A metal oxide semiconductor fet (MOSFET) is a voltage-controlled transistor that uses an insulated gate to regulate current flow between its drain and source terminals, acting as a highly efficient electronic switch or amplifier. What this changes in a real circuit is the ability to switch high-power loads—like 12V DC motors, solenoid valves, or 50W LED arrays—using ultra-low current signals from 3.3V or 5V microcontrollers, completely eliminating the need for the heavy, continuous base-current drive required by older bipolar transistors.
Core Theory: Voltage-Controlled Switching
Unlike older transistor designs that rely on current flow through the control terminal, the metal oxide semiconductor fet relies on an electric field. The 'O' and 'S' in the name refer to the thin layer of silicon dioxide (glass) insulating the metal gate from the semiconductor channel. Because of this insulation, the gate draws virtually zero continuous current. You only need to supply enough voltage to create an electric field that pulls charge carriers into the channel, bridging the gap between the drain and source.
To visualize this, think of a water valve where the gate is a flexible rubber diaphragm. Applying air pressure (voltage) to the diaphragm pushes it down to open the pipe (channel), allowing water (current) to flow. Crucially, the water never leaks back into the air pressure control line because the rubber diaphragm (the oxide insulation) keeps the two systems physically separated.
Two parameters dictate real-world performance:
- Gate-Source Threshold Voltage ($V_{GS(th)}$): The minimum voltage where the device just begins to turn on (usually allowing a tiny 250µA current). This is not the voltage needed to fully turn it on.
- Drain-Source On-Resistance ($R_{DS(on)}$): The resistance of the channel when fully enhanced. Lower is always better for power switching, as it dictates heat generation.
Worked Example: Sizing for a 12V DC Motor
Let us look at a common bench scenario: switching a 12V DC motor that draws 8A continuous and 15A at stall, controlled by a 5V Arduino Nano. Many beginners reach for the IRF520 module because it is cheap and widely available. This is a critical mistake.
The IRF520 is a standard-level device. Its datasheet only guarantees a low $R_{DS(on)}$ when the gate is driven at 10V. At 5V, it remains partially in its linear (resistive) region, resulting in an $R_{DS(on)}$ well over 1Ω. Using the power formula $P = I^2 \times R$, the dissipation would be $8^2 \times 1 = 64W$. The TO-220 package will melt and fail in seconds.
Instead, we select a logic-level N-channel MOSFET like the IRLB8721. According to the semiconductor theory outlined by All About Circuits, we must check the $R_{DS(on)}$ test conditions in the datasheet. For the IRLB8721, $R_{DS(on)}$ is specified at $V_{GS} = 4.5V$, measuring just 4.6 mΩ (0.0046 Ω).
The Math:
- Power Dissipation: $P = 8^2 \times 0.0046 = 0.294W$
- Thermal Resistance ($R_{\theta JA}$) for a bare TO-220 in free air is roughly 62°C/W.
- Temperature Rise: $\Delta T = 0.294W \times 62°C/W = 18.2°C$
- Final Junction Temp: 25°C (ambient) + 18.2°C = 43.2°C.
At 43.2°C, the IRLB8721 runs comfortably cool without a heatsink, proving why checking the $V_{GS}$ test condition is the most important step in MOSFET selection.
Where You Meet This in Practice
You will rarely build a discrete motor driver from scratch for commercial products today, but the metal oxide semiconductor fet remains the underlying workhorse in almost all modern power electronics:
- DC-DC Buck/Boost Converters: Synchronous rectifiers use MOSFETs instead of diodes on the low-side to eliminate the 0.4V forward voltage drop, pushing efficiency past 95%.
- Battery Management Systems (BMS): High-current LiFePO4 packs use back-to-back MOSFETs as solid-state disconnect switches for over-current and short-circuit protection.
- Solid State Relays (SSRs): DC-output SSRs are essentially optocouplers driving high-power MOSFETs, providing galvanic isolation between a PLC logic output and a 48V industrial load.
- Brushless DC (BLDC) Motor Inverters: The electronic speed controllers (ESCs) in drones and EVs use six MOSFETs arranged in a 3-phase bridge to commutate the motor windings.
MOSFET vs. BJT: Clearing Up the Confusion
What people most commonly confuse the metal oxide semiconductor fet with is the Bipolar Junction Transistor (BJT), such as the 2N2222 or TIP31. While both have three terminals and can act as switches, their control physics are entirely different. A BJT is a current-controlled device; to keep a BJT saturated (fully on), you must continuously feed current into the base terminal (typically 1/10th of the collector current). A MOSFET is voltage-controlled; once the gate capacitance is charged, it requires zero continuous current to stay on.
| Feature | Metal Oxide Semiconductor FET | Bipolar Junction Transistor (BJT) |
|---|---|---|
| Control Mechanism | Voltage (Electric Field) | Current (Minority Carrier Injection) |
| Continuous Drive Power | Near Zero (only switching losses) | High ($V_{BE} \times I_B$) |
| On-State Voltage Drop | Very Low ($I \times R_{DS(on)}$) | Fixed Saturation Voltage ($V_{CE(sat)}$ ~0.2V) |
| High-Frequency Switching | Excellent (Majority carrier device) | Poor (Storage time delays turn-off) |
| Parallel Operation | Self-balancing (Positive temp coefficient) | Prone to thermal runaway (Negative temp coefficient) |
For a deeper dive into the switching dynamics and gate charge requirements, the Texas Instruments gate driver fundamentals guide (SLUP169) remains the definitive reference for calculating the exact peak current your gate driver must supply to achieve nanosecond switching times.
Frequently Asked Questions
How do I know if my metal oxide semiconductor fet is logic-level?
Ignore the $V_{GS(th)}$ (threshold voltage) on the front page of the datasheet, as this only indicates when the device begins to leak a few hundred microamps. Instead, look at the Static Electrical Characteristics table for the $R_{DS(on)}$ specification. If the 'Test Conditions' column lists $V_{GS} = 4.5V$ or $V_{GS} = 5V$, it is a true logic-level device safe for 5V Arduinos. If you are using a 3.3V ESP32, you need an ultra-logic-level part specifically tested at $V_{GS} = 2.5V$, such as the CSD17571Q5A.
Why does my metal oxide semiconductor fet get hot when switching PWM?
This is caused by switching losses. During the transition between the OFF and ON states, the MOSFET passes through its linear region where both the drain-source voltage and the drain current are simultaneously high. If your microcontroller GPIO pin cannot source enough current to charge the gate capacitance ($Q_g$) rapidly, the transition takes too long. The device spends excessive time in this high-dissipation zone, generating massive heat even if the $R_{DS(on)}$ is tiny. The fix is to add a dedicated gate driver IC (like the TC4420) between the microcontroller and the gate.
What is the difference between an N-channel and P-channel metal oxide semiconductor fet?
N-channel devices use electrons as charge carriers, making them faster, cheaper to manufacture, and capable of much lower $R_{DS(on)}$ values for a given die size. They are the standard choice for 'low-side' switching (placed between the load and ground). P-channel devices use 'holes' as carriers, resulting in higher resistance and higher costs. They are used for 'high-side' switching (placed between the power supply and the load) because their gate can be pulled to ground to turn on. However, in modern high-power designs, engineers often use an N-channel MOSFET on the high-side combined with a charge-pump circuit to generate the required elevated gate voltage, avoiding the performance penalty of P-channel silicon.
