Switching Transistor
A switching transistor is a transistor designed for fast, clean on-off operation rather than for linear amplification, so that it behaves as nearly as possible like a closed or open switch controlled by a small signal. Almost every digital circuit relies on this mode: in the off state it blocks current with only a trace of leakage, and in the on state it passes current with a small voltage drop across it. What separates a good switching transistor from an ordinary one is how quickly it moves between those states, and the main obstacle is not the turn-on at all but the delay in turning off.
The delay comes from saturation, in which the transistor stores charge it must later remove. Driving a bipolar transistor hard enough that the collector voltage falls to a few tenths of a volt, typically 0.1 to 0.3 volts, forward-biases the collector junction too, and excess minority carriers pile up in the base and collector regions. When the input is removed, collector current does not stop until this stored charge is swept away or recombines, an interval called the storage time that can be much longer than the rise or fall times. Switching speed is therefore described by delay, rise, storage and fall times, and datasheets give them for specified drive conditions. The speed problem is much the same for diodes, which is why fast switching diodes face the same design choices.
Fabrication tricks aim to reduce stored charge without sacrificing the other properties. One classic method is gold doping: diffusing a small amount of gold into the silicon creates deep-level recombination centres that shorten the minority carrier lifetime, so stored charge disappears faster, at the price of somewhat lower gain and higher leakage. Thin, lightly doped base regions and small junctions also help, as do geometries that keep capacitances low. In circuits the same idea appears as a clamp, such as a Schottky diode from base to collector in Schottky TTL, which prevents the transistor from going deep into saturation at all, so there is little charge to remove. A widely used small-signal NPN of this type was the 2N2369, built for saturated switching with turn-off times in the range of a few tens of nanoseconds.
The switching transistor's most consequential role was in the first generations of digital logic. Early transistor computers in the 1950s used germanium devices, including the surface-barrier types from Philco, because they could switch quickly enough to compete with vacuum tubes while being smaller, cooler and more reliable. Diode-transistor logic (DTL) followed, with diodes forming the gate and a transistor providing inversion and gain, and then transistor-transistor logic (TTL) in the early 1960s, whose multi-emitter input transistor became the basis of a huge family of standard integrated circuits. Each logic gate in these families is built around saturating switches, and their propagation delay is largely determined by how quickly a transistor can be turned off.
A common misconception is that a faster transistor is simply one with a higher gain or a higher transit frequency, when the limit in saturated switching is often charge storage. A part with an excellent cutoff frequency can still turn off slowly if it was driven deeply into saturation, because its stored charge, not its gain, controls delay. Another trap is to design for typical base drive: switching circuits are normally given a forced beta, with base current set to a tenth or so of collector current, to guarantee saturation across tolerances, and this overdrive directly lengthens storage time. Designers must therefore balance a low saturation voltage against switching speed, rather than maximizing either alone.
Modern relevance is large, even though the bipolar saturated switch has been displaced in most logic. CMOS logic, which uses MOSFETs and has no minority-carrier storage in the same sense, dominates digital integrated circuits, with billions of transistors acting as switches in a processor. Yet discrete bipolar switching transistors remain common as small drivers for relays, LEDs and motors, and the same trade between conduction loss and switching loss governs power devices. The central design question has stayed constant from germanium to silicon and from TTL to CMOS: how cleanly can a small signal flip a device between two states?
Read switching transistor as an amplifier forced into two states: its quality is judged not by gain but by how little voltage it drops when on, how little it leaks when off, and how fast it sheds stored charge to turn off.