A MOSFET and a BJT can both switch the same load in many circuits, and choosing between them comes down to what the rest of the circuit asks. The BJT is a current-controlled valve: it turns on with base current, and its on-state drop stays relatively fixed. The MOSFET is a voltage-controlled gate: it turns on with gate voltage, and its on-state drop scales with the current through its resistance. That single difference drives everything else — the drive signal, the efficiency at different currents, the cost, and the failure behavior. This article builds the one-paragraph model of each, compares the drive requirements and the losses at two current levels, and gives an eight-question decision flow for picking the switch.
A Valve vs a Gate: The One-Paragraph Model of Each
A BJT behaves like a current-controlled valve: a small current into the base opens the collector-emitter path, and once saturated the collector-emitter voltage is a roughly fixed drop — about 0.2–0.3 V for a power device. A MOSFET behaves like a voltage-controlled gate: a voltage on the gate opens a channel, and the on-state is a resistance, so the drop is the current times the on-resistance. The mental model is the difference between a valve that opens fully for a fixed drop and a gate whose resistance it is.
The consequence is that the two devices ask different things of the driver. The BJT wants base current, and the driver must supply it continuously while the load is on. The MOSFET wants gate voltage, and the gate draws current only during the switching transition — the steady state needs no current at all. The MOSFET gate-drive requirement and the BJT base-current math are the two drive-side references; this article positions the whole choice.
The one-paragraph model is the reusable tool: valve versus gate, current-driven versus voltage-driven, fixed drop versus resistive drop.
What the Drive Signal Asks From You: Current vs Voltage
The drive requirement is where the two devices split most cleanly. A BJT switch needs a base resistor and a continuous base current sized from the collector current and the minimum hFE; a driver that cannot source that current leaves the BJT half-saturated and dissipating. A MOSFET switch needs a gate drive that can charge and discharge the gate capacitance fast enough, and a gate voltage above the threshold; the steady-state gate current is negligible, which is why a logic gate can drive a MOSFET directly where it would struggle to sink continuous base current, and why battery-powered designs favor the MOSFET’s near-zero steady drive.
The drive difference shows up in the interface. A microcontroller output that can source a few milliamps comfortably drives a logic-level MOSFET’s gate but may undersupply a power BJT’s base at a heavy load. Conversely, a BJT is more forgiving of a slow, low-voltage drive that a MOSFET might leave below threshold. The choice of drive hardware often decides the device rather than the other way around.
The practical read: if the controlling signal is a clean voltage source with little current to spare, the MOSFET wins; if the circuit already has a defined base-drive current path, the BJT slots in naturally. The NPN vs PNP polarity guide and the logic-level MOSFET article cover the two drive-side families.
Loss Comparison at 1 A and 50 A: Two Tables That Tell the Story
The loss comparison is where the two devices cross over, and the crossover follows the drop-versus-resistance difference. At low current, the BJT’s fixed drop can look efficient; at high current, the MOSFET’s resistive drop, which grows with current, can overtake or fall behind depending on the on-resistance.
| Current | BJT loss (drop 0.3 V) | MOSFET loss (RDS 8 mΩ) | Winner |
|---|---|---|---|
| 1 A | 0.3 W | 0.008 W | MOSFET |
| 10 A | 3 W | 0.8 W | MOSFET |
| 50 A | 15 W | 20 W (hot, higher RDS) | Depends on hot RDS |
The 1 A row shows the MOSFET’s advantage is already large because its resistive drop at 1 A is tiny. The 10 A row keeps the MOSFET ahead. The 50 A row is the interesting case: the MOSFET’s hot RDS(on) grows with temperature, and at very high current a BJT’s fixed drop can actually lose less if the MOSFET’s resistance is high — but in practice a well-chosen MOSFET with low RDS(on) usually still wins. The MOSFET thermal reality article explains why the hot RDS(on) is the number to use, and the low RDS(on) analysis shows when paying for a lower resistance pays.
The loss tables tell the story the model predicts: the MOSFET is efficient across the range when its on-resistance is low, and the comparison must use the hot resistance at the real current, not the 25 °C headline. The tables also reveal the other half of the trade — at high current the MOSFET’s loss grows with I², so a marginal part can cross the BJT’s fixed-drop curve sooner than a naive 1 A comparison suggests, which is why the hot-resistance row matters as much as the low-current one.
The failure behavior is a second real difference that the model predicts but beginners often miss. A BJT in saturation has a roughly fixed collector-emitter drop, so an overcurrent shows up as excess heat in the transistor and eventually a thermal failure — a gradual, heat-driven death. A MOSFET’s on-state is a resistance, so an overcurrent raises the voltage drop and the dissipation in a way that compounds with the temperature rise of its own RDS(on), a faster, more thermal-runaway-like failure at the same margin. The two fail differently under the same overload, and the circuit context decides which failure mode is more acceptable.
The drive interface also has a subtlety worth naming: a MOSFET gate is a capacitor, and driving it fast enough for a clean edge means the driver must source and sink the transient gate current. A design that switches the MOSFET slowly — because the drive is weak — spends the transition time in the resistive region, where both voltage and current are present, and dissipates far more than the conduction-loss table suggests. This is why a MOSFET chosen for low RDS(on) can still run hot if its gate drive is underpowered; the switching loss, not the conduction loss, is the hidden term. The loss sheet method frames the switching term the same way it does for diodes, and the MOSFET switching analysis adds the gate-charge side that sizes the driver.
Cost and Sourcing: Where the BJT Still Wins in 2026
The MOSFET wins most of the electrical comparison at low to moderate current, but the BJT still holds a real edge on cost and simplicity in specific roles. A small-signal BJT is cheaper and more forgiving in low-frequency, low-current switching where the loss difference is irrelevant; a discrete MOSFET adds a gate-charge consideration and often a driver for the faster edges. The cost gap narrows with scale, but the BJT’s simplicity — one base resistor, no gate driver, no threshold margin to manage — keeps it the default for many small loads.
The sourcing picture is the other side. Power MOSFETs have a wider selection and a stronger high-current catalog, while the BJT remains dominant in the small-signal and logic-switching space. The honest 2026 read is that the MOSFET owns the power-switching domain, and the BJT owns the cheap, simple, low-current end. The MOSFET families and transistor families on the site span both, and the MOSFET selection guide walks the power side in full.
Switching Decision Flow: Eight Questions Before You Pick
The decision flow reduces the choice to a short list of questions. Question one, what current does the switch carry — above a few amps the MOSFET’s low resistance wins. Question two, what voltage does it block — the MOSFET’s voltage range is broader. Question three, how fast is the switching — the MOSFET handles fast edges; the BJT’s storage delay is a limit. Question four, what drive signal is available — a voltage source favors the MOSFET, a current source the BJT. Question five, what is the load type — a resistive, inductive, or capacitive load changes the switching-loss and snubber demands.
Questions six through eight complete the picture: what cost target — the BJT wins the cheap low-current end; what failure behavior is acceptable — the two fail differently and the circuit context decides; and what is the thermal budget — the loss tables at the real current decide it. The MOSFET datasheet hub and the IGBT boundary article extend the flow to the next device tier when the MOSFET hits its voltage or frequency limit, and the LV/MV MOSFET families supply the parts the flow lands on for the power-switching role. The closing judgment is that the MOSFET and BJT are not rivals but neighbors on a spectrum: the BJT owns the cheap, simple, low-current switch, and the MOSFET owns the power and speed domain — and the eight questions locate any real design on that spectrum in a few minutes, with no memorized part numbers required along the way.