Transistors are three-terminal semiconductor devices. One terminal controls electrical resistance or current flow between the other two terminals, giving transistors a valvelike operation. Transistors can amplify signals over some linear range of voltage and current or they can rapidly switch signals between two or more levels. Typical switching circuits include inverters, converters, voltage regulators, and relay and solenoid drivers.

Transistors are grouped into two categories: bipolar-junction transistors (BJTs) and field-effect transistors (FETs). The two types of field-effect transistors are metal-oxide semiconductor FETs, or MOSFETs, and junction FETs, or JFETs.

BJTs: Bipolar junction transistors are composed of two p-n junctions formed by three layers of semiconductor material. The layers are called the emitter, base, and collector. An electrical signal (current) applied to the base terminal controls the current between the emitter and collector.

The emitter and collector are both made from the same type of material, n-type in npn or p-type in pnp transistors. N-type material is negatively doped to have excess electrons, and p-type positively doped to be deficient of electrons. In either case, the base material is opposite that of the emitter and collector. The two types of BJTs can be differentiated in schematic diagrams by their circuit symbols.

Among the most important transistor characteristics are maximum Vce (voltage between collector and emitter), Ic (current flowing through the collector), power dissipation, and maximum allowable junction temperature (collector junction). Exceeding these values may damage the transistor.

JFETs: Junction field-effect transistors are composed of an n or p-type semiconductor channel with two electrical contacts, called source and drain, at opposite ends. Two junctions, of opposite polarity with respect to the channel, laterally bound the channel along its length. The two junction contacts are normally tied together and called the gate. In schematic diagrams an inward-pointing arrowhead on the gate lead signifies an n-channel JFET. An outward pointing arrowhead is used for p-channel JFETs.

An electrical signal applied to the gate terminal controls the channel resistance between the source and drain. The gate voltage VG reverse biases the gate/channel junction. Because the reverse-biased channel region is depleted of electrons (n-channel JFET), it does not conduct current flow. Thus, the width of the "conductive" channel is decreased and its resistance is increased.

Channel voltage comes from the drain to source voltage Vds. Because the source is normally grounded, the channel voltage is greatest near the drain, causing the depletion region to be wider and the "conductive" channel to be narrower there. As Vds is increased, the depletion region widens and the channel resistance increases. At pinchoff, the channel resistance reaches a maximum and the current flowing into the drain saturates.

MOSFETs: Metal-oxide semiconductor field-effect transistors have three or four terminals. A fourth terminal can provide a second connection to the gate, which can minimize gate-drain capacitance in high-frequency applications. Although several circuit symbols are used to denote MOSFETs, inward-pointing arrowheads always signify n-channel devices and outward-pointing arrowheads p-channel.

Three-terminal MOSFETs have the same terminal names as JFETs; source, drain, and gate. Like the JFET, a voltage applied to the gate terminal controls MOSFET channel resistance. But the gate region is not a junction. It is a metal-oxide sandwich running the length of the channel surface. MOSFETs also differ in that their source and drain regions are junctions and not simply contacts.

Source and drain are shallow n-type regions in an n-channel MOSFET. The channel is p-type. With no bias applied to the gate, the resistance of the path between the source and drain is high and current does not flow. A positive voltage on the gate induces a buildup of electrons below the gate oxide. When the gate voltage is sufficiently large, the p-type channel is inverted to n-type material and it connects the n-type source and drain. As VG increases, the inversion channel widens lowering the channel resistance.

Because little current flows into the gate -- it is an insulator -- MOSFETs consume the least amount of power of the three transistors. And for the same reason, it has the highest input resistance. Input resistance typically is in excess of 1010 \#189>, allowing one device to operate a large number of MOSFETs.

There are two types of MOSFETs, which differ in construction and in operation. One type is called a depletion-mode MOSFET and the other enhancement mode. A depletion-mode MOSFET conducts current without a gate bias. In an n-channel device, a thin n-type region exists under the oxide in the absence of an applied bias. It connects the source and drain allowing current flow. In fact, a negative voltage is required to drive the electrons out of (deplete) the region to increase channel resistance and reduce current flow.

In an enhancement-mode MOSFET there is no current conducted in the absence of a gate voltage. Without a gate bias, the channel under the gate is the opposite polarity of the source and drain. Increasing negative gate voltage to "drive more electrons out of the channel" has no effect. A positive voltage induces an n-channel and allows current to flow.

MOSFET devices must be protected from static electrical charges because excessive charges on the gate terminal can destroy the device. For example, the static charge generated on a person's body can weaken or break the gate oxide. Some MOSFETs incorporate internal protective diodes to combat static charges. Others are packaged with all leads shorted together with a metal ring that should not be removed until the device is permanently installed in the circuit.

MOSFETs are gaining on BJTs in terms of new applications. They are faster and require less drive power, and low-voltage versions exhibit less on-resistance. Also, prices of low-voltage MOSFETs have dropped rapidly because of new designs and improved manufacturing techniques.

MOSFET manufacturers have cut product cost by increasing cell density. MOSFET chips contain a large number of identical cells connected in parallel. The first generation of power MOSFETs typically contained 500,000 cells/in.2, but second-generation devices contain 800,000 cells. Third-generation MOSFETs have 1 million or more cells/in.2 Higher cell densities are advantageous because in devices rated at 100 V or less, on-resistance drops rapidly as cell density increases and channel length decreases.

New fabrication techniques also provide faster fall times because of the smaller volume of material. Fall time for a new 50-V chip, for example, is 30 nsec, down from the 90 nsec typical for earlier devices. The small chips also turn on at about half the gate voltage. Input and output capacitance is lower by about the same amount.

Despite these advantages, MOSFETs have not been widely used in pulse-width-modulated (PWM) motor controls. A parasitic bipolar transistor, inherent in the device structure, made MOSFETs unreliable for this application. Unless each MOSFET was protected by a pair of diodes, the bipolar transistor could turn on unintentionally and destroy the MOSFET. The diodes, however, are costly and use precious chip space.

Excessive current in the body turns on the bipolar in forward or reverse mode depending on the direction of flow. The current may result from excessive dv/dt or di/dt. The reverse bias safe operating area (RBSOA) of the bipolar is much smaller than for the MOSFET. Thus, when a parasitic bipolar turns on and then attempts to turn off, it may fail if the voltage across it exceeds 50 or 60% of the MOSFET's rated value. Avalanching can also start a secondary breakdown where sufficient external energy is available.

Above 200 V, the epitaxial channel material primarily determines on-resistance, thus, epi thickness must be increased. Resistance rises rapidly with increasing voltage. A result is that for a given current rating, high-voltage MOSFETs are much larger and, thus, more costly than low-voltage components. Nevertheless, MOSFETs sometimes compete with bipolar transistors in the 200 to 500-V range because drive circuits are less expensive.

Further, MOSFETs are now competing with BJTs in PWM motor-control applications. Newly designed power MOSFETs virtually eliminate the parasitic bipolar transistor, allowing MOSFETs to withstand large unclamped inductive loads. The design uses a continuous interdigitated metal source and gate instead of a closed-source cellular structure which eliminates overlap input capacitance and has large die size to achieve minimum on-resistance. The larger die size lowers the thermal resistance and reduces current density.

Presently, input capacitance limits switching speed of most MOSFETs to 200 kHz for applications such as motor control, switching power supplies, and PWM amplifiers. The lower input capacitance of the new devices will take the switching speeds up to 1 MHz at high voltage and current. The consensus is that such devices will become the preferred power switch for applications demanding over 500-V withstand, and possibly for those up to 1,000 V or more.

IGBTs: Insulated-gate bipolar transistors (IGBT) combine the strengths of MOSFET and bipolar transistors. They have a MOS-like input with a BJT-like output. Voltage-controlled drive consumes little power because of the insulated gate. A large reverse blocking capability is possible because there is no body diode as in MOSFETs.

IGBT output capability is on par with power-bipolar Darlington pairs -- 50 A and 1,000 V. But switching times are 10 times faster because of a lack of junction effects at the input. Turnoff times range from 0.1 to 10 ∝sec, for example. IGBTs also have a larger safe operating area. Further, current handling capability is increased by a low saturation voltage.

IGBT applications include motor controls and other industrial electronic systems. Although wide acceptance has been hampered by reports of slow fall times and overcurrent latching, several manufacturers claim to have overcome the latch problem. IGBT modules that handle 200 A and others that switch 1,200 V are now available.

Thyristors: Thyristors have turn-on characteristics that can be controlled by externally applied voltage or current. These devices generally consist of interconnected layers of p and n-type semiconductor material.

Silicone-controlled rectifier (SCR): These are sometimes called reverse-blocking thyristors and have three terminals: anode, cathode, and gate. With a small forward voltage applied between the anode and cathode, the SCR is normally off and conducts no current (high-resistance state). If forward voltage increases sufficiently, the SCR begins to conduct. SCR voltage at the conduction point is called the breakover voltage. A signal applied to the gate terminal controls the level of breakover voltage. The higher the gate current, the lower the breakover voltage.

A gate signal need last from a few microseconds to milliseconds to fire an SCR into the forward-conducting state. Once it fires, the gate signal has no effect. The reverse characteristics of the SCR resemble those of a conventional reverse-biased diode. SCRs must be turned off by interrupting the anode current or reducing it below a minimum value called the holding current IH. When gate current is zero, the minimum holding current is often referred to as the latching current IL.