Current-controlled vs. voltage-controlled devices — and why one's drive circuit won't correctly drive the other.
Both a bipolar junction transistor (BJT) and a MOSFET use a third terminal to control current flowing between the other two. Past that, the resemblance mostly ends. A BJT's control terminal — the base — is fundamentally current-controlled: a small base current Ib, pushed continuously into a forward-biased base-emitter junction, controls a much larger collector current Ic = β·Ib, and Ib has to keep flowing for as long as the device is meant to conduct. A MOSFET's control terminal — the gate — is fundamentally voltage-controlled: the gate sits on the far side of a thin insulating oxide, isolated from the channel, so once the gate-source capacitance is charged to the voltage that opens the channel, essentially zero steady-state gate current is needed to hold it open. The gate draws real current only transiently, while that capacitance is being charged or discharged during a switching edge.
A BJT's base is connected to the emitter through a real, forward-biased PN junction — the same kind of junction as a diode. Forward-biasing that junction and keeping it forward-biased requires a real, continuous DC current, exactly like keeping a diode conducting requires continuous current through it. That base current doesn't just turn the device on once; it has to be resupplied every instant the transistor is meant to keep conducting, and the resulting collector current Ic tracks it in real time, scaled by the current gain β (Ic ≈ β·Ib). Stop supplying Ib and Ic collapses almost immediately. A MOSFET's gate, by contrast, is separated from the channel by a thin layer of insulating oxide — there is no DC conduction path into the channel at all. The gate and channel instead form one plate pair of a capacitor (often called Ciss, the input capacitance). Charging that capacitor to a voltage above the threshold voltage Vt opens a conductive channel between source and drain; once charged, the channel stays open with the gate simply holding a static voltage, drawing no further current, the same way a charged capacitor holds its voltage on an open circuit.
A gate that draws essentially zero steady-state current, with very high DC input impedance, is exactly what digital logic and most modern power-switching applications want: the driving stage doesn't need to continuously source current just to hold a device on, so gate drive can be simple and low-power, and a single logic output can hold many gates at once with no ongoing current penalty. That's the core reason MOSFETs dominate digital ICs and the vast majority of modern switching power converters. A BJT, needing continuous base current proportional to the collector current it must sustain, is a heavier steady-state load on its drive circuit — but that same current-in, current-out relationship gives BJTs some enduring advantages in certain analog roles: BJTs can conduct a given current with less required input voltage swing (a base-emitter junction drop of roughly 0.6–0.7V regardless of current over a wide range, versus a MOSFET's gate-source voltage that must rise further above threshold to conduct more current), and their exponential Ic-Vbe relationship gives very predictable, tightly-matched behavior that some analog circuits — precision current mirrors, certain low-noise or high-linearity amplifier stages — are specifically designed around.
A BJT's base sees a forward-biased PN junction on the other side — a real diode-like conduction path that, like any forward-biased diode, needs continuous current to stay conducting. A MOSFET's gate sees an insulating oxide layer on the other side — an actual dielectric with no DC conduction path at all, forming a capacitor between the gate and the channel. Charging that capacitor to the right voltage is a one-time (per switching event) transfer of charge, Q = Ciss·ΔV; holding it there afterward costs nothing but leakage, which is why steady-state gate current is essentially zero. Neither device is a "better" version of the other — they simply have different structures behind the same three-terminal package, and those structures dictate current-controlled versus voltage-controlled behavior directly.
They're both three-terminal current-steering devices, but the resemblance stops at the package outline. A drive circuit built for a BJT is designed to continuously source real currentinto the base for as long as the device needs to conduct — typically a resistor or current source sized for a specific steady Ib. Feed that same circuit into a MOSFET's gate and, once the gate capacitance charges up, essentially no further current flows to it — the gate will sit at whatever voltage that continuous small current happens to settle it to, which may not reliably hold the channel fully enhanced, and none of the fast-charge, low-impedance drive a MOSFET needs for a clean, low-loss switching edge is present. Run it the other way — drive a BJT's base from a circuit designed only to charge a capacitance briefly during a MOSFET-style switching edge — and the base current isn't resupplied once that transient ends; the BJT starves for continuous Ib, Ic collapses, and the device never reaches (or stays in) a fully conducting state. Swapping device families always means redesigning the drive stage, not just replacing a part.
Explains why a BJT is fundamentally a current-controlled device — a small, continuous base current controls a much larger collector current through current gain β — while a MOSFET is fundamentally a voltage-controlled device — a gate voltage charges a capacitance to open a channel, after which essentially zero steady-state gate current is required. Covers why this distinction drives where each device family dominates, and why the two are not interchangeable in the same drive circuit.
Both devices are drawn as three-terminal symbols, both can switch or amplify, and modern discrete and IC catalogs list them side by side, which invites treating a MOSFET as simply a newer, better BJT — swap one in for the other and it should just work. It doesn't. A BJT's base terminal sits behind a real forward-biased PN junction that needs continuous current to stay conducting; a MOSFET's gate sits behind an insulating oxide layer with no DC conduction path at all, needing only a one-time charge to open the channel. A drive circuit engineered to continuously supply base current has no mechanism to deliver the low-impedance, fast charge/discharge a MOSFET gate needs for clean switching — and a drive circuit that only delivers a brief charging pulse has no mechanism to keep resupplying the continuous current a BJT's base requires to stay on.
A BJT's base-emitter junction is forward-biased like a diode; sustaining conduction across it requires a real, continuous DC current Ib, and the resulting collector current tracks it as Ic ≈ β·Ib, where β is the current gain. Interrupt Ib and Ic collapses almost immediately, since there is no charge storage mechanism holding the junction open on its own. A MOSFET's gate is separated from the channel by a thin dielectric (the gate oxide), forming a capacitor (Ciss) rather than a conduction path. Charging that capacitor to a gate-source voltage above the threshold voltage Vt opens a conductive channel between source and drain; the charge, once placed, stays there with the gate acting purely as a static voltage source, and drain current Id continues to flow — set by the held Vgs — independent of any further gate current. Gate current is nonzero only transiently, proportional to Ciss·(dV/dt), while the gate voltage is actively being changed.
A device whose control terminal draws essentially zero steady-state current and presents very high DC input impedance is ideal wherever a drive signal needs to control many devices cheaply and with minimal ongoing power — digital logic gates and the vast majority of modern power-switching converters, both dominated by MOSFETs (and their relative, the IGBT, which combines a MOSFET gate with a BJT-like output stage). A BJT's continuous current-in/current-out relationship instead gives a highly predictable, exponential relationship between Vbe and Ic that some analog design specifically leans on — precision current mirrors, certain low-noise or high-linearity amplifier stages, and applications wanting a given conduction current at a smaller required control-terminal voltage swing than an equivalent MOSFET would need. Because the two devices' control terminals behave so differently at DC, any circuit driving one has to be redesigned, not just re-populated with the other part, when switching device families.
No — a MOSFET gate draws essentially zero current only in steady state, once the gate-source capacitance (Ciss) has already been charged to the desired voltage. While that voltage is actively being changed — every switching transition — the gate draws a real, sometimes substantial transient current proportional to Ciss times the rate of voltage change (dV/dt), which is exactly why fast, high-power MOSFET switching needs a low-impedance gate driver capable of sourcing and sinking sizable peak currents, even though the average/steady-state current is negligible.
A gate driver is built to deliver a brief charging pulse and then supply almost no current afterward, since that's all a MOSFET gate needs. A BJT's base needs the opposite: a real, continuously resupplied current for as long as the device should conduct. Once the gate-driver-style pulse ends, the BJT's base current isn't sustained, the base-emitter junction stops being adequately forward-biased, and the collector current collapses — the BJT effectively turns back off shortly after being turned on.
Yes, in the sense that a higher β means less base current Ib is required to sustain a given collector current Ic, since Ic ≈ β·Ib. But the base current still has to be continuous and real, no matter how small β makes it — a high-β BJT still needs an uninterrupted DC path supplying that (smaller) Ib, unlike a MOSFET gate, which needs no continued current at all once charged, regardless of any device parameter.
BJTs offer a lower, more predictable control-terminal voltage for a given conduction current (base-emitter voltage stays near 0.6–0.7V across a wide current range, versus a MOSFET's gate-source voltage that must rise further above threshold to conduct more current), and their exponential Ic-Vbe relationship gives very tight, predictable device matching that some analog circuits — precision current mirrors, certain high-linearity or low-noise amplifier stages — are specifically designed to exploit. In many switching-power and digital-logic applications, though, the MOSFET's near-zero steady gate-drive requirement makes it the clearly preferred choice.
Neither exactly — an IGBT (insulated-gate bipolar transistor) is a hybrid: it has a MOSFET-style, voltage-controlled, near-zero-steady-current gate for its control terminal, but its output stage behaves more like a BJT's, often giving lower conduction losses at high voltage and current than a MOSFET alone. It's a deliberate combination built to get the MOSFET's easy, low-power gate drive together with some of the BJT's high-voltage/high-current conduction characteristics.
Try our Semiconductor & Electronics Studio
More calculators, simulators, and guides for this discipline.