V times 100 kHz = 0.02 W, small at 100 kHz but rising linearly into watts at megahertz speeds. There is a fundamental tension with on-resistance: a bigger MOSFET has lower R_DS(on) but more gate charge, so picking a part means trading conduction loss against switching effort.

Two parts, same on-resistance: one with Qg = 15 nC switches in half the time of one with Qg = 30 nC for the same gate driver, so the low-charge part wins in a high-frequency converter.

Switching time is gate charge divided by drive current.

Low on-resistance and low gate charge pull in opposite directions: a bigger die lowers R_DS(on) but raises Qg. There is no free lunch; pick for your switching frequency.

Also called
Qg閘極電荷