Thyristors


A thyristor is a semiconductor device acting as an on-off switch triggered by a signal of very small power. The thyr- part of the name comes from the Greek qura, door, alluding to the function of the device as a gate or door. Thyristors are most used to control AC, either as an on-off or static switch, or with phase control allowing the power to be regulated without producing large amounts of heat, or requiring mechanical contacts. The basic thyristor is the silicon controlled rectifier, or SCR, but the usual device for AC control is the triac. In spite of the great usefulness of thyristors in everyday applications, they are not usually covered in an electronics course, except in passing.

The SCR is a pnpn device, as shown schematically in the figure. When the anode A is negative and the cathode K is positive, the SCR will not conduct until its reverse breakdown voltage is reached, like any diode. When A is positive and K is negative, the two pn junctions at anode and cathode are forward-biased, but current is blocked by the pn junction in the middle. Again, the SCR acts like a normal diode, not conducting until it breaks down. However, in this state it can be triggered to conduct freely, the voltage across the device dropping to about that of a forward-biased pn junction.

The figure shows that the SCR can be imagined as a pair of transistors. This is only to show the action of the device--transistors connected this way do not an SCR make. If a little current i is injected into the base of the npn transistor on the bottom, it appears as β1i in the collector. This, however, feeds the base of the pnp transistor, which again amplifies the current so that a current β2β1i flows in its collector lead. This is a feedback loop, and if the gain β2β1 is greater that 1, the currents will go on increasing and the device will conduct from A to K. The triggering action depends on the fact that the β of a transistor is very small for very small base currents, since all the base currents are used up by impurities and defects in the base-emitter junction, and little transistor action occurs. As the base current increases, however, beta quickly increases, and soon breakdown occurs.

There are three ways to provide the necessary current for breakdown. First, simply increasing the voltage across the diode increases the leakage currents, and the SCR will eventually break down, like any diode, but helped by the transistor action. Second, if the voltage across the SCR is increased rapidly, there will be sufficient current charging the reverse-biased pn junction to raise beta enough to cause breakdown. This is called dV/dt breakdown, and is usually unwanted. Finally, a current injected into the n region near the cathode can provide the current needed for triggering. This current can be supplied by an actual conductive connection, or by a photocurrent as in a phototransistor. Light-activated SCR's (LASCR) are available by themselves, or as part of an optocoupler, and can be very convenient because of the isolation of the controlling circuit.

The circuit symbol and the V-I characteristic of an SCR are shown in the figure. With the anode positive, the SCR will trigger when sufficient current flows in the gate, as the gate is made positive with respect to the cathode. Once the SCR has triggered, the gate loses control, and the SCR continues to conduct as long as the current is above some small holding value. If AC is supplied, the SCR must be triggered on every positive half-cycle, since it stops conducting when the voltage reverses. This explains the great utility of the SCR in AC circuits, where turn-off is not a problem because of the polarity reversal. The SCR is valuable as a controlled rectifier, when it is triggered at some selected point in the cycle, called phase control.

A device that contains two SCR's back-to-back, but with a single trigger, is called a triac. The circuit symbol and V-I characteristic are shown in the figure. It is not actually two discrete SCR's, but a very ingenious single crystal with various p and n layers. The triac gives full-wave AC control with no extra components, and is very convenient. The main terminals are no longer called anode and cathode, since there is no polarity, but simply MT1 and MT2. MT1 is associated with the gate, while MT2 is the one associated with the metal tab that helps cool the device in high-power examples. The forward conducting voltage falls only to 0.7V or so, so if high currents are to be handled, the resulting heat must be dissipated. This small voltage drop is of little importance in 120V circuits. In an isolated triac, there is no electrical connection between MT2 and the metal tab, which can be very convenient.

There is a third device, usually made as a pnp, that is meant to break down at a certain voltage in either direction, more or less symmetrically. There is no trigger for this device, which is normally used to trigger a triac itself. This is called a diac, and its symbol is like that for a triac, but without the gate lead. Various symbols are in use. The diac will be seen below in the circuit for phase control.

Circuits

The connections for a small SCR in a TO-92 package are shown at the right, looking at the leads. The connections for a 4A triac in a TO-220 package are shown at the left. Any SCR or triac within reason can be used for these experiments, but be sure to check the pinout. The 2N5060 has a breakdown voltage of 30V (other members of the family have breakdown voltages up to 400V) and can handle an average current of 0.5A. The L2004L3 is a sensitive-gate (3mA will trigger it) triac with isolated cooling tab that will handle 4A rms, with a holding current of 5 mA.

The same simple test circuit can be used for either an SCR or a triac. The LED load makes it obvious when the device turns on. R2 can be varied to study how much trigger current is necessary, and R1 varied to find the holding current. If you reverse the supply voltage (turn the LED around!) the triac circuit will still trigger, but, of course, the SCR will not. The triggering with MT2 positive, as in the figure, is called Quadrant I, while with the supply reversed, it is called Quadrant III. A triac is most sensitive in these two quadrants, but will also respond to gate currents in the other directions. Many modern triacs seem to be symmetrical with respect to MT1 and MT2. To study triggering in Quadrants II and IV, simply interchange the connections of MT1 and MT2.

A static switch, controlled by an optocoupler, is shown in the figure. The MOC3032 optocoupler has an IRED and a small triac with a zero-crossing detector trigger. When the IRED is on, the triac will fire at each zero-crossing of the voltage across it, and the current is used to trigger the main triac. The 33Ω resistor protects the triac against excess current, and does not need to be very large. The triac in the MOC3032 can stand the 120V (actually, it is rated at 250V) when it is not conducting, and when it is, the main triac is also on, so the voltage across it is small. A 1k resistor can also be connected between the gate and MT1 to help turn-off, but it is not shown in the diagram and is not necessary here. This circuit is best studied with AC applied, which can be a 10V amplitude sine wave from a function generator. I used a buffered sine wave from an ICL8038 chip, and it worked excellently. Also, the action of the device can only really be appreciated by using an oscilloscope, so I shall presume you have one available.

Use one channel of the scope to show the applied voltage, and the other to show the voltage across the load, the "neutral" terminal being taken as ground. When there is no current in the optocoupler, the voltage across the load is zero. When you ground node "a" the triac will be triggered, and the load voltage will be the same as the applied voltage, less the drop across the triac, which is easily visible at these voltages. This circuit is essentially that of the well-known "solid state relay" which is just an optocoupler and a triac with a few protective additions. You can make your own solid state relays much more cheaply. Note that the control is completely isolated from the load, but also that the load is not absolutely disconnected when "off." There will always be some small leakage, which may or may not be a bother. Also note that in AC wiring, it is always the "hot" wire, not the grounded one, that must be switched, so that a device that is turned off is not a hazard to someone who may be grounded.

Alternating current can also be applied to the test circuit for the SCR or triac, but without the LED, of course, and only a resistive load. Here, it is convenient to look at the voltage at MT2 with MT1 grounded. Note the triggering point and the drop-out point and the beginning and end of each half-cycle. When the triggering is not absolutely symmetrical in the positive and negative cycles, there will be a DC component in the load (which is normally not very deleterious).

The figure shows a phase control for 120V AC, with some refinements. A diac D is used for triggering. What happens, essentially, is that the capacitor charges up when the diac is nonconducting at the beginning of each half-cycle. The diac fires at 27 to 32V, and the capacitor is dumped into the trigger of the triac T. The potentiometer controls the rate of charging of the capacitor. Note that it is a good idea to use a resistor like the 3.3k one to prevent a problem when the potentiometer is turned to zero ohms. The 15k resistor and 0.1 μF capacitor just in front of the diac provides a second time constant that gives a greater range of control and improved symmetry, both of which are matters of importance in a dimming switch. The resistor and capacitor in series across the triac is called a snubber. It prevents dV/dt from becoming too large with certain kinds of loads (inductive), triggering the triac when it is not wanted. This combination is quite common, but a 39Ω resistor in series with an 0.01 μF capacitor is also seen.

This circuit must be tested with 120V AC, and the load could well be a 60W light bulb. Take great care with such tests with wires in the open, using one hand only near the breadboard, as in the days of vacuum tubes. You could even work with rubber gloves, but that is being extravagant. If you have an isolation transformer, use it. 120V is quite safe if you are not grounded. Fuse your AC supply for a small current, such as 1A, so you do not rely on the 15A fuses in your fusebox if something goes wrong. The rapid switching at each half-cycle generates a large amount of high-frequency noise on the AC line. To eliminate this very annoying interference, use a 100 μH choke in the hot wire and a 0.1 μF capacitor across the line on the triac side. Power input modules can be purchased with built-in RFI filters (but at a cost). The reference gives more circuits and refinements.

Reference

Teccor Electronics, Thyristor Product Catalog (Irving, TX: no date)


Return to Electronics Index

Composed by J. B. Calvert
Created 14 July 2001
Last revised