An up-to-date guide to power semiconductors: how they work and their characteristics. This can help you evaluate and select adjustable-speed drives.
Since the introduction of solid state control of ac motors, a number of power semiconductors have been developed, each with a unique set of characteristics. These semiconductors include the silicon controlled rectifier (SCR), gate turn-off device (GTO), power (giant) transistors (GTR), and insulated gate bipolar transistors (IGBT).
Under normal operating conditions, power semiconductors function as oneway switches; current flows in one direction, from the highest voltage potential to a lower voltage potential. Their three basic connections include terminals for input power, output power and a control terminal. A current or voltage signal applied to the control terminal closes (turns on) the semiconductor switch.
The most appropriate power semiconductor for an application depends on its speed (response to a control signal), gain or amplification, efficiency, and control method (voltage or current). These four basic characteristics are compared in the Table.
With its introduction in the 1950s as a variable voltage-controlling device in acto- dc conversion applications, SCRs are the oldest controllable power semiconductor. GTOs became available for widespread usage in the late 1970s, then GTRs in the early 1980s and, most recently, IGBTs.
As speed of switching increases from SCRs through IGBTs, waveforms are produced that contain increasingly higherfrequency components. This produces correspondingly higher leakage currents, which must be considered in the the circuit design.
Silicon controlled rectifier (SCR)
To easily understand how an SCR works, consider its function in a typical dc drive, which was the first general industrial application. In these drives, SCRs control dc motor speed and torque by converting ac line voltage to adjustable dc voltage. To do this, the controller determines when in the ac half cycle to turn the SCR on so it conducts to the end of that half cycle. The earlier in the half cycle the SCR turns on, the higher the average output voltage is to the motor. To turn the SCR on, the controller applies a current to the SCR gate, Figure 1.
If the load (such as a capacitor) can store energy in the form of a voltage, the SCR can turn on only when the instantaneous value of ac line voltage exceeds the value of stored voltage.
SCRs are capable of handling large amounts of current — 5,000 A to 10,000 A. But large devices are needed to handle high currents; and the larger the device, the slower its switching characteristics.
Typically, the time it takes an SCR to turn fully on after receiving a gate pulse is in the order of 10s of μsec. To turn on, SCRs require a heavy pulse of gate current (about 50 A/μsec ) for a short period of time and then lower-current sustaining pulses to keep it on. If the device is starved (insufficient carriers or electrons to start conduction), high-resistance areas can be developed leading to hot spots and failure. Such failures can also occur if, once turned on, sufficient current through the device is not maintained.
Several approaches are available involving external circuit components (such as saturable reactors) to ensure that the device will safely turn on. Some schemes for turning SCRs on involve continuous pulse firing; others provide a heavy turn-on pulse followed by a sustained pulse to maintain conduction of the SCR.
Turn-on characteristics are the most significant limitation of an SCR in a power-control application. The circuit must be capable of handling the heavy turn-on current, if even for a short time.
Since SCRs are rather slow to turn on and require line commutation to turn off, other devices such as SCR-rated, fastblow fuses are necessary to protect the SCR and turn it off if some event occurs, such as a short or a fast impedance change, while the device is conducting.
In the early 1960s, SCRs were used to control the frequency of ac drives. Now, GTOs, GTRs and IGBTs have replaced SCRs for this function, particularly in low-voltage ac drives, because of their greater efficiency and controllability.
Efficiency of SCRs is fairly low because the voltage drop across the device (typically 2 to 2.5 V) produces high losses when conducting heavy current.
In PWM ac drives and in low-voltage ac motor-control applications, SCRs function as rectifiers to convert ac line voltage to a dc voltage, which, in turn, is connected to the ac motor by other types of semiconductor switches.
Continue on page 2
For dc drives, SCRs are the least expensive power semiconductor for applications below a power range of 200 to 500 hp. In the larger power ranges, GTOs become practical. For ac drives, SCRs are the least expensive power semiconductors, but the drives require many capacitors to commutate the SCRs off to produce an ac output.
Gate turn off device (GTO)
A GTO, like an SCR, can handle high current levels — in the thousands of amperes. Plus, a GTO, Figure 2, is similar to an SCR except that its control connection provides the means to both turn on and turn off the device. Turn on occurs when a positive voltage is applied to the gate. Applying a negative voltage to the gate turns the GTO off and interrupts current flow in the main circuit. Once turned on, a GTO will stay on and does not require sustaining pulses as does an SCR. To turn a GTO off, the negative or turn-off control signal must be greater than the current flowing through the GTO. Typically, turn-off time of a GTO is relatively slow—about 50 μsec. Like SCRs, GTOs require fast-acting, current-limiting fuse to protect against overcurrent situations.
Efficiency of a GTO is relatively low due to the power source required to supply the negative turn-off bias. Because GTOs turn on relatively slowly (like SCRs), high switching losses reduce efficiency.
Giant transistor (GTR)
A GTR differs from an SCR and a GTO in that the control or base signal must be maintained to keep the device turned on. How much control signal is needed depends on how the GTR is operated. For large-gain, low-loss operation, the control signal must be large. For large-gain, large-loss operation, the control signal can be small.
Current-carrying capability of a GTR is in the range of 1,000 to 2,000 A.
Switching speed of a GTR is fast, about 5 μsec. Because GTRs are low-gain devices (as are all transistors), they are typically mounted with one device feeding another in a Darlington configuration, Figure 3.
During the 1980s, GTRs became popular for control of ac motor speed and torque because of their ability to respond rapidly to control input signals. The GTR was responsible for advancing the initial commercial success of PWM ac drives.
A GTR’s gain depends on the component’s design and the resulting voltage drop across the device. For example, with a drop of about 2 V, the IGBT’s gain would be about five. But, if the drop is two or three times greater (4 to 7 V), a gain of 200 would be produced. Higher voltage across the device eliminates the need for a large base driver or amplification of the control signal to turn the device on. Thus, the control package can be reduced, but a larger heat sink is required.
Lower voltage drop across a GTR means that relatively larger values of base-driver current are required. For instance, for a device with a gain of five, a 10-A base-driver current is required for a 50-A output. The problem with this approach is that inductance in the circuit can delay power-current flow affecting the switching characteristics of the device. Even a switching delay of a few microseconds per ampere can adversely affect critical applications where high values of current must be supplied in a short time.
Although GTRs are dependable components for use in the manufacture of ac drives, they are gradually being replaced by IGBTs. There are no significant technical reasons for this trend, but the faster switching speeds of an IGBT produce a so-called “quiet ac motor controller.” Commercial pressure for smaller controller packages and a simpler control-topower interface is forcing the early retirement of GTRs for ac motor control.
Insulated-gate bipolar transistors (IGBT)
IGBTs , Figure 4, differ from SCRs, GTOs and GTRs by the control scheme. To turn on an IGBT, a voltage is applied to its gate. The MOSFET-type input converts the voltage to the current required to turn on the output portion (transistor) of the IGBT.
Thus, the interface from control to power is simplified because it requires very little control current to handle large amounts of power current. Because control current is small, time delays associated with large values of control current are avoided. Thus, a fast control change produces a fast power change.
Current-carrying capability of an IGBT is in the same range as that of a GTR — 1,000 to 2,000 A.
An IGBT’s fast response to signal changes — less than 1 μsec — reduces audible levels in an ac motor while controlling torque and speed. And, its high switching frequency (carrier frequency) provides a responsive control of current. Also, an IGBT’s low losses result in compact packaging of the ac motor controller.
Since the introduction of IGBTs, concerns have arisen about the effects of high-frequency switching and harmonicrich waveforms during transmission of power to an ac motor. Reflected waves, RFI and grounding issues are current technical hot buttons and, as IGBTs are used more widely, more issues will surface. However, these will be resolved through an in-depth understanding of how an IGBT functions and how surrounding components respond to it.
Howard G. Murphy, P.E., is manager of applications for standard drives at Allen-Bradley Co., Inc., Mequon, Wisc.