Text preview for : Power Control with Thyristors and Triacs.pdf part of Philips Semiconductors Thyristors and Triacs Semiconductor Applications. (Thyristors and Triacs)
Back to : Power Control with Thyris | Home
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
CHAPTER 6
Power Control with Thyristors and Triacs
6.1 Using Thyristors and Triacs 6.2 Thyristor and Triac Applications 6.3 Hi-Com Triacs
485
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Using Thyristors and Triacs
487
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
6.1.1 Introduction to Thyristors and Triacs
Brief summary of the thyristor family
The term thyristor is a generic name for a semiconductor switch having four or more layers and is, in essence, a p-n-p-n sandwich. Thyristors form a large family and it is helpful to consider the constituents which determine the type of any given thyristor. If an ohmic connection is made to the first p region and the last n region, and no other connection is made, the device is a diode thyristor. If an additional ohmic connection is made to the intermediate n region (n gate type) or the intermediate p region (p gate type), the device is a triode thyristor. If an ohmic connection is made to both intermediate regions, the device is a tetrode thyristor. All such devices have a forward characteristic of the general form shown in Fig. 1. There are three types of thyristor reverse characteristic: blocking (as in normal diodes), conducting (large reverse currents at low reverse voltages) and approximate mirror image of the forward characteristic (bidirectional thyristors). Reverse blocking devices usually have four layers or less whereas reverse conducting and mirror image devices usually have five layers. The simplest thyristor structure, and the most common, is the reverse blocking triode thyristor (usually simply referred to as the 'thyristor' or SCR 'silicon controlled rectifier'). Its circuit symbol and basic structure are shown in Fig. 2. The most complex common thyristor structure is the bidirectional triode thyristor, or triac. The triac (shown in Fig. 3) is able to pass current bidirectionally and is therefore an a.c. power control device. Its performance is that of a pair of thyristors in anti-parallel with a single gate terminal. The triac needs only one heatsink, but this must be large enough to remove the heat caused by bidirectional current flow. Triac gate triggering circuits must be designed with care to ensure that unwanted conduction, ie. loss of control, does not occur when triggering lasts too long. Thyristors and triacs are both bipolar devices. They have very low on-state voltages but, because the minority charge carriers in the devices must be removed before they can block an applied voltage, the switching times are comparatively long. This limits thyristor switching circuits to low frequency applications. Triacs are used almost exclusively at mains supply frequencies of 50 or 60Hz, while in some applications this extends up to the 400Hz supply frequency as used in aircraft. The voltage blocking capabilities of thyristors and triacs are quite high: the highest voltage rating for the Philips range is 800V, while the currents (IT(RMS)) range from 0.8A to 25A.
Gate
The devices are available as surface mount components, or as non-isolated or isolated discrete devices, depending on the device rating.
Forward current On-state characteristic
I G> 0 Reverse voltage Avalanche breakdown region IL IH
IG = 0 Forward voltage
V (BO) Reverse characteristic Off-state characteristic
Reverse current
Fig. 1 Thyristor static characteristic
Anode
Anode
p Gate n p n J1 J2 J3
Cathode
Cathode
Fig. 2 Thyristor circuit symbol and basic structure
MT1
MT1
Gate
Gate
n p n n p
n
MT2
MT2
Fig. 3 Triac circuit symbol and basic structure
489
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Thyristor operation
The operation of the thyristor can be understood from Fig. 2. When the thyristor cathode is more positive than the anode then junctions J1 and J3 are reverse biased and the device blocks. When the anode is more positive than the cathode, junctions J1 and J3 are forward biased. As J2 is reverse biased, then the device still blocks forward voltage. If the reverse voltage across J2 is made to reach its avalanche breakdown level then the device conducts like a single forward-biased junction. The 'two transistor' model of Fig. 4 can be used to consider the p-n-p-n structure of a thyristor as the interconnection of an npn transistor T1 and a pnp transistor T2. The collector of T1 provides the base current for T2. Base current for T1 is provided by the external gate current in addition to the collector current from T2. If the gain in the base-collector loop of T1 and T2 exceeds unity then the loop current can be maintained regeneratively. When this condition occurs then both T1 and T2 are driven into saturation and the thyristor is said to be 'latched'. The anode to cathode current is then only limited by the external circuit.
Anode i A
Thyristor turn-on methods
Turn-on by exceeding the breakover voltage
When the breakover voltage, VBO, across a thyristor is exceeded, the thyristor turns on. The breakover voltage of a thyristor will be greater than the rated maximum voltage of the device. At the breakover voltage the value of the thyristor anode current is called the latching current, IL. Breakover voltage triggering is not normally used as a triggering method, and most circuit designs attempt to avoid its occurrence. When a thyristor is triggered by exceeding VBO the fall time of the forward voltage is quite low (about 1/20th of the time taken when the thyristor is gate-triggered). As a general rule, however, although a thyristor switches faster with VBO turn-on than with gate turn-on, the permitted di/dt for breakover voltage turn-on is lower.
Turn-on by leakage current
As the junction temperature of a thyristor rises, the leakage current also increases. Eventually, if the junction temperature is allowed to rise sufficiently, leakage current would become large enough to initiate latching of the regenerative loop of the thyristor and allow forward conduction. At a certain critical temperature (above Tj(max)) the thyristor will not support any blocking voltage at all.
T2
Turn-on by dV/dt
Gate i G Cathode T1
Any p-n junction has capacitance - the larger the junction area the larger the capacitance. If a voltage ramp is applied across the anode-to-cathode of a p-n-p-n device, a current will flow in the device to charge the device capacitance according to the relation:
Fig. 4 'Two transistor' model of a thyristor
iC = C.
There are several mechanisms by which a thyristor can be latched. The usual method is by a current applied to the gate. This gate current starts the regenerative action in the thyristor and causes the anode current to increase. The gains of transistors T1 and T2 are current dependent and increase as the current through T1 and T2 increases. With increasing anode current the loop gain increases sufficiently such that the gate current can be removed without T1 and T2 coming out of saturation. Thus a thyristor can be switched on by a signal at the gate terminal but, because of the way that the current then latches, the thyristor cannot be turned off by the gate. The thyristor must be turned off by using the external circuit to break the regenerative current loop between transistors T1 and T2. Reverse biasing the device will initiate turn-off once the anode current drops below a minimum specified value, called the holding current value, IH.
dv dt
(1)
If the charging current becomes large enough, the density of moving current carriers in the device induces switch-on.
Turn-on by gate triggering
Gate triggering is the usual method of turning a thyristor on. Application of current to the thyristor gate initiates the latching mechanism discussed in the previous section. The characteristic of Fig. 1 showed that the thyristor will switch to its on-state condition with forward bias voltages less than VBO when the gate current is greater than zero. The gate current and voltage requirements which ensure triggering of a particular device are always quoted in the device data. As thyristor triggering characteristics are temperature dependant, the amplitude and duration of the gate pulse must be sufficient to ensure that the thyristor latches under all possible conditions.
490
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
During gate turn-on, the rate of rise of thyristor anode current dIF/dt is determined by the external circuit conditions. However, the whole active area of the thyristor (or triac) cannot be turned on simultaneously: the area nearest to the gate turns on first, followed by the remainder of the device. At turn-on it is important that the rate of rise of current does not exceed the specified rating. If dIF/dt is excessive then only a limited area of the device will have been turned on as the anode current increases. The resulting localised heating of the device will cause degradation and could lead to eventual device failure. A suitably high gate current and large rate of rise of gate current (dIG/dt) ensures that the thyristor turns on quickly (providing that the gate power ratings are not exceeded) thus increasing the thyristor turn-on di/dt capability. Once the thyristor has latched then the gate drive can be reduced or removed completely. Gate power dissipation can also be reduced by triggering the thyristor using a pulsed signal.
Forward current
On-state
T2+
IG > 0 Reverse voltage Off-state V (BO) IL IH IG = 0 Forward voltage IH IL IG = 0 IG > 0 V (BO) Off-state
T2On-state
Reverse current
Fig. 6 Triac static characteristic
Quadrant 1 2 3 4 (1+) (1-) (3-) (3+)
Polarity of MT2 wrt MT1 MT2+ MT2+ MT2MT2-
Gate polarity G+ GGG+
Triac operation
The triac can be considered as two thyristors connected in antiparallel as shown in Fig. 5. The single gate terminal is common to both thyristors. The main terminals MT1 and MT2 are connected to both p and n regions of the device and the current path through the layers of the device depends upon the polarity of the applied voltage between the main terminals. The device polarity is usually described with reference to MT1, where the term MT2+ denotes that terminal MT2 is positive with respect to terminal MT1.
MT2
Table 1. Operating quadrants for triacs
MT2+ Quadrant 2 + + Quadrant 1
I
G
I -
G G+
G-
MT1
Fig. 5 Anti parallel thyristor representation of a triac
I
The on-state characteristic of the triac is similar to that of a thyristor and is shown in Fig. 6. Table 1 and Fig. 7 summarise the different gate triggering configurations for triacs. Due to the physical layout of the semiconductor layers in a triac, the values of latching current (IL), holding current (IH) and gate trigger current (IGT) vary slightly between the different operating quadrants. In general, for any triac, the latching current is slightly higher in the second (MT2+, G-) quadrant than the other quadrants, whilst the gate trigger current is slightly higher in fourth (MT2-, G+) quadrant.
G
I + MT2+
G
Quadrant 3
Quadrant 4
Fig. 7 Triac triggering quadrants For applications where the gate sensitivity is critical and where the device must trigger reliably and evenly for applied voltages in both directions it may be preferable to use a negative current triggering circuit. If the gate drive circuit is arranged so that only quadrants 2 and 3 are used (i.e. Goperation) then the triac is never used in the fourth quadrant where IGT is highest.
491
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
For some applications it is advantageous to trigger triacs with a pulsating signal and thus reduce the gate power dissipation. To ensure bidirectional conduction, especially with a very inductive load, the trigger pulses must continue until the end of each mains half-cycle. If single trigger pulses are used, one-way conduction (rectification) results when the trigger angle is smaller than the load phase angle. Philips produce ranges of triacs having the same current and voltage ratings but with different gate sensitivities. A device with a relatively insensitive gate will be more immune to false triggering due to noise on the gate signal and also will be more immune to commutating dv/dt turn-on. Sensitive gate triacs are used in applications where the device is driven from a controller IC or low power gate circuit.
Gate voltage, V G (V)
Gate power ratings Gate-cathode characteristic
Failure to trigger V GT
PGM(max) 5W = = 0.1 P = 0.5W G(AV) IGT Gate current, I G (A) = 1.0
Fig. 9 Thyristor gate characteristic The gate triggering characteristic is limited by the gate power dissipation. Figure 9 also shows the continuous power rating curve (PG(AV)=0.5W) for a typical device and the peak gate power curve (PGM(max)=5W). When designing a gate circuit to reliably trigger a triac or thyristor the gate signal must lie on a locus within the area of certain device triggering. Continuous steady operation would demand that the 0.5W curve be used to limit the load line of the gate drive circuit. For pulsed operation the triggering locus can be increased. If the 5W peak gate power curve is used, the duty cycle must not exceed
max = PG(AV) 0.5 = 0.1 = PGM 5 (2)
The diac
It is also worthwhile to consider the operation and characteristics of the diac in the context of multilayer bipolar devices. The diac is more strictly a transistor than a thyristor, but has an important role in many thyristor and triac triggering circuits. It is manufactured by diffusing an n-type impurity into both sides of a p-type slice to give a two terminal device with symmetrical electrical characteristics. As shown in the characteristic of Fig. 8, the diac blocks applied voltages in either direction until the breakover voltage, VBO is reached. The diac voltage then breaks back to a lower output voltage VO. Important diac parameters are breakover voltage, breakover current and breakback voltage as shown in the figure.
Breakback voltage
Forward current
I Reverse voltage V (BO) V O
(BO)
Forward voltage VO I (BO) V (BO)
At the other end of the scale, the level below which triggering becomes uncertain is determined by the minimum number of carriers needed in the gate-cathode junction to bring the thyristor into conduction by regenerative action. The trigger circuit load line must not encroach into the failure to trigger region shown in Fig. 9 if triggering is to be guaranteed. The minimum voltage and minimum current to trigger all devices (VGT and IGT) decreases with increasing temperature. Data sheets for Philips thyristors and triacs show the variation of VGT and IGT with temperature.
Reverse current
Thyristor commutation
A thyristor turns off by a mechanism known as 'natural turn-off', that is, when the main anode-cathode current drops below the holding value. It is important to remember, however, that the thyristor will turn on again if the reapplied forward voltage occurs before a minimum time period has elapsed; this is because the charge carriers in the thyristor at the time of turn-off take a finite time to recombine. Thyristor turn-off is achieved by two main methods - self commutation or external commutation.
Fig. 8 Diac static characteristic and circuit symbol
Gate requirements for triggering
To a first approximation, the gate-to-cathode junction of a thyristor or triac acts as a p-n diode. The forward characteristic is as shown in Fig. 9. For a given thyristor type there will be a spread in forward characteristics of gate junctions and a spread with temperature.
492
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Self Commutation
In self-commutation circuits the thyristor will automatically turn off at a predetermined time after triggering. The thyristor conduction period is determined by a property of the commutation circuit, such as the resonant cycle of an LC-circuit or the Volt-Second capability of a saturable inductor. The energy needed for commutation is delivered by a capacitor included in the commutation circuit.
+
I
thyristor
L E C R R leakage
LC circuit in series with the thyristor
When the thyristor is triggered, the resulting main current excites the resonant circuit. After half a resonant cycle, the LC circuit starts to reverse the anode current and turns the thyristor off. The thyristor conduction interval is half a resonant cycle. It is essential for proper commutation that the resonant circuit be less than critically damped. Fig. 10 shows the circuit diagram and the relevant waveforms for this arrangement.
LC Circuit in parallel with the thyristor
Initially the capacitor charges to the supply voltage. When the thyristor is triggered the load current flows but at the same time the capacitor discharges through the thyristor in the forward direction. When the capacitor has discharged (i.e. after one resonant half-cycle of the LC circuit), it begins to charge in the opposite direction and, when this charging current is greater than the thyristor forward current, the thyristor turns off. The circuit diagram and commutation waveforms are shown in Fig. 11.
Fig. 10 Commutation using a series LC circuit
IR + I thyristor L E C R
External commutation
If the supply is an alternating voltage, the thyristor can conduct only during the positive half cycle. The thyristor naturally switches off at the end of each positive half cycle. The circuit and device waveforms for this method of commutation are shown in Fig. 12. It is important to ensure that the duration of a half cycle is greater than the thyristor turn-off time.
Reverse recovery
In typical thyristors the reverse recovery time is of the order of a few micro-seconds. This time increases with increase of forward current and also increases as the forward current decay rate, dIT/dt, decreases. Reverse recovery time is the period during which reverse recovery current flows (t1 to t3 in Fig. 13) and it is the period between the point at which forward current ceases and the earliest point at which the reverse recovery current has dropped to 10% of its peak value.
Fig. 11 Commutation using a parallel LC circuit
493
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Reverse recovery current can cause high values of turn-on current in full-wave rectifier circuits (where thyristors are used as rectifying elements) and in certain inverter circuits. It should also be remembered that, if thyristors are connected in series, the reverse voltage distribution can be seriously affected by mismatch of reverse recovery times.
i thyristor
Turn-off time
Turn-off time is the interval between the instant when thyristor current reverses and the point at which the thyristor can block reapplied forward voltage (t1 to t4 in Fig. 13). If forward voltage is applied to a thyristor too soon after the main current has ceased to flow, the thyristor will turn on. The circuit commutated turn-off time increases with: -junction temperature -forward current amplitude -rate of fall of forward current -rate of rise of forward blocking voltage -forward blocking voltage. Thus the turn-off time is specified for defined operating conditions. Circuit turn-off time is the turn-off time that the circuit presents to the thyristor; it must, of course, be greater than the thyristor turn-off time.
R
Triac commutation
Unlike the thyristor, the triac can conduct irrespective of the polarity of the applied voltage. Thus the triac does not experience a circuit-imposed turn-off time which allows each anti-parallel thyristor to fully recover from its conducting state as it is reverse biased. As the voltage across the triac passes through zero and starts to increase, then the alternate thyristor of the triac can fail to block the applied voltage and immediately conduct in the opposite direction. Triac-controlled circuits therefore require careful design in order to ensure that the triac does not fail to commutate (switch off) at the end of each half-cycle as expected. It is important to consider the commutation performance of devices in circuits where either dI/dt or dV/dt can be large. In resistive load applications (e.g. lamp loads) current surges at turn-on or during temporary over-current conditions may introduce abnormally high rates of change of current which may cause the triac to fail to commutate. In inductive circuits, such as motor control applications or circuits where a dc load is controlled by a triac via a bridge rectifier, it is usually necessary to protect the triac against unwanted commutation due to dv(com)/dt.
dV D dt
Fig. 12 Thyristor commutation in an a.c. circuit
I
T
dI T dt
I
R
V D
V R
t
0
t t 1 2
t
3
t
4
Fig. 13 Thyristor turn-off characteristics
The commutating dv(com)/dt limit for a triac is less than the static dv/dt limit because at commutation the recently conducting portion of the triac which is being switched off has introduced stored charge to the triac. The amount of stored charge depends upon the reverse recovery characteristics of the triac. It is significantly affected by junction temperature and the rate of fall of anode current prior to commutation (dI(com)/dt). Following high rates of change of current the capacity of the triac to withstand high reapplied rates of change of voltage is reduced. Data sheet specifications for triacs give characteristics showing the
494
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
maximum allowable rate of rise of commutating voltage against device temperature and rate of fall of anode current which will not cause a device to trigger.
Current
Supply voltage Load current
Time dV /dt com VDWM Time
The usual method is to place a dv/dt-limiting R-C snubber in parallel with the triac. Additionally, because commutating dv/dt turn-on is dependent upon the rate of fall of triac current, then in circuits with large rates of change of anode current, the ability of a triac to withstand high rates of rise of reapplied voltage is improved by limiting the di/dt using a series inductor. This topic is discussed more fully in the section entitled 'Using thyristors and triacs'.
-dI/dt Voltage across triac Trigger pulses
Conclusions
This article has presented the basic parameters and characteristics of triacs and thyristors and shown how the structure of the devices determines their operation. Important turn-on and turn-off conditions and limitations of the devices have been presented in order to demonstrate the capabilities of the devices and show the designer those areas which require careful consideration. The device characteristics which determine gate triggering requirements of thyristors and triacs have been presented. Subsequent articles in this chapter will deal with the use, operation and limitations of thyristors and triacs in practical applications, and will present some detailed design and operational considerations for thyristors and triacs in phase control and integral cycle control applications.
Time
Fig. 14 Inductive load commutation with a triac Consider the situation when a triac is conducting in one direction and the applied ac voltage changes polarity. For the case of an inductive load the current in the triac does not fall to its holding current level until some time later. This is shown in Fig. 14. At the time that the triac current has reached the holding current the mains voltage has risen to some value and so the triac must immediately block that voltage. The rate of rise of blocking voltage following commutation (dv(com)/dt) can be quite high.
495
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
6.1.2 Using Thyristors and Triacs
This chapter is concerned with the uses, operation and protection of thyristors and triacs. Two types of circuit cover the vast majority of applications for thyristors and triacs: static switching circuits and phase control circuits. The characteristics and uses of these two types of circuit will be discussed. Various gate drive circuits and protection circuits for thyristor and triacs are also presented. The use of these circuits will enable designers to operate the devices reliably and within their specified limits. Circuit-generated RFI can be almost completely eliminated by ensuring that the turn-on switching instants correspond to the zero-crossing points of the a.c. mains supply. This technique is known as synchronous (or zero voltage) switching control as opposed to the technique of allowing the switching points to occur at any time during the a.c. cycle, which is referred to as asynchronous control. In a.c. circuits using thyristors and triacs the devices naturally switch off when the current falls below the device holding current. Thus turn-off RFI does not occur.
Thyristor and triac control techniques
There are two main techniques of controlling thyristors and triacs - on-off triggering (or static switching) and phase control. In on-off triggering, the power switch is allowed to conduct for a certain number of half-cycles and then it is kept off for a number of half-cycles. Thus, by varying the ratio of "on-time" to "off-time", the average power supplied to the load can be controlled. The switching device either completely activates or deactivates the load circuit. In phase control circuits, the thyristor or triac is triggered into conduction at some point after the start of each half-cycle. Control is achieved on a cycle-by-cycle basis by variation of the point in the cycle at which the thyristor is triggered.
Asynchronous control
In asynchronous control the thyristor or triac may be triggered at a point in the mains voltage other than the zero voltage crossover point. Asynchronous control circuits are usually relatively cheap but liable to produce RFI.
Synchronous control
In synchronous control systems the switching instants are synchronised with zero crossings of the supply voltage. They also have the advantage that, as the thyristors conduct over complete half cycles, the power factor is very good. This method of power control is mostly used to control temperature. The repetition period, T, is adjusted to suit the controlled process (within statutory limits). Temperature ripple is eliminated when the repetition period is made much smaller than the thermal time constant of the system. Figure 1 shows the principle of time-proportional control. RFI and turn-on di/dt are reduced, and the best power factor (sinusoidal load current) is obtained by triggering synchronously. The average power delivered to a resistive load, RL, is proportional to ton/T (i.e. linear control) and is given by equation 1.
Static switching applications
Thyristors and triacs are the ideal power switching devices for many high power circuits such as heaters, enabling the load to be controlled by a low power signal, in place of a relay or other electro-mechanical switch. In a high power circuit where the power switch may connect or disconnect the load at any point of the mains cycle then large amounts of RFI (radio frequency interference) are likely to occur at the instants of switching. The large variations in load may also cause disruptions to the supply voltage. The RFI and voltage variation produced by high power switching in a.c. mains circuits is unacceptable in many environments and is controlled by statutory limits. The limits depend upon the type of environment (industrial or domestic) and the rating of the load being switched. RFI occurs at any time when there is a step change in current caused by the closing of a switch (mechanical or semiconductor). The energy levels of this interference can be quite high in circuits such as heating elements. However, if the switch is closed at the moment the supply voltage passes through zero there is no step rise in current and thus no radio frequency interference. Similarly, at turn-off, a large amount of high frequency interference can be caused by di/dt imposed voltage transients in inductive circuits.
Pout =
2 V(RMS) ton . RL T
(1)
where: T is the controller repetition period ton is controller 'on' time V(RMS) is the rms a.c. input voltage. Elsewhere in this handbook the operation of a controller i.c. (the TDA1023) is described. This device is specifically designed to implement time-proportional control of heaters using Philips triacs.
497
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Triac
Supply voltage
Input voltage
Voltage
O = wt
Trigger signal
Triac Current
Output current
O = wt
tON T
Trigger
Fig. 1 Synchronous time-proportional control
O = wt
Phase control
Phase control circuits are used for low power applications such as lamp control or universal motor speed control, where RFI emissions can be filtered relatively easily. The power delivered to the load is controlled by the timing of the thyristor (or triac) turn-on point. The two most common phase controller configurations are 'half wave control', where the controlling device is a single thyristor and 'full wave control', where the controlling device is a triac or a pair of anti-parallel thyristors. These two control strategies are considered in more detail below:
Device triggers
Trigger angle,
Conduction angle,
Fig. 2 Phase controller - resistive load
IT(AV) = 2.IT(MAX) = 0.637 IT(MAX) IT(MAX)
IT(RMS) =
2
= 0.707 IT(MAX)
(2)
where
IT(MAX) = VT(MAX) 2 V(RMS) = RL RL (3)
Resistive loads
The operation of a phase controller with a resistive load is the simplest situation to analyse. Waveforms for a full wave controlled resistive load are shown in Fig. 2. The triac is triggered at angle , and applies the supply voltage to the load. The triac then conducts for the remainder of the positive half-cycle, turning off when the anode current drops below the holding current, as the voltage becomes zero at =180°. The triac is then re-triggered at angle (180+)°, and conducts for the remainder of the negative half-cycle, turning off when its anode voltage becomes zero at 360°. The sequence is repeated giving current pulses of alternating polarity which are fed to the load. The duration of each pulse is the conduction angle , that is (180-)°. The output power is therefore controlled by variation of the trigger angle . For all values of other than =180° the load current is non-sinusoidal. Thus, because of the generation of harmonics, the power factor presented to the a.c. supply will be less than unity except when =0. For a sinusoidal current the rectified mean current, IT(AV), and the rms current, IT(RMS), are related to the peak current, IT(MAX), by equation 2.
From equation 2 the 'crest factor', c, (also known as the 'peak factor') of the current waveform is defined as:
Crest factor, c = IT(MAX) IT(RMS) (4)
The current 'form factor,' a, is defined by:
Form factor, a = IT(RMS) IT(AV) (5)
Thus, for sinusoidal currents:
a= IT(RMS) = 1.111; IT(AV) c= IT(MAX) = 1.414 IT(RMS) (6)
For the non-sinusoidal waveforms which occur in a phase controlled circuit, the device currents are modified due to the delay which occurs before the power device is triggered. The crest factor of equation 4 and the form factor of equation 5 can be used to describe variation of the current waveshape from the sinusoidal case.
498
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Half wave controller
Figure 3a) shows the simplest type of thyristor half-wave phase controller for a resistive load. The load current waveform is given in Fig. 3b). The variation of average load current, IT(AV), rms load current, IT(RMS) and load power over the full period of the a.c mains, with trigger angle are given in equation 7.
IT(AV) = IT(AV) max. (1 - cos) 2
IT(AV) max = IT(MAX)
a)
b)
Supply voltage
Thyristor Voltage
I T(MAX) Thyristor Current
Trigger
1
- 1 sin 2 2 2 IT(RMS) = IT(RMS) max.
IT(RMS) max =
IT(MAX) 2
a)
Fig. 3 Half wave control
b)
- 1 sin 2 2 P(out) = P(out) max.
P(out) max =
I
2 T(MAX) L
Supply voltage
R
4
(7)
Triac Voltage
N.B. When using equation 7 all values of must be in radians. For each case the maximum value occurs when =180° (= radians). At =180° the crest factor and form factor for a half wave controller are given by:
a= IT(RMS) = 1.571; IT(AV) c= IT(MAX) = 2.0 IT(RMS) (8)
I
T(MAX) Triac Current
I
T(MAX) Trigger
Fig. 4 Full wave control The variation of normalised average current, IT(AV)/IT(AV)max, rms current IT(RMS)/IT(RMS)max, and power, P(out)/P(out)max, for equations 7 and 9 are plotted in Fig. 5. Figure 6 shows the variation of current form factor with conduction angle for the half wave controller and the full wave controller of Figs. 3 and 4.
1
Full wave controller
Figure 4 shows the circuit and load current waveforms for a full-wave controller using two antiparallel thyristors, or a triac, as the controlling device. The variation of rectified mean current, IT(AV), rms current, IT(RMS), and load power with trigger angle are given by equation 9.
IT(AV) = IT(AV) max. (1 - cos) 2
IT(AV) max = 2IT(MAX)
0.8
1
Amplitude
0.6
- 1 sin 2 2 2 IT(RMS) = IT(RMS) max.
IT(RMS) max =
IT(MAX)
0.4
2
0.2
- sin 2 2 P(out) = P(out) max.
1
P(out) max =
2 IT(MAX)RL 2
0
(9)
0 I T(AV) I
30
60 I T(RMS) IT(RMS)max
90 P (OUT) P (OUT)max
120
150 Conduction angle
180
N.B. When using equation 9 all value of must be in radians. For each case the maximum value occurs when =180° (= radians).
T(AV)max
Fig. 5 Current and power control using conduction angle
499
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
6 Form factor 5 4 3 Half-wave rectifier 2 Full-wave rectifier 1 0
Triac Voltage
Supply voltage
I T(RMS) I T(AV)
Triac Current
Inductor iron core saturation
0
30
60
90
120
150 Conduction angle
180
Trigger
Fig. 6 Variation of form factor with conduction angle
Device triggers
Device fails to trigger
Conduction angle
Fig. 7 Triac triggering signals - single pulse
Inductive loads
The circuit waveforms for a phase controller with an inductive load or an active load (for example, a motor) are more complex than those for a purely resistive load. The circuit waveforms depend on the load power factor (which may be variable) as well as the triggering angle. For a bidirectional controller (i.e triac or pair of anti-parallel thyristors), maximum output, that is, sinusoidal load current, occurs when the trigger angle equals the phase angle. When the trigger angle, , is greater than the load phase angle, , then the load current will become discontinuous and the triac (or thyristor) will block some portion of the input voltage until it is retriggered. If the trigger angle is less than the phase angle then the load current in one direction will not have fallen back to zero at the time that the device is retriggered in the opposite direction. This is shown in Fig. 7. The triac fails to be triggered as the gate pulse has finished and so the triac then acts as a rectifier. In Fig. 7 the triac is only triggered by the gate pulses when the applied supply voltage is positive (1+ quadrant). However, the gate pulses which occur one half period later have no effect because the triac is still conducting in the opposite direction. Thus unidirectional current flows in the main circuit, eventually saturating the load inductance. This problem can be avoided by using a trigger pulse train as shown in Fig. 8. The triac triggers on the first gate pulse after the load current has reached the latching current IL in the 3+ quadrant. The trigger pulse train must cease before the mains voltage passes through zero otherwise the triac will continue to conduct in the reverse direction.
Triac Voltage
Supply voltage
Triac Current
Fails to trigger Trigger
Device triggers
Device triggers
Conduction angle
Fig. 8 Triac triggering signals - pulse train
Gate circuits for thyristors and triacs
As discussed in the introductory article of this chapter, a thyristor or triac can be triggered into conduction when a voltage of the appropriate polarity is applied across the main terminals and a suitable current is applied to the gate. This can be achieved using a delay network of the type shown in Fig. 9a). Greater triggering stability and noise immunity can be achieved if a diac is used (see Fig. 9b). This gives a trigger circuit which is suitable for both thyristors and triacs. Figure 10 shows several alternative gate drive circuits suitable for typical triac and thyristor applications. In each circuit the gate-cathode resistor protects the device from false triggering due to noise - this is especially important for sensitive gate devices. In addition opto-isolated thyristor and triac drivers are available which are compatible with the Philips range of devices.
500
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Load 12V 220R
Load 180R BC337 1k0 BT145 10k 1k0 BT145
Load 12V 4k7 100nF 12V
Load 2:1 BAW62 BT145 1k0 1k0 BT145
10k
BC337
10k
Fig. 10 Alternative triac triggering circuits
a)
+
E
IR
the Q1/R2/R3 stage is that the BC547 is on at all instants in time when the applied voltage waveform is high and thus holds the BT169 off. If the BT169 is off then no gate signal is applied to the triac and the load is switched off.
IR
R
+
BT169
BT151
b)
IR
E
R
R
Fig. 11 Master-slave thyristor triggering circuit Fig. 9 Basic triac triggering circuits In some applications it may be necessary to cascade a sensitive gate device with a larger power device to give a sensitive gate circuit with a high power handling capability. A typical solution which involves triggering the smaller device (BT169) from a logic-level controller to turn on the larger device (BT151) is shown in Fig. 11. Figure 12 shows an isolated triac triggering circuit suitable for zero voltage switching applications. This type of circuit is also known as a solid state relay (SSR). The function of
R2 R4 Q1 R3 BC547 1K0 BT169 100R
R1
+
BT138
100R 100nF
Fig. 12 Opto-isolated triac triggering circuit
501
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
If the input signal is switched high then the photo-transistor turns on. If this occurs when the mains voltage is high then Q1 remains on. When the line voltage passes through its next zero crossing in either direction the photo transistor ensures that Q1 stays off long enough for the BT169 to trigger. This then turns the triac on. Once the thyristor turns on, the drive circuit is deprived of its power due to the lower voltage drop of the BT169. The triac is retriggered every half cycle.
commutating dv(com)/dt. A small choke in the a.c circuit will limit the di(com)/dt to an acceptable level. An alternative topology which avoids triac commutation problems is to control the load on the d.c. side.
Snubber networks
Snubber networks ensure that the device is not exposed to excessive rates of change of voltage during transient conditions. This is particularly important when considering the commutation behaviour of triacs, which has been discussed elsewhere.
Voltage transient protection
There are three major sources of transient which may affect thyristor and triac circuits: -the mains supply (e.g. lightning) -other mains and load switches (opening and closing) -the rectifying and load circuit (commutation) In order to ensure reliable circuit operation these transients must be suppressed by additional components, removed at source or allowed for in component ratings. Three types of circuit are commonly employed to suppress voltage transients - a snubber network across the device, a choke between the power device and external circuit or an overvoltage protection such as a varistor.
Choke
Load Varistor Snubber
Fig. 13 Triac protection The following equations can be used to calculate the values of the snubber components required to keep the reapplied dv/dt for a triac within the dv(com)/dt rating for that device. The parameters which affect the choice of snubber components are the value of load inductance, frequency of the a.c. supply and rms load current. The value of the snubber resistor needs to be large enough to damp the circuit and avoid voltage overshoots. The snubber capacitor should be rated for the full a.c. voltage of the system. The snubber resistor needs to be rated at 0.5W. For circuits where the load power factor, cos, 0.7 the snubber values are given approximately by:
fIT(RMS) 2 C 25L dV(com)/dt R=
Series line chokes
A series choke may be used to limit peak fault currents to assist in the fuse protection of thyristors and triacs. If the choke is used in conjunction with fuse protection, it must retain its inductance to very large values of current, and so for this reason it is usually an air-cored component. Alternatively, if the choke is only required to reduce the dv/dt across non-conducting devices then the inductance needs only to be maintained up to quite low currents. Ferrite-cored chokes may be adequate provided that the windings are capable of carrying the full-load current. Usually only a few microhenries of inductance are required to limit the circuit di/dt to an acceptable level. This protects the devices from turning on too quickly and avoids potential device degradation. For instance, a 220V a.c. supply with 20µH source inductance gives a maximum di/dt of (2202)/20=16A/µs. Chokes used to soften commutation should preferably be saturable so as to maintain regulation and avoid deterioration of the power factor. As their impedance reduces at high current, they have very little effect on the inrush current. The addition of di/dt limiting chokes is especially important in triac circuits where the load is controlled via a bridge rectifier. At the voltage zero-crossing points the conduction transfers between diodes in the bridge network, and the rate of fall of triac current is limited only by the stray inductance in the a.c. circuit. The large value of commutating di/dt may cause the triac to retrigger due to
C
3L
(9)
where: L is the load inductance f is the supply frequency IT(RMS) is the rms device current dv(com)/dt is the device commutating dv/dt rating. The presence of a snubber across the device can improve the turn-on performance of the triac by using the snubber capacitor discharge current in addition to the load current to ensure that the triac latches at turn-on. The value of the snubber resistor must be large enough to limit the peak capacitor discharge current through the triac to within the turn-on di/dt limit of the device.
502
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Varistor
The use of a metal oxide varistor (MOV), as shown in Fig. 13, protects the device from transient overvoltages which may occur due to mains disturbances.
Short-circuit condition
Fuses for protecting triacs should be fast acting, and the amount of fuse I2t to clear the circuit must be less than the I2t rating of the triac. Because the fuses open the circuit rapidly, they have a current limiting action in the event of a short-circuit. High voltage fuses exhibit low clearing I2t but the fuse arc voltage may be dangerous unless triacs with a sufficiently high voltage rating are used.
Overcurrent protection
Like all other semiconductor devices, triacs have an infinite life if they are used within their ratings. However, they rapidly overheat when passing excessive current because the thermal capacitance of their junction is small. Overcurrent protective devices (circuit breakers, fuses) must, therefore, be fast-acting.
Conclusions
This paper has outlined the most common uses and applications of thyristor and triac circuits. The type of circuit used depends upon the degree of control required and the nature of the load. Several types of gate circuit and device protection circuit have been presented. The amount of device protection required will depend upon the conditions imposed on the device by the application circuit. The protection circuits presented here will be suitable for the majority of applications giving a cheap, efficient overall design which uses the device to its full capability with complete protection and confidence.
Inrush condition
Motors, incandescent lamp or transformer loads give rise to an inrush condition. Lamp and motor inrush currents are avoided by starting the control at a large trigger angle. Transformer inrush currents are avoided by adjusting the initial trigger angle to a value roughly equal to the load phase angle. No damage occurs when the amount of inrush current is below the inrush current rating curve quoted in the device data sheet (see the chapter 'Understanding thyristor and triac data').
503
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
6.1.3 The Peak Current Handling Capability of Thyristors
The ability of a thyristor to withstand peak currents many times the size of its average rating is well known. However, there is little information about the factors affecting the peak current capability. This section will investigate the effect of pulse duration on the peak current capability of thyristors. Data sheets for thyristors always quote a figure for the maximum surge current that the device can survive. This figure assumes a half sine pulse with a width of either 10 ms or 8.3 ms, which are the conditions applicable for 50/60 Hz mains operation. This limit is not absolute; narrow pulses with much higher peaks can be handled without damage but little information is available to enable the designer to determine how high this current is. This section will discuss some of the factors affecting a thyristor's peak current capability and review the existing prediction methods. It will go on to present the results of an evaluation of the peak current handling capabilities for pulses as narrow as 10 µs for the BT151, BT152 and BT145 thyristors. It will also propose a method for estimating a thyristor's peak current capability for a half sine pulse with a duration between 10 µs and 10 ms from its quoted surge rating.
Expected Results
I2t is normally quoted at 10 ms, assuming that the surge is a half sine pulse, and is derived from the surge current from:
ITSM 2 0.01 I 2t = 2
This calculates the RMS current by dividing ITSM by 2 Under the simplest of analyses I2t would be assumed to be constant so a device's peak current capability could be calculated from:
1
I pk = ITSM
0.01 2 tp
where Ipk is the peak of a half sine current pulse with a duration of tp. However, experience and experiments have shown that such an approach is inaccurate. To overcome this, other 'rules' have been derived. One of these 'rules' suggests that it is not I2t which is constant but I3t or I4t. Another suggestion is that the 'constancy' continuously changes from I2t to I4t as the pulses become shorter. All these rules are expressed in the general equation:
1
Energy Handling
In addition to the maximum surge current, data sheets often quote a figure called "I2t for fusing". This number is used to select appopriate fuses for device protection. I2t represents the energy that can be passed by the device without damage. In fact it is not the passage of the energy which causes damage, but the heating of the crystal by the energy absorbed by the device which causes damage. If the period over which the energy is delivered is long, the absorbed energy has time to spread to all areas of the device capable of storing it - like the edges of the crystal, the plastic encapsulation, the mounting tab and for very long times the heatsink - therefore the temperature rise in the crystal is moderated. If, however, the delivery period is short - say a single half sine pulse of current with a duration of <10 ms - the areas to which the energy can spread for the actual duration of the pulse are limited. This means that the crystal keeps all the energy giving a much bigger temperature rise. For very short pulses (<0.1 ms) and large crystal, the problem is even worse because not all of the active area of a thyristor crystal is turned on simultaneously - conduction tends to spread out from the gate area - so the current pulse passes through only part of the crystal resulting in a higher level of dissipation and an even more restricted area for absorbing it.
I pk = ITSM
0.01 N tp
where is N is either constant or a function of the pulse width, for example:
1 N = log tp
The graph shown in Fig. 1 shows what several of these 'rules' predict would happen to the peak current capability if they were true. Unfortunately little or no real information currently exists to indicate the validity of these rules. Tests have been performed on three groups of devices - BT151, BT152 and BT145 - to gather the data which would, hopefully, decide which was correct.
Test Circuit
The technique chosen to measure the peak current capability of the devices was the stepped surge method. In this test, the device is subjected to a series of current pulses of increasing magnitude until it receives a surge which causes measurable degradation.
505
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
15 14
Choice of L & C
The width of the half sine pulse from an LC circuit is:
tpulse = C L
Peak Current Multiplying Factor
13 12 11 10 9 8 7 6 5 4 3 2 1 10us 100us 1ms 10ms
and the theoretical peak value of the current is:
I peak = V
L
C
Width of Half Sine Pulse I t = const. 3 I t = const.
2
I t = const. log(1/t) I t = const.
4
Fig. 1 Predicted ITSM multiplying factors
These equations assume that the circuit has no series resistance to damp the resonant action which would result in a longer but lower pulse. Minimising these effects was considered to be important so care was taken during the building of the circuits to keep the resistance to a minimum. To this end capacitors with low ESR were chosen, the inductors were wound using heavy gauge wire and the loop C / L / DUT / R3 was kept as short as possible. It was decided to test the devices at three different pulse widths - 10 µs, 100 µs and 1 ms - so three sets of L and C were needed. The values were selected with the help of a 'spreadsheet' program running on an PC compatible computer. The values which were finally chosen are shown in Table 1. Also given in Table 1 are the theoretical peak currents that the L / C combination would produce for a initial voltage on C of 600 V.
Circuit Description
The circuits used to perform the required measurements were of the form shown in Fig. 2. They produce half sine pulses of current from the resonant discharge of C via L. Triggering of the device under test (DUT) itself is used to initiate the discharge. The gate signal used for all the tests was a 100 mA / 1 µs pulse fed from a pulse generator in single-shot mode. The magnitude of the current pulse is adjusted by changing the voltage to which C is initially charged by varying the output of the PSU. The pulse is monitored by viewing the voltage across R3 on an digital storage oscilloscope. R1 and D protect the power supply. R1 limits the current from the supply when DUT fails and during the recharging of C. D attempts to prevent any high voltage spikes being fed back into the PSU.
Test Procedure
As mentioned earlier, the test method called for each device to be subjected to a series of current pulses of increasing amplitude. The resolution with which the current capability is assessed is defined by the size of each increase in current. It was decided that steps of approximately 5% would give reasonable resolution. Experimentation indicated that the clearest indication of device damage was obtained by looking for changes in the off-state breakdown voltage. So after each current pulse the DUT was removed from the test circuit and checked on a curve tracer. This procedure did slow the testing but it was felt that it would result in greater accuracy. Pulse Width C (µF) 13.6 100 660 L (µH) 0.75 10 154 Ipeak (A) 2564 1885 1244
D
R1
S1
L
Vak
DC PSU
DUT R2 C R3 Ia Trigger Pulse
10 µs 100 µs 1 ms
0-600V
Fig. 2 Surge current test circuit Table 1. Inductor and Capacitor Values Pushbutton S1 and resistor R2 are a safety feature. R2 keeps C discharged until S1 is pressed. The trigger pulse needs a button on the pulse generator to be pressed which means both hands are occupied and kept away from the test circuit high voltages. It was also decided that, since this work was attempting to determine the current that a device could survive - not which killed it, the figure actually quoted in the results for a device's current capability would be the value of the pulse prior to the one which caused damage.
506
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
Peak Current (amps)
@ 1000 @ #
* * #
@ *
* * * *
correlation between any of the predicted factors and the measured factors. In fact the variation in the factors between the various device types would indicated that no rule based on an Int function alone can give an accurate prediction. This implies that something else will have to be taken into account.
* *
@ @
# # #
# # #
100
10us 100us 1ms 10ms
Width of half-sine pulse
# BT151 @ BT152 * BT145
Fig. 3 Peak current capability measurements
Test Results
Figure 3 is a graph showing the measured current capabilities of all of the tested devices. Table 2 summarises the measurements by giving the mean of the results for the three device types at each of the pulse widths. Table 3 expresses the mean values as factors of the device ITSM rating. This table also gives the factors that the various 'rules' would have predicted for the various pulse widths. Mean Peak Current Capability (Amps) Pulse Width 10 µs 100 µs 1 ms BT151 912 595 264 BT152 1092 1021 490 BT145 1333 1328 697
Further study of Fig. 3 reveals that the difference in the peak current capability of the three device types is becoming less as the pulses become shorter. This could be explained by a reduction in the active area of the larger crystals, making them appear to be smaller than they actually are. This is consistent with the known fact that not all areas of a thyristor turn on simultaneously - the conduction region tends to spread out from the gate. If the pulse duration is less than the time it takes for all areas of the device to turn on, then the current flows through only part of the crystal, reducing the effective size of the device. If the rate at which the conduction area turns on is constant then the time taken for a small device to be completely ON is shorter than for a large device. This would explain why the performance increase of the BT145 starts falling off before that of the BT151.
Proposed Prediction Method
The above interpretation leads one to believe that the original energy handling rule, which says that I2t is a constant, may still be correct but that the performance it predicts will 'roll off' if the pulse duration is less than some critical value. The equation which was developed to have the necessary characteristics is:
1 1
I pk = ITSM
2 0.01 2 tp tp tp + tcrit
Table 2. Measured Current Capability Measured Factor Pulse Width 10 µs 100 µs 1 ms Predicted Factor (by Int rule) n= log(1/t) 4.0 3.2 2.2
which simplifies to:I pk = ITSM
0.01 tp + tcrit
BT BT BT n=2 n=3 n=4 151 152 145 9.1 6.0 2.6 5.5 5.1 2.4 4.4 31.6 10.0 5.6 4.4 10.0 4.6 2.3 3.2 2.2 3.2 1.8
where tcrit is proportional to - but not necessarily equal to the time taken to turn on all the active area of the crystal and is calculated from:tcrit = A R
Table 3. Measured and Predicted ITSM Multiplication Factors
where: A = crystal area R = constant expressing the rate at which the area is turned on. Preferably, A should be the area of the cathode but this information is not always available. As an alternative the total crystal area can be used if the value of R is adjusted accordingly. This will inevitably introduce an error because cathode and crystal areas are not directly proportional, but it should be relatively small.
Interpretation of Results
It had been hoped that the measurements would give clear indication of which of the 'rules' would give the most accurate prediction of performance. However, an inspection of Table 3 clearly shows that there is no
507
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
R was determined empirically to be approximately 0.02 m2/s Using this value of R gives the values of tcrit shown in Table 3. Using these values in the above equation predicts that the peak current handling capability of the BT151, BT152 and BT145 would be as shown in Fig. 4. Device BT151 BT152 BT145 tcrit 148 µs 410 µs 563 µs
In this section, an equation has been proposed which takes crystal size into account by using it to calculate a factor called tcrit. This time is then used to 'roll off' the performance increase predicted by the original energy handling equation - I2t = constant. This results in what is believed to be a more accurate means of estimating the capability of a device for a half sine pulse with a duration between 10 µs and 10 ms.
Table 3. Calculated Values of tcrit
Conclusions
The first conclusion that can be drawn from this work is that a thyristor, with average rating of only 7.5A, is capable of conducting, without damage, a peak current greater than 100 times this value in a short pulse. Furthermore the power required to trigger the device into conducting this current can be <1 µW. This capability has always been known and indeed the surge rating given in the data sheet gives a value for it at pulse widths of around 10 ms. What has been missing is a reliable method of predicting what the peak current capability of a device is for much shorter pulses. The results obtained using the test methods indicate that the previously suggested 'rules' fail to take into account the effect that crystal size has on the increase in performance.
Peak Current (amps)
1000
100 10us
100us
1ms
10ms
Width of half-sine pulse
BT151 BT152 BT145
Fig. 4 Predicted peak current handling using 'Rolled-off I2t' rule
508
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
6.1.4 Understanding Thyristor and Triac Data
The importance of reliable and comprehensive data for power semiconductor devices, together with the advantages of the absolute maximum rating system, is clear. This present article describes the data sheet descriptions of Philips thyristors and triacs, and aims to enable the circuit designer to use our published data to the full and to be confident that it truly describes the performance of the devices. A brief survey of short-form catalogues is an insufficient method of comparing different devices. Published ratings and characteristics require supporting information to truly describe the capabilities of devices; thus comparisons between devices whose performance appears to be similar should not be made on economic grounds alone. Manufacturers have been known to quote ratings in such a way as to give a false impression of the capabilities of their devices. Ratings and characteristics given in published data should always be quoted with the conditions to which they apply, and these conditions should be those likely to occur in operation. Furthermore, it is important to define the rating or characteristic being quoted. Only if data is both complete and unambiguous can a true comparison be made between the capabilities of different types. forward voltage reaches the breakover voltage V(BO), turn-on is initiated by avalanche breakdown and the voltage across the thyristor falls to the on state voltage VT. However, turn-on can occur when the forward (anode-to-cathode) voltage is less than V(BO) if the thyristor is triggered by injecting a pulse of current into the gate. If the device is to remain in the on state, this trigger pulse must remain until the current through the thyristor exceeds the latching current IL. Once the on state is established, the holding current IH is the minimum current that can flow through the thyristor and still maintain conduction. The load current must be reduced to below IH to turn the thyristor off; for instance, by reducing the voltage across the thyristor and load to zero.
Anode
Anode
p Gate Gate n p n
Thyristors
Thyristor is a generic term for a semiconductor device which has four semiconductor layers and operates as a switch, having stable on and off states. A thyristor can have two, three, or four terminals but common usage has confined the term thyristor to three terminal devices. Two-terminal devices are known as switching diodes, and four-terminal devices are known as silicon controlled switches. The common, or three-terminal, thyristor is also known as the reverse blocking triode thyristor or the silicon controlled rectifier (SCR). Fig. 1 shows the circuit symbol and a schematic diagram of the thyristor. All Philips thyristors are p-gate types; that is, the anode is connected to the metal tab. The thyristor will conduct a load current in one direction only, as will a rectifier diode. However, the thyristor will only conduct this load current when it has been 'triggered'; this is the essential property of the thyristor. Fig. 2 shows the static characteristic of the thyristor. When a small negative voltage is applied to the device, only a small reverse leakage current flows. As the reverse voltage is increased, the leakage current increases until avalanche breakdown occurs. If a positive voltage is applied, then again a small forward leakage current flows which increases as the forward voltage increases. When the
Cathode Cathode
Fig. 1 Thyristor circuit symbol and basic structure
Forward current
On-state characteristic
Reverse voltage Avalanche breakdown region
IL IH
Forward voltage V (BO)
Reverse characteristic
Off-state characteristic
Reverse current
Fig. 2 Thyristor static characteristic Thyristors are normally turned on by triggering with a gate signal but they can also be turned on by exceeding either the forward breakover voltage or the permitted rate of rise
509
Thyristors and Triacs
Power Semiconductor Applications Philips Semiconductors
of anode voltage dVD/dt. However, these alternative methods of switching to the conducting state should be avoided by suitable circuit design.
Triacs
The triac, or bidirectional triode thyristor, is a device that can be used to pass or block current in either direction. It is therefore an a.c. power control device. It is equivalent to two thyristors in anti-parallel with a common gate electrode. However, it only requires one heatsink compared to the two heatsinks required for the anti-parallel thyristor configuration. Thus the triac saves both cost and space in a.c. applications. Figure 3 shows the triac circuit symbol and a simplified cross-section of the device. The triac has two main terminals MT1 and MT2 (the load connections) and a single gate. The main terminals are connected to both p and n regions since current can be conducted in both directions. The gate is similarly connected, since a triac can be triggered by both negative and positive pulses.
MT1 MT1
The on state voltage/current characteristic of a triac resembles that of a thyristor. The triac static characteristic of Fig. 4 shows that the triac is a bidirectional switch. The condition when terminal 2 of the triac is positive with respect to terminal 1 is denoted in data by the term 'T2+'. If the triac is not triggered, the small leakage current increases as the voltage increases until the breakover voltage V(BO) is reached and the triac then turns on. As with the thyristor, however, the triac can be triggered below V(BO) by a gate pulse, provided that the current through the device exceeds the latching current IL before the trigger pulse is removed. The triac, like the thyristor, has holding current values below which conduction cannot be maintained. When terminal 2 is negative with respect to terminal 1 (T2-) the blocking and conducting characteristics are similar to those in the T2+ condition, but the polarities are reversed. The triac can be triggered in both directions by either negative (G-) or positive (G+) pulses on the gate, as shown in Table 1. The actual values of gate trigger current, holding current and latching current may be slightly different in the different operating quadrants of the triac due to the internal structure of the device. Quadrant Polarity of T2 wrt T1 Gate polarity 1 2 3 4 (1+) (1-) (3-) (3+) T2+ T2+ T2T2G+ GGG+
Gate
Gate
n p n n p
n
Table 1. Operating quadrants for triacs
Device data Anode to cathode voltage ratings
MT2
MT2
Fig. 3 Triac circuit symbol and basic structure
Forward current
The voltage of the a.c. mains is usually regarded as a smooth sinewave. In practice, however, there is a variety of transients, some occurring regularly and others only occasionally (Fig. 5). Although some transients may be removed by filters, thyristors must still handle anode to cathode voltages in excess of the nominal mains value. The following reverse off-state voltage ratings are given in our published data: VRSM: the n