MOSFET Circuits

A MOSFET is a voltage-controlled transistor used in analog, digital and power circuits. A physical MOSFET has gate, drain, source and body terminals, although the body is often tied to the source inside a three-terminal discrete package. Modern integrated circuits can contain large numbers of MOSFET devices.

Device dimensions, structure and manufacturing complexity vary widely by process and voltage rating. Circuit analysis starts from the nonlinear characteristics of MOSFET. Engineers use a large-signal model to find the DC operating point and switching states, then a linearised small-signal model for incremental gain and impedance around a chosen bias point. Both models apply to MOSFET circuits.

The large-signal model describes nonlinear currents and voltages. The small-signal model uses local derivatives of that model and is valid only for changes small enough to stay near the bias point. For an enhancement-mode n-channel device, cutoff, triode and saturation are set by gate-to-source voltage V_GS, threshold voltage V_tn and drain-to-source voltage. Amplifiers commonly bias the device in saturation. Switches use cutoff for off and the low-resistance triode region for on, with gate drive selected from datasheet conditions rather than threshold voltage alone.

MOSFET Driver Circuits

A gate driver sources and sinks current to charge and discharge the MOSFET gate within the required switching time. Slow transitions can increase the interval in which drain voltage and current overlap, raising switching loss and junction temperature. Drive strength, gate resistance, layout inductance, switching frequency and electromagnetic-interference limits determine the useful transition time. Some floating high-side n-channel drivers use a bootstrap supply, which imposes duty-cycle and refresh constraints. Other MOSFET circuits use ground-referenced, isolated or charge-pump drive supplies.

The insulated gate presents nonlinear capacitances to the driver. Datasheets combine their switching effect into gate-charge curves, including the Miller plateau. A driver turns the device on or off by moving this charge, so treating the gate as one fixed ideal capacitor is only a first approximation.

MOSFET Switching Circuits

MOSFETs used as switches move between cutoff and a low-resistance on state in the triode region. They pass through saturation during each transition, so driver design controls switching loss as well as steady-state conduction loss.

MOSFET switching circuits include the power device, gate driver, load and current return path. When the transistor is on, current flows through its channel and the load. N-channel devices often offer lower on-resistance for the same die area, while p-channel devices can simplify some high-side drives. Selection also depends on voltage, current, gate charge, reverse recovery, thermal limits and cost.

A common n-channel low-side circuit connects the load between the positive supply and drain, then connects source to ground. A high-side circuit uses a different arrangement and must drive the gate relative to the moving source node. Gate-source voltage must reach the datasheet value that guarantees the required on-resistance; merely crossing threshold produces only the small test current used to define that specification. A MOSFET is not universally a better switch than a BJT. Compare conduction loss, drive loss, switching speed, voltage rating, temperature and safe operating area when selecting MOS switches.

MOSFET Inverter Circuits

A logic inverter produces a low output for a high input and a high output for a low input. It is different from a power inverter, which converts DC power into AC power. Logic inverters form logic gates and more complex digital circuits. The diagram below shows an idealised voltage-transfer characteristic.

Early MOS logic used p-channel devices, followed by n-channel logic and then complementary MOS. Electron mobility is normally higher than hole mobility in comparable silicon channels, so an n-MOS device can provide more drive for the same geometry. CMOS pairs n-channel and p-MOS devices to reduce static current at valid logic levels. The following sections compare three historical inverter forms.

Resistive-load n-MOS inverter:
This circuit places load resistance R between the supply and output, with an n-channel MOSFET between output and ground.

When V_in is below threshold, the transistor is off and the resistor pulls output toward the supply. Any output capacitor charges through that resistor. A valid high input turns the transistor on and pulls output toward ground, but the low level is not exactly zero because the device has finite on-resistance. The circuit draws static current in the low-output state and an integrated resistor can occupy substantial area.

Active-load n-MOS inverter:
This topology uses an n-channel MOSFET as the load instead of a resistor. A pull-down transistor draws the output toward the lower rail, while the load transistor draws it toward the upper rail. Logic levels, static current and transition speed depend on the two device sizes and bias arrangement.

In the following active-load circuit, the pull-up gate is tied to the supply so that device remains on.

CMOS inverter:
A CMOS inverter uses an n-channel and p-channel pair with common gates and drains. The p-channel device pulls output up, while the n-channel device pulls it down.

When V_in is low, the n-channel device turns off and the p-channel device turns on. The output capacitor charges toward the supply and produces a high output.
When V_in is high, the n-channel device turns on and the p-channel device turns off. The capacitor discharges toward ground and produces a low output.

CMOS provides a steep transfer transition and very low static power at valid logic levels. Real circuits still dissipate leakage power, dynamic power while charging and discharging capacitance and short-circuit power while both devices conduct during an input transition. Complementary fabrication also requires both device polarities.