Type a branch, a subject or a topic — “round robin”, “paging”, “civil”.

The MOSFET

The gate is a capacitor, not a junction — which is why CMOS logic burns nothing at rest, why MOSFETs parallel safely, and why static destroys them.

Skip to the animation

A MOSFET's gate is a capacitor rather than a junction, so it draws no steady current and controls the channel by field alone — which is why CMOS logic dissipates almost nothing at rest, why MOSFETs parallel safely, and why static electricity destroys them.

The insulated gate

The gate is separated from the channel by a thin oxide layer — it is a capacitor, not a junction. Steady-state gate current is picoamps against a BJT base's milliamps.

That one fact is why CMOS logic burns almost no power while it is sitting still, and therefore why modern digital electronics at billions of transistors per chip is possible at all.

How the channel forms

  1. 1With no gate voltage, the n-source and n-drain sit in a p-body, so one junction is always reverse-biased. The device is genuinely off.
  2. 2Raise the gate voltage: its field repels holes from the surface and attracts electrons.
  3. 3Past the threshold voltage, the surface inverts from p-type to n-type, forming a channel.
  4. 4Source and drain are now connected by a conducting path.

No carriers were injected across a junction — the field rearranged the ones already there. That is why the gate needs charge to change state and none to hold it.

The two regions

RegionConditionBehaves asUsed for
Cut-offV_GS < V_thOpen circuitThe off state of a switch
Linear / triodeV_DS < V_GS − V_thVoltage-controlled resistorThe on state of a switch
SaturationV_DS > V_GS − V_thVoltage-controlled current sourceAmplifiers

The terminology is a trap: a MOSFET in saturation is amplifying, while a BJT in saturation is a closed switch. The words mean opposite things in the two devices.

As a switch

On, a MOSFET is a small resistance R_DS(on), so its loss is I²R and falls away at low current — unlike a bipolar device's fixed 0.7 V floor. Off, it leaks nanoamps and stores no charge to clear, so it turns off quickly.

The remaining cost is switching loss: charging and discharging the gate capacitance every cycle. That is what limits how fast it is worth switching, and why gate drivers exist.

Why MOSFETs parallel and BJTs do not

R_DS(on) rises with temperature. A device carrying more current heats, becomes more resistive, and sheds current to its neighbours — so parallel MOSFETs balance automatically.

Two paralleled BJTs do the opposite: the hotter one conducts harder and takes more current still. That is thermal runaway, and it is why paralleled bipolar devices need individual emitter resistors to force sharing.

The vulnerability

The gate oxide is a few nanometres thick and breaks down around 20 V. A person walking across a carpet carries kilovolts, and the insulated gate offers no path for that charge to escape — the same insulation that made the device attractive.

Worse, the damage is often latent: the part passes test and fails months later in the field. Internal protection diodes help and do not remove the need for wrist straps and conductive packaging.

Against the BJT

MOSFETBJT
Drive requirementCharge only, no steady currentContinuous base current
Transconductance per mALowerHigher
Matching and noiseWorseBetter
ParallelingSelf-balancingNeeds emitter resistors
Dominates inDigital, power switchingPrecision analogue front ends

BiCMOS processes put both on one die because the argument has no single winner. The IGBT does the same at device level: a MOSFET gate driving bipolar conduction, for high-voltage power.

The numbers you will be asked for

Saturation current

I_D = ½k(V_GS − V_th)²

Transconductance

g_m = 2 I_D / (V_GS − V_th)

On resistance loss

P = I² · R_DS(on)

Switching loss

P ≈ Q_g · V_gs · f_sw

gate charge per cycle

Threshold

V_th ≈ 1–4 V for a power device

Watch it work

loading visualisation…

Check yourself

question 1 / 4

One question at a time. Pick an answer to see why it is right or wrong, then move on — there is no score to keep and nothing is saved.

What single property of the MOSFET makes CMOS logic possible?
Two MOSFETs are put in parallel to share a load. Why does that work when two BJTs would not?
A MOSFET in 'saturation' is doing what?
Why are MOSFETs so vulnerable to static discharge?

0 / 4

4 still unanswered — the dots above jump straight to them.