An introduction to Power Electronic Devices

The work of solid-state relays and solid-state modules is inseparable from their internal power electronic devices, so it is necessary to introduce some basic knowledge of power electronic devices. Through this article, you will learn What are the power electronic devices? How do they work? What are their basic characteristics? How to use them?

You can quickly navigate to the chapters you are interested in through the Directory below, and the Quick Navigator on the right side of the browser.

CONTENTS

§1. What is a Power Electronic Device?

§2. What is an Uncontrollable Device?

§3. What is a Transistor?

§4. What is an Half-controlled Device?

§5. What is a Fully-controlled Device?

§6. How to choose Power Electronic Devices?

§7. How to use Power Electronic Devices?

§1.What is a Power Electronic Device?

In the power equipment or power system, the main circuit is used to realize the change or control of electric energy, and the Power Electronic Device (PED) is the core of the main circuit. Earlier, power electronic devices included power vacuum devices (such as mercury arc rectifiers, thyratrons) and power semiconductor devices (such as power diodes, thyristors). Due to the obvious advantages of power semiconductor devices in cost and performance, they have gradually replaced the position of power vacuum devices, thus so-called power electronic devices nowadays usually refer to power semiconductor devices with silicon as the main material.

Compared with information semiconductor devices (or Information Electronic Devices, IED) that also use silicon as the main material, power electronic devices have the following characteristics:

● Power electronic devices have a large power processing capacity, but due to the large power loss, they need to be equipped with a radiator for cooling.

● Power electronic devices often require information electronic devices to provide control signals.

● Power electronic devices generally work in the switching state, not in the amplification state, in order to reduce their power consumption.

Due to the characteristics of semiconductors, power electronic devices will inevitably produce certain power losses during operation. These power losses will not only reduce the conversion efficiency of electrical energy, but also cause permanent damage to the power electronic devices due to overheating. The main losses of power electronic devices can be divided into on-state loss, off-state loss, and switching loss. The on-state loss is the loss caused by the on-state voltage drop, and if the switching frequency of the power electronic device is not high, the on-state loss will account for a high proportion of the total loss. The off-state loss is the loss caused by the off-state leakage current, and the off-state loss usually accounts for a very small proportion of the total loss and is often ignored. The switching loss refers to the loss generated during the switching process of power electronic devices, and the switching loss is greatly affected by the switching frequency, that is to say, the higher the switching frequency, the greater the proportion of the switching loss in the total loss.

Because of the differences in the materials and structures used, the performance of different types of power electronic devices may be completely different. Therefore, before choosing power electronic devices, we must first understand their categories and characteristics to give full play to their advantages.

According to the degree of control, power electronic devices can be divided into uncontrollable type, half-controlled type, and fully-controlled type.

According to the effective trigger signal of driving circuit, power electronic devices can be divided into pulse-triggered type, and level-triggered type. Pulse triggering means that the working status of a power electronic device is triggered by a short pulse signal, such as SCR and GTO. Level triggering means that the working status of a power electronic device is triggered by holding an input signal at a specific level, such as GTR, MOSFET and IGBT. In short, level triggering has a longer duration and pulse triggering has a shorter duration.

According to the carrier involved in the conduction process, power electronic devices can be divided into unipolar type, and bipolar type. When unipolar type devices work, only one type of carrier (free electron or hole) participates in the conduction process, such as MOSFET, JFET and SIT. When bipolar type devices work, free electrons and holes are involved together in the conduction process, such as thyristor, GTO, GTR, IGBT, SITH, TRIAC, RCT and LTT.

According to the driving circuit signal, power electronic devices can be divided into current-driven type, and voltage-driven type. The working state of the current-driven devices is controlled by the input current, such as thyristor, GTO and GTR. The working state of the voltage-driven devices is controlled by the electric field effect generated by the input voltage, such as MOSFET, JFET and IGBT. Most of the voltage-driven devices are unipolar type, and most of the current-driven devices are bipolar type. Voltage-driven devices usually have the characteristics of high input impedance, low driving power, simple driving circuit, and high operating frequency. Current-driven devices usually have a conductance modulation effect, so the on-state voltage drop and the on-state loss is small, but the operating frequency is low, the required driving power is large, and the driving circuit is more complicated.

§2. What is an Uncontrollable Device?

2.1 Introduction to Uncontrollable Devices

Uncontrollable devices refer to devices that cannot be turned on and off by control signals, so no driving circuit is required. Uncontrollable devices generally refer to power diodes, and their basic structure and working principle are very similar to diodes. The power diode is encapsulated by a large-area PN junction and lead wires at both ends. According to the shape, power diodes can be divided into bolt type and flat type. According to the carriers involved in the conduction process, power diodes can be divided into unipolar power diodes and bipolar power diodes. Power diodes are widely used in power equipment in various fields because of their simple structure and low price. Especially fast recovery diodes and Schottky barrier diodes have an irreplaceable position in the rectification and inverter of low voltage, intermediate frequency and high frequency fields. With the development of modularization and integration technology, modular power diodes are becoming more and more common in the market (click to view more power diode modules).

2.2 How does Power Diode work?

The essence of the power diode is the PN junction formed by the contact between the P-type semiconductor and the N-type semiconductor. Therefore, in order to understand the working principle of the power diode, it is necessary to understand the basic characteristics and working principle of the PN junction.

2.2.1 Basic Structure of PN Junction

The power diode is composed of a P-type semiconductor and an N-type semiconductor. N-type semiconductors and P-type semiconductors are composed of doped intrinsic semiconductors, that is to say, the concentration of free electrons in N-type semiconductors is high, and the concentration of holes in P-type semiconductors is high. The connection area between the P-type semiconductor and the N-type semiconductor is called the PN junction. Free electrons and holes are also called free carriers (referred to as carriers). The movement and recombination of free electrons in the semiconductor will inevitably lead to the generation and recombination of holes. From a macro point of view, this process is more like negatively charged free electrons and positively charged holes moving in opposite directions in the semiconductor at the same time. The movement of free electrons and holes in semiconductors is very fast and random, so it is almost impossible to predict the trajectory of a certain free electron or hole and accurately know its position at a certain moment. However, the movement of a large number of holes and free electrons is not without rules.

Majority carriers in semiconductors will diffuse from high-concentration regions to low-concentration regions, that is to say, the majority carrier free electrons (n_n) in the N region diffuse from the high-concentration N region through the PN junction to the low-concentration P region, and at the same time, the majority carrier holes (p_p) in the P region diffuse from the high concentration P region through the PN junction to the low concentration N region. The carriers near the PN junction are depleted due to the diffusion movement, leaving only space charges (positive ions and negative ions) that cannot be moved, so this area is called the space charge region (also known as the depletion region). Since there are no free moving carriers in the space charge region, it is similar to an insulator. The space charge in the space charge region will generate a built-in electric field to prevent carriers from passing through the PN junction (the built-in electric field will be formed within a few nanoseconds after the PN junction is manufactured). Even so, there are still very few carriers that pass through the PN junction and become minority carriers in the opposite region, that is to say, free electrons become minority carriers in the P region (p_n), and holes become minority carriers in the N region (n_p). This phenomenon is called the quantum tunneling effect. The difference in carrier concentration on both sides of the space charge region produces a built-in potential difference (also called a built-in potential, or contact potential difference). Minority carriers will continue to diffuse into the lower concentration area. The average distance that the minority carriers can reach in the process of diffusion and recombination is called the diffusion length. The diffusion length is affected by the minority carrier lifetime, that is to say, the longer the minority carrier lifetime, the longer the diffusion length. When reaching the edge of the diffusion region, the minority carriers will pass through the PN junction and return to their original region under the action of the built-in electric field.

* Single Crystal Semiconductor and Polycrystalline Semiconductor

Single crystal semiconductor refers to a semiconductor with pure chemical composition, no impurities and no lattice defects, that is, intrinsic semiconductor, such as silicon (Si), germanium (Ge) and gallium arsenide (GaAs). The structure of single crystal semiconductors is very regular, and its macroscopic properties are anisotropic, that is, their physical properties are different in different directions. Single crystal semiconductors are the materials for most semiconductor devices.

Polycrystalline semiconductor refers to a semiconductor material composed of a large number of tiny single crystal semiconductor particles with different orientations. The structure of polycrystalline semiconductors does not have regularity, and their macroscopic properties are often isotropic, that is, their physical properties are the same in different directions. Polycrystalline semiconductors can be used to make narrow-film transistor switch matrices for solar cells, liquid crystal displays, and gate materials for MOSFET.

* Energy Band Theory

Energy band theory is a theory that uses quantum mechanics to study the movement of electrons inside a solid. In a coordinate system with energy as the ordinate, the energy of electrons in the crystal can be represented by a horizontal line (i.e., energy level), that is to say, the greater the energy, the higher the position of the line. Energy levels that are very close to each other within a certain energy range form the energy band. The vertical distance between the highest energy level and the lowest energy level in the energy band is called the energy band width. The position and width of the energy band are affected by the temperature and the crystal type (such as metal, semiconductor and insulator). The energy band of a semiconductor is shown in Figure 04(a).

Full Band: It refers to the energy band that is completely occupied by electrons when T = 0K. The electrons in the full band are valence electrons (that is, electrons that are bound by the valence bonds on the crystal atoms and cannot move freely), so it does not have any conductivity.

Empty Band: It refers to the energy band that is not occupied by electrons when T = 0K. There are no electrons in the empty band, so it does not have any conductivity. The empty band becomes the conduction band when there are electrons in it.

Conduction Band: It refers to the energy band that is not completely occupied by electrons when T > 0K. The electrons in the conduction band are free electrons, so it has conductivity.

Valence Band: It refers to the energy band occupied by valence electrons when T > 0K. The electrons in the valence band are valence electrons, so it does not have any conductivity.

Forbidden Band: It refers to the energy range between the top of the valence band and the bottom of the conduction band. There is no energy level of shared electrons (that is, electrons shared by multiple atoms in the crystal) in the forbidden band, but there are energy levels of non-shared electrons (that is, localized electrons in impurities and defects). The forbidden band width (band gap) reflects the bondage degree of valence electrons or the strength of the valence bond, that is, the minimum average energy required for intrinsic excitation. The middle line of forbidden band is Fermi level.

Common metallic materials generally have a narrow band gap (the conduction band and valence band even overlap each other), and their electrons easily gain energy at room temperature and transition to the conduction band, so they are very conductive and often used as conductors. Insulating materials generally have a wide band gap (usually greater than 9 electron volts, or 9 eV), and their electrons have a difficult time transitioning to the conduction band, so they are poorly conductive and often used as insulators. The bandgap of semiconductor materials is between conductor and insulator (about 1-3 eV, e.g. 0.67 eV for germanium and 1.12 eV for silicon at room temperature), so a semiconductor can be made to conduct electricity by giving the right amount of energy excitation (breaking the valence bond to make the valence electrons transition to the conduction band to produce free electrons and holes) or changing the width of the band gap (reducing the energy required for the electron transition).

* Intrinsic Excitation and Free Carriers

Intrinsic excitation means that by giving certain excitation conditions, the electrons in the intrinsic semiconductor will cross the forbidden band from the lower energy band (full band or valence band) into the higher energy band (empty band or conduction band) and becomes the free electrons. It should be noted that the free electrons in the conduction band of the intrinsic semiconductor refer to the approximately free electrons in the solid, which can move freely in the entire solid, but cannot run out of the solid. Due to the lack of electrons in the lower energy band, a positively charged vacancy is formed, which is called a hole. Free electrons in the conduction band and holes in the valence band are collectively called electron-hole pairs. In intrinsic semiconductors, free electrons and holes generated by intrinsic excitation can move freely, so they are called free carriers, and their concentrations are equal to each other, and as the temperature rises, their concentration will increase exponentially. The directional movement of free electrons and holes will form electrons flow and holes flow. The free electrons in the conduction band will fall into the holes, causing the electron-hole pairs to disappear. This process is called recombination. The energy generated during recombination is released in the form of electromagnetic radiation (emitting photon) or thermal vibration of the lattice (emitting phonon). At a certain temperature, the generation and recombination of electron-hole pairs exist simultaneously and reach a dynamic equilibrium. At this time, the intrinsic semiconductor has a certain carrier concentration and thus it has a certain electrical conductivity. Through intrinsic excitation, more electron-hole pairs are generated, thereby increasing the carrier concentration, which can effectively increase the electrical conductivity of the semiconductor. According to this principle, semiconductor devices such as semiconductor thermistors and semiconductor photo-resistors can be manufactured. The electrical conductivity of intrinsic semiconductors at room temperature is small, and the carrier concentration is sensitive to temperature changes, so it is difficult to effectively control the semiconductor characteristics of their semiconductors through temperature.

Intrinsic excitation methods can generally be divided into intrinsic thermal excitation, intrinsic photo excitation, and collisional ionization intrinsic excitation.

Intrinsic Thermal Excitation: It refers to the infrared photons radiated by thermal motion of molecules as the temperature increases, which makes valence electrons to gain enough energy to break free from the bondage of valence bonds and become free electrons. The energy required for intrinsic thermal excitation is equal to the band gap. The thermal excitation efficiency at room temperature is usually very limited, as a very high temperature is required to allow enough carriers to transition to the conduction band.

Intrinsic Photo Excitation: It refers to the photons radiated by light (generally referring to visible light or ultraviolet light), which makes valence electrons to gain enough energy to break free from the bondage of valence bonds and become free electrons. The energy required for intrinsic photo excitation is greater than intrinsic thermal excitation. Since the photons of visible light have higher energy than the infrared photons usually generated by thermal motion, the energy level of electrons after intrinsic photo excitation is usually located at a higher position in the conduction band. Because the momentum of the photon can be ignored, the intrinsic photo excitation does not change the momentum of the electron, so this process is also called the vertical transition. But if there is a phonon (referring to the simple harmonic vibration of the crystal lattice) involved, the momentum of the electron will change, so it is also called a non-vertical transition.

Collisional Ionization Intrinsic Excitation: It refers to the collision and ionization of valence electrons by high-energy electrons (that is, free electrons accelerated by an electric field) to become free electrons. The average energy required for collisional ionization intrinsic excitation is about 1.5 times the band gap. The electrons produced by the intrinsic excitation of collisional ionization are ionized electrons, which are the truly free electrons that can leave the solid, and their energy exceeds the free electrons with the highest energy level in the conduction band.

* Fermi Level and Fermi-Dirac Distribution

The Fermi level is the energy level that has a 50% chance of being occupied by electrons at any temperature, that is to say, below the Fermi level, the farther the distance is, the greater the possibility of being occupied by the electrons, and in the same vein, above the Fermi level, the farther the distance is, the less the possibility of being occupied by the electrons. For semiconductors, especially intrinsic semiconductors, the Fermi level is in the middle line of the forbidden band. When the temperature T = 0K, the full band is filled with electrons (the electron occupancy probability is 1), and the empty band has no electrons at all (the electron occupancy probability is 0), then their Fermi level is exactly at the middle line of the forbidden band (the electron occupancy probability is 1/2). Even when the temperature rises T > 0K, intrinsic excitation will produce electron-hole pairs, but since the number of electrons increased in the conduction band is equal to the number of electrons decreased in the valence band, the Fermi level is still in the middle line of the forbidden band (the electron occupancy probability is 1/2). Therefore, the position of the Fermi level of the intrinsic semiconductor does not change with temperature and is always at the middle line of the forbidden band. The Fermi-Dirac distribution of electrons can be calculated from the Fermi level and temperature, as shown in Figure 04(b).

* Semiconductor Doping

Generally, intrinsic semiconductor will be doped to introduce new energy levels to increase its electrical conductivity. The doped semiconductor is more susceptible to external influences, such as light and temperature rise.

By doping the silicon crystal (or germanium crystal) with phosphorus element (or antimony element), the phosphorus atom (or antimony atom) will occupy the position of the silicon atom. Then a set of full energy levels will be added to the position in the forbidden band which is very close to the conduction band. The electrons on these energy levels can easily transition to the conduction band to become free electrons. Therefore, the phosphorus element (or antimony element) is called a donor impurity (or N-type impurity), and a semiconductor doped with an N-type impurity is called an N-type semiconductor.

By doping the silicon crystal (or germanium crystal) with boron element (or indium element), the boron atom (or indium atom) will occupy the position of the silicon atom. Then a set of empty energy levels will be added to the position in the forbidden band which is very close to the valence band. The electrons in the valence band can easily transition to these energy levels and leave holes in the valence band. Therefore, boron element (or indium element) is called acceptor impurities (or P-type impurities), and semiconductors doped with P-type impurities are called P-type semiconductors.

2.2.2 Unidirectional Conductivity of PN Junction

The essence of the working principle of the power diode is the unidirectional conductivity of the PN junction. In the case of constant external conditions, like temperature and radiation, the external circuit will supplement the carriers consumed during the operation of the power diode, so the electrical conductivity of the power diode is mainly affected by its internal carrier concentration.

Forward Conduction State: If a forward bias voltage is applied across the power diode, the majority carriers will move closer to the PN junction, which will narrow the space charge region and weaken the built-in electric field, but the PN junction will still maintain dynamic equilibrium state. Only when the forward bias voltage is greater than the built-in electric field, the dynamic equilibrium can be broken, and a superimposed electric field in the opposite direction of the drift current is generated. The diffusion current I_D is greater than the drift current I_d. The current flowing through the PN junction is the forward current I_F. Due to the built-in potential difference, when the power diode is in the forward conduction state, an on-state voltage drop will be generated at both ends of it, which makes the power diode present a low impedance state. The on-state voltage drop of the power diode is not a fixed value, it is proportional to the current flowing.

Reverse Cut-off State: If a reverse bias voltage is applied across the power diode, the majority carriers will move away from the PN junction, which will widen the space charge region and enhance the built-in electric field. Then the dynamic equilibrium of the PN junction will be broken, and a superimposed electric field in the same direction as the drift current is generated. The drift current I_d is greater than the diffusion current I_D. The current flowing through the PN junction is the reverse saturation current I_sat. Because the number of minority carriers is too small, the reverse saturation current of the power diode is usually negligible, which makes the power diode present a high impedance state.

Reverse Breakdown State: If the reverse bias voltage across the power diode is continuously raised to a certain critical value, the number of carriers in the power diode increases rapidly, resulting in a significant increase in the reverse current I_R. In the reverse breakdown state, the power diode presents a no-impedance state, and its reverse current I_R and reverse voltage U_R are both very large. The reverse breakdown of PN junction is mainly divided into avalanche breakdown and Zener breakdown. Both of these breakdowns will increase the temperature of the PN junction, and eventually lead to thermal breakdown, which will cause permanent damage to the PN junction. If the cooling measures are done well enough, even if the power diode is reverse broken down, but the PN junction is not destroyed, then after limiting or closing the reverse voltage, the PN junction can still be restored to its original state.

* Diffusion Current and Drift Current

Diffusion motion refers to the movement of majority carriers from a high-concentration area to a low-concentration area. The diffusion motion is determined by the concentration gradient. Drift motion refers to the movement of minority carriers returning to the original area under the action of the built-in electric field. The drift motion is determined by the built-in electric field. Diffusion motion and drift motion are the basic motions of carriers in the PN junction. The current generated by the diffusion motion is called the diffusion current I_D, and the current generated by the drift motion is called the drift current I_d. When no external voltage is applied (or the forward bias voltage is less than the built-in electric field), the diffusion motion will cause the built-in electric field to enhance and the drift current will increase; the drift motion will cause the built-in electric field to weaken and the diffusion current to increase. Finally, the diffusion current is equal to the drift current, and the PN junction will be in a state of dynamic equilibrium with zero total current.

* Avalanche Breakdown and Zener Breakdown

Avalanche breakdown usually occurs in a low-doped PN junction with a wide depletion layer. When the reverse bias voltage is large, the free electrons in the semiconductor will continue to accelerate under the action of the electric field force, and due to the wide depletion layer, the free electrons will gain a long acceleration distance, thus gaining a large amount of kinetic energy. These high-energy electrons collide with valence electrons, freeing them from the bondage of valence bonds and thereby generating new electron-hole pairs. These newly generated free electrons continue to repeat this process under the action of the electric field force, causing the free carriers in the semiconductor to increase rapidly like an avalanche, leading to a sharp increase in drift current. The essence of the avalanche breakdown is the collisional ionization intrinsic excitation, so the avalanche breakdown voltage is generally higher than 6V. The avalanche breakdown voltage increases with the increase of temperature, mainly because the irregular thermal movement of carriers increases with the increase of temperature, so a larger reverse voltage is required to provide a large enough electric field to make the carriers do directional acceleration motion.

Zener breakdown usually occurs in a highly doped PN junction with a narrow depletion layer. Because the depletion layer is narrow, there is not enough distance for the free electrons to accelerate, so the free electrons cannot gain a large amount of kinetic energy, and thus the avalanche breakdown does not occur. However, even if the reverse bias voltage is not large, a strong electric field can still be generated in the PN junction due to the narrow depletion layer, which pulls the electrons out of the valence bond to create new electron-hole pairs. This phenomenon is also called field-induced excitation. Field-induced excitation will greatly increase the number of carriers in the semiconductor, thereby significantly increasing the drift current. Zener breakdown voltage is generally less than 4V. The Zener breakdown voltage decreases with increasing temperature, mainly because the electrons in the valence bond become more active with increasing temperature, so it is more easily for the electric field force to pull them out.

2.2.3 Capacitance Effect of PN Junction

The amount of charge in the PN junction changes with the applied voltage, exhibiting a capacitance effect. This capacitance is called the junction capacitance C_J (also known as differential capacitance). According to the different generation mechanism and function, the junction capacitance C_J can be divided into the barrier capacitance C_B and the diffusion capacitance C_D, and they conform to the calculation formula C_J = C_B + C_D. Both the barrier capacitance and the diffusion capacitance are non-linear capacitance.

1- Barrier Capacitance

The narrow layer of ions in the space charge region (depletion layer) forms the barrier region. The number of space charges in the barrier region changes with the applied bias voltage, which is equivalent to the charge and discharge effect of the capacitor, that is to say, when the forward bias voltage increases, the barrier region decreases, which is equivalent to storing free electrons or holes into the barrier region, and in the same vein, when the forward bias decreases, the barrier region increases, which is equivalent to taking out free electrons or holes from the barrier region. The equivalent capacitance of the barrier region is called the barrier capacitance C_B. If the frequency of the applied bias voltage is higher, the effect of the barrier capacitance is more obvious. Regardless of low-frequency operation or high-frequency operation, the barrier capacitance may deteriorate the unidirectional conductivity of the semiconductor device, or even fail to work. In fact, the maximum operating frequency of a semiconductor device is often determined by the barrier capacitance. It is worth noting that the barrier capacitance is the capacitance effect related to the majority carrier (p_p and n_n), and neither the forward bias nor the reverse bias can be ignored. In forward bias, when the forward voltage is low, the barrier capacitance is much larger than the diffusion capacitance, so the barrier capacitance is the main component of the junction capacitance, C_J ≈ C_B.

2- Diffusion Capacitance

When the PN junction is forward biased, the built-in electric field is weakened, and the drift motion of minority carriers is weakened. The diffusion current is greater than the drift current. Therefore, the carriers that diffuse to the opposite area will accumulate at the barrier boundary to form a certain concentration of non-equilibrium minority carriers (p_n and n_p), that is to say, the closer to the PN junction, the higher the concentration, and in the same vein, the farther away from the PN junction, the lower the concentration. The amount of charge of such non-equilibrium minority carriers changes with the forward bias, which is equivalent to the charge and discharge effect of the capacitor, that is to say, when the forward bias increases, the non-equilibrium minority carriers are increased, which is equivalent to the charging of the capacitor, and in the same vein, when the forward bias decreases, the non-equilibrium minority carriers are reduced, which is equivalent to the discharging of the capacitor. The equivalent capacitance at the boundary of the barrier region is called the diffusion capacitance C_D. The diffusion capacitance has a great influence on the switching speed of the PN junction when working at low frequencies, and can be ignored when working at high frequencies. In forward bias, if the forward voltage is high, the diffusion capacitance is much larger than the barrier capacitance, so the diffusion capacitance is the main component of the junction capacitance, C_J ≈ C_D. In reverse bias, there are too few non-equilibrium minority carriers, and the diffusion capacitance can be ignored, so the barrier capacitance is the main component of the junction capacitance, C_J ≈ C_B.

2.3 Main Parameters of Power Diodes

1- Maximum Forward Average Current I_FM(AV)

The maximum forward average current I_F(AV) is the rated current of the power diode, which refers to the average value of the maximum power frequency half-wave sine current allowed to flow through the power diode under the specified case temperature T_C and heat dissipation conditions. If it exceeds I_F(AV), the diode will be burned out. Since the waveform of some power diodes is not necessarily a half-sine wave, and some power diodes do not have resistance characteristics, I_F(AV) is defined according to the thermal effect of current, that is, find a resistor with similar heat generation according to the principle of equal effective value. Considering that the heat dissipation conditions will affect the ability of the power diode to withstand current, it is recommended to leave a certain margin to avoid damage to the power diode due to heat dissipation problems.

2- Threshold Voltage U_TO

The threshold voltage U_TO (also known as the dead region voltage) is the lowest forward voltage at which the power diode can be turned on. The threshold voltage is the lowest forward voltage drop of the power diode. The threshold voltage of germanium crystal is about 0.1V; the threshold voltage of silicon crystal is about 0.5V.

3- On-state Voltage Drop U_CO

The on-state voltage drop U_CO (also known as the conduction voltage) is the forward voltage drop when the power diode is turned on and works stably. Ideally, the on-state voltage drop of the power diode is equal to the built-in potential. The built-in potential is related to the degree of semiconductor doping and is approximately equal to the half of the band gap. The on-state voltage drop of the power diode is proportional to the current flowing. The on-state voltage drop of germanium crystals is usually around 0.1-0.3V; the on-state voltage drop of silicon crystals is usually around 0.5-0.8V.

4- Maximum Forward Voltage Drop U_FM

The maximum forward voltage drop U_FM is the forward voltage drop corresponding to the maximum forward average current I_FM(AV) at a specified temperature.

5- Reverse Saturation Current I_sat

When an appropriate reverse voltage is applied, a very small leakage current will be generated, which is called the reverse saturation current I_sat. The reverse saturation current is generated by the drift motion of minority carriers, so it is greatly affected by temperature.

6- Reverse Repetitive Peak voltage U_RRM

The reverse repetitive peak voltage U_RRM (also known as the maximum reverse voltage U_RM) is the rated voltage of the power diode, which is the highest reverse voltage that the power diode can withstand and can be applied repeatedly. If it exceeds this value, the power diode will be reverse broken down and damaged. Taking into account factors such as overvoltage in the circuit, it is generally recommended to choose a power diode with twice the rated value as the margin. For example, a power diode with a rated voltage of 1000V can only be used in a 500V operating environment.

7- Reverse Recovery Time t_rr

The reverse recovery process is caused by the capacitance effect of the power diode. When the circuit switched from the on-state to the off-state, the power diode needs to release the charge stored in the junction capacitance before blocking the reverse current. This discharge time is called the reverse recovery time t_rr, that is, the time from when the forward conduction current is zero to when it enters the fully turn-off state. The reverse recovery time of power diodes of different specifications is different, so we need to fully consider when designing the circuit, otherwise it may cause unnecessary trouble. For example, the reverse recovery time of a power diode is T_rr. If a continuous PWM wave with a period of T₁ (T₁ < T_rr) passes through the power diode, the PWM wave cannot be blocked when the power diode is reversely biasing.

8- Maximum Operating Junction Temperature T_JM

The junction temperature T_J refers to the average temperature of the PN junction. The maximum operating junction temperature T_JM refers to the highest average temperature that the PN junction can withstand without damage (usually the highest junction temperature of germanium transistors is about 75°C, and the highest junction temperature of silicon transistors is about 150°C). Temperature has a very significant impact on the working characteristics of power diodes, so sufficient heat dissipation conditions must be provided to avoid damage to the power diode due to overheating.

9- Maximum Operating Frequency f_M

The maximum operating frequency f_M is the upper turn-off frequency of the power diode. If the frequency is too high, the power diode will easily lose its ability to block reverse current due to the capacitance effect. At the same time, if the frequency is too high, it will also cause the power diode to be burned due to the increase of on-state power consumption.

10- Surge Current I_FSM

The surge current I_FSM refers to the maximum continuous overcurrent of one or several power frequency cycles that the power diode can withstand.

2.4 Basic Characteristics of Power Diodes

2.4.1 Static Characteristics of Power Diodes

The static characteristics of the power diode mainly refers to the volt-ampere characteristic curve of the power diode, shown in Figure 10.

When a forward bias voltage is applied to both ends of the power diode, the power diode will not be turned on immediately. The power diode will be turned on only when the forward voltage is greater than its threshold voltage U_TO. At this time, the forward current I_F begins to increase significantly, until the power diode is in a stable conduction state, at which time its conduction voltage is U_CO. If the forward current reaches I_FM, the corresponding voltage drop is U_FM, and the power diode will be burned out due to excessive current.

When a reverse bias voltage is applied to both ends of the power diode, the power diode will not conduct, but will generate a small constant value current, that is, reverse leakage current. The power diode will be reverse broken down when the reverse voltage reaches its reverse breakdown voltage U_BR, and its reverse current will become very large at this instant.

2.4.2 Dynamic Characteristics of Power Diodes

The dynamic characteristics of the power diode refer to its switching characteristics, that is, the voltage-current characteristic of the power diode during the switching between on-state and off-state. Because of the junction capacitance, the voltage-current characteristics of power diodes change with time.

1- Turn-on process

The dynamic characteristic of the power diode in the turn-on process is shown in Figure 11. When the voltage changes from zero bias to forward bias, the forward current I_F of the power diode will rise from 0 to I_F1. Due to the large di/dt, under the action of the line inductance, a forward peak voltage U_FP will be generated at both ends of the power diode. After a certain period of time, the forward voltage U_F will gradually drop from U_FP to the stable value U_F1 (that is, the on-state voltage drop). In this process, the time when the forward current rises from 0 to I_F1 is called the forward recovery time t_fr.

2- Turn-off process

The dynamic characteristic of the power diode in the turn-off process is shown in Figure 12. Because of the junction capacitance, the power diode will not turn off immediately even if the forward bias is switched to the reverse bias, and it will take some time to regain its reverse blocking capability.

When the power diode is switched from forward bias to reverse bias at t_F, and the forward current I_F decreases rapidly, and drops to 0 at t₀, and di_F/dt will be large at this instant. During the period from t₀ to t₁, the current not only does not disappear, but instead becomes the reverse current I_R and rises rapidly until the maximum I_RP is reached. This time period is called the delay time t_d. During the period from t₁ to t₂, the reverse current begins to drop sharply to a very small value. This time period is called the fall time t_f. During the period afte t₂, the reverse current begins to slowly decrease until it drops to 0 (in fact, there still exists a very small reverse leakage current). The time from t₀ to t₂ is called the reverse recovery time t_rr, during which the power diode is reverse conducting. The reverse recovery time t_rr determines the operating frequency of the power diode. If the operating frequency of the external circuit is too high, the power diode cannot enter the reverse cut-off state when reverse biased, and there is a large reverse current, which is equivalent to the power diode losing its reverse blocking ability.

Before the reverse current rises to the maximum value, the voltage across the power diode drops rapidly from the on-state voltage drop U_F1 to 0. Moreover, since t_f is usually very short, di_R/dt will be very large. Under the action of the line inductance, a reverse peak voltage U_RP is quickly generated at both ends of the power diode, and then it begins to drop to a stable value U_R1. The reverse peak voltage is usually very large and may break down the power diode. Therefore, increasing the proportion of t_f in t_rr will help reduce the reverse peak voltage. The recovery coefficient (Sr = t_f / t_d) is usually used to express the softness of the reverse recovery characteristics of the power diode.

2.5 Main Types of Power Diodes

1- General Purpose Diode

General purpose diodes (GPD), also known as rectifier diodes, have a long recovery time, high forward current rating and high reverse voltage rating. They are mostly used in rectifier circuits with low switching frequency (below 1kHz), and generally cannot be used in medium and high frequency circuits.

2- Fast Recovery Diode

The internal structure of the fast recovery diode (FRD) is different from that of the general purpose diode. It adds a base region I between the P-type and N-type silicon materials to form a P-I-N structure. Because the base region is very thin and the reverse recovery charge is few, it not only greatly reduces t_rr and the transient forward voltage drop, but also improves its reverse voltage withstand capability. The recovery time of the fast recovery diode is usually several hundred nanoseconds (t_rr > 100ns), its forward voltage drop is about 0.6V, the forward current is several amperes to several thousand amperes, and the reverse peak voltage can reach several hundred to several thousand volts. Ultra-fast recovery diodes (UFRD), also known as fast recovery epitaxial diodes (FRED), have less reverse recovery charge than that of fast recovery diodes, and their recovery times are as short as 20~30 nanoseconds (t_rr < 100ns). The forward voltage drop of ultra-fast recovery diodes is also very low (about 0.9V), but its reverse withstand voltage capability is usually not high (less than 1200V).

3- Schottky Barrier Diode

Schottky Barrier Diode (SBD) is a kind of power diode based on the barrier formed by the contact between metal and semiconductor. Compared with general purpose diodes and fast recovery diodes, Schottky barrier diodes have the advantages of short reverse recovery time, no obvious forward voltage overshoot, and high reverse withstand voltage, but their reverse leakage current is large. The forward voltage drop of a Schottky barrier diodes is affected by the reverse withstand voltage, and if the reverse withstand voltage increases, the forward voltage drop will increase significantly. But when the reverse withstand voltage is low, the forward voltage drop of Schottky barrier diodes is significantly lower than that of general purpose diodes and fast recovery diodes, so the switching losses and on-state losses are very low. Therefore, Schottky barrier diodes are usually used in rectifier circuits below 200V. However, it should be noted that Schottky barrier diodes are very sensitive to temperature, so their operating temperature must be strictly limited.

§3. What is a Transistor?

Before introducing half-controlled devices and fully-controlled devices, it is necessary to briefly introduce Bipolar Junction Transistors (BJT).

3.1 Introduction to Transistors

The transistor is a semiconductor bipolar device device with three terminals and two PN junctions, so it is also known as the bipolar junction transistor or BJT. The transistor is one of the basic components of semiconductor devices and also one of the core components. Since its birth in the 1940s, the transistor has completely changed the structure of electronic circuits, triggered a solid-state revolution, and promoted the emergence of integrated circuits and large-scale integrated circuits. The transistor has a current amplifying function, and can control a large change in collector current with a very small change in base current, so it is often used as a contact-less switch in electronic circuits. The switching frequency of the transistor is high, and there is no mechanical service life, so it has a significant advantage over electromagnetic relays and mechanical switches.

3.2 How does Transistor work?

3.2.1 Basic Structure of Transistors

The transistor is a three-layer semiconductor structure, which has one more PN junction than Power Diode. These two closely spaced PN junctions divide the transistor into three regions with different areas and doping concentrations, the one is the base region that has very small area with 3-30μm thickness and low doping concentration, the second is the emitter region that has small area and high doping concentration, and the third is the collector region that has large area and low doping concentration. The PN junction between the collector region and the base region is called the collector junction J₁. The PN junction between the emitter region and the base region is called the emitter junction J₂.

According to the material, transistors can be divided into silicon transistors and germanium transistors. According to the doping composition, transistors can be divided into PNP transistors and NPN transistors, that is to say, under forward bias, the emitter region of PNP transistors will emit holes, and its direction is the same as that of the current, so the arrow in the electrical symbol goes from the emitter to the base, and in the same vein, under forward bias, the emitter of the NPN transistors will emit free electrons, and its direction is opposite to that of the current, so the arrow in the electrical symbol goes from the base to the emitter.

3.2.2 Working Principle of Transistors

Take the NPN transistor as an example. The NPN transistor can be regarded as two equivalent diodes (VD1 and VD2), as shown in Figure 13(a). Because the N- region of VD1 has low a doping concentration and a large area, it is not prone to avalanche breakdown, so it can withstand a large reverse voltage. However, in forward bias, VD1 has a very small forward current, so overall, VD1 is well suited for operation in the reverse cut-off state. When VD1 is in the reverse state, VD1 produces a reverse saturation current I_CBO, but the doping concentration in the N- and P regions is very low, so the I_CBO is very small. Due to the high doping concentration and small area of the N+ region in VD2, it is prone to avalanche breakdown, so its reverse withstand voltage capability is very poor. However, in forward bias, VD2 has a very large forward current, so overall, VD2 is well suited for operation in the forward conduction state. When VD2 operates in a forward bias, two currents are generated, one is the current I_EP generated by the flow of holes in the P region and the other is the current I_EN generated by the flow of free electrons in the N+ region. Since the doping concentration of the P region of VD1 is lower than that of the N+ region, I_EN is greater than I_EP. On the basis of the above information, if we further understand how these two equivalent diodes work, then it is easy to understand how NPN transistors work.

The connection method of transistors in a actual circuit can be divided into common emitter, common collector and common base. For the sake of explanation, we take an NPN transistor with a common emitter connection method as an example, as shown in Figure 13(b), that is to say, the collector power supply E_C and the collector resistor R_C are connected in series to the collector and emitter, and moreover, the base power supply E_B and the base resistor R_B are connected in series to the base and emitter. In this circuit, current flows in from the collector and base, and flows out of the emitter. Similarly, the total current flowing in from the collector is called the collector current I_C, the total current flowing in from the base is called the base current I_B, and the total current flowing out of the emitter is called the emitter current I_E. The linear relationship between I_C and I_E is common base current gain α, and the linear relationship between I_C and I_B is common emitter current gain β. It is worth noting that because of the difference doping concentration, the collector junction J₁ is not suitable for forward bias, and the emitter junction J₂ is not suitable for reverse bias. If the collector and emitter are reversely connected, the possibility of NPN transistor breakdown will increase significantly.

Cut-off State: The NPN structure is such that there is always a PN junction in a reverse bias state. When no voltage is applied to the base, even if a large voltage smaller than the breakdown voltage BV_CEO) is applied to the collector and emitter, the NPN transistor cannot be turned on, but it still outputs a very small leakage current I_CEO.

Active State: VD1 and VD2 must work at the same time to turn on the NPN transistor, so a certain voltage needs to be applied to the base to make J₁ reverse biased (U_BC < 0) and J₂ forward biased (U_BE > U_TO). When the NPN transistor is turned on, its internal current is a bit different from that as the equivalent diodes worked separately, as shown in Figure 13(b). The free electrons injected from the N+ region into the P region do not completely recombine with the holes in the P region. As a result of the reverse bias of J₁, a part of the free electrons will pass through the P region and be directly injected into the N- region to generate a large reverse current I_CN. Then we can conclude that by injecting a small amount of free electrons into the P region through the base, the ICN can be effectively increased, that is, when the NPN transistor is working in an active state, a small change in the base current I_B will lead to a large change in the collector current I_C. This phenomenon is called conductance modulation effect. The conductance modulation effect is manifested as an amplifier that is able to amplify a small input current into a large output current, so the active state is also vividly known as the amplification state.

Saturation State: With the increase of I_B, the concentration of holes in the P region decreases, I_EP decreases, and the depletion region of J₁ keeps increasing, as shown in Figure 13(c). At the meantime, as fewer and fewer free electrons are injected into the P region from the N+ region, I_BN begins to approach a certain critical value infinitely, and the amplification effect of I_B on I_C begins to weaken dramatically. When I_B and I_C no longer have a linear relationship, the NPN transistor begins to be in the saturation state. In this case, even if the I_B increases, the increase in I_C is very small, and the saturation depth of the NPN transistor begins to deepen. When almost all free electrons in the N+ region are injected into the N- region, the base potential is the same as the collector potential (U_BC = 0). At this time, the NPN transistor is in a deep saturation state, and I_C is completely unaffected by I_B. It is important to note that the deeper the saturation depth of the NPN transistor, the larger the depletion region of J1 and the greater the likelihood of avalanche breakdown on J₁.

* Calculation Formula of Transistor

I_C = I_CN + I_CBO, (1)

I_B = I_BN + I_EP - I_CBO, (2)

I_E = I_C + I_B = I_CN + I_BN + I_EP, (3)

because I_C > 0, then we get I_E / I_C = I_B / I_C + 1; (4)

α = I_CN / I_E = (I_C - I_CBO) / I_E, (5)

β = I_CN / (I_B + I_CBO) = (I_C - I_CBO) / (I_B + I_CBO), (6)

because I_C > I_B >> I_CEO >> I_CBO ≈ 0, if ignoring all the leakage current, we can get α ≈ I_C / I_E, β ≈ I_C / I_B, (7)

then we can get 1/α = 1/β + 1, (8)

so the relation between α and β is: α = β / (1 + β), β = α / (1 - α). (9)

* Leakage Current

Both the collector junction reverse saturation current I_CBO and the penetration current I_CEO are unavoidable leakage currents in the transistor. The value of I_CBO can be measured by opening the emitter of the transistor (I_E = 0) and applying voltage to its collector and base, as shown in Figure 14(a). The value of I_CEO can be measured by opening the base of the transistor (I_B = 0) and applying voltage to its collector and emitter, as shown in Figure 14(b).

The generation mechanism of the collector junction reverse saturation current I_CBO is shown in Figure 13(a).

The generation mechanism of the penetration current I_CEO is as follows: Under the action of an external electric field, the majority carriers in the collector region move away from the PN junction, widening the space charge region, thereby enhancing the built-in electric field of the collector junction J₁, which is conducive to drift motion; the majority carriers in the emitter region move closer to the PN junction, narrowing the space charge region, thereby weakening the built-in electric field of the emitter junction J₂, which is not conducive to drift motion. Therefore, under the action of the built-in electric field, the minority carriers in the base region drift to the collector region through J₁. Simultaneously, the minority carriers in the collector region drift to the base region through J₁, some of which participate in the recombination of the base region, and the others of which diffuse to the emitter region through J₂. However, because of the low doping concentration of the base region, the proportion of minority carriers participating in the recombination of the base region is very low. It is not difficult to find that this process is very similar to the generation mechanism of I_EN when the transistor is turned on. Therefore, there is a linear relationship between I_CEO and I_CBO, I_CEO = (1 + β) * I_CBO. Nevertheless, due to the low doping concentration of the collector region and the base region, the value of I_CEO is very low and can usually be ignored. The I_CEO of silicon transistors is generally less than 100nA; the I_CEO of germanium transistors is generally less than 100μA.

* Conductance Modulation Effect

Conductance (G) is the reciprocal of resistance, and the unit is Siemens (S). Conductance modulation effect (also known as the base region conductivity modulation effect, or Webster effect) is one of the basic characteristics of bipolar transistors (BPT), which refers to the phenomenon that the conductivity of the base region increases significantly (or the resistivity of the base region decreases significantly) when the working current of the bipolar transistor is large. Besides BJT, other bipolar transistors such as SCR, GTO, GTR and parasitic transistors in IGBT also have conductivity modulation effect. In addition, when the working current of the bipolar transistor is large, the Early effect (the phenomenon that the changes of the collector junction voltage will lead to the changes of the width of the base region) and the Kirk effect (the phenomenon that the width of the base region increases) will also appear.

3.3 Main Parameters of Transistors

1- Common Base Current Gain α

Common base current gain α (the full name is "hybrid parameter forward current gain, common base", HFB), which is determined by the emitter efficiency factor and the base region transport factor, α = F_E * F_B. When the base is zero-biased (U_BC = 0), the base short-circuit amplification factor α₀ is determined by the emitter efficiency factor, the base region transport factor, the collector efficiency factor and the avalanche multiplication factor, α₀ = F_E * F_B * F_C * M.

The emitter efficiency factor F_E is the ratio of the electron current I_EN injected into the base region to the emitter current I_E, F_E = I_EN / I_E = I_EN / (I_EN + I_EP) = 1 / [1 + (I_EP / I_EN) ]. By reducing the doping concentration of the base region, the total amount of impurities in the base region is much smaller than that in the emitter region, which can effectively increase the number of minority carriers injected into the base region from the emitter region. The closer the ratio of I_EP to I_EN is to 0, the higher the emission efficiency of the transistor.

The base region transport factor F_B is the ratio of the electron current I_CN that reaches the collector region to the electron current I_EN injected into the base region, F_B = I_CN / I_EN. By reducing the width of the base region, the time that carriers from the emitter region stay in the base region can be effectively shortened, thereby increasing the number of minority carriers that transit the base region. The smaller the width of the base region, the less the recombination loss of electrons from the emission region in the base region.

The collector efficiency factor F_C is the ratio of the collector current I_C to the electron current I_CN that reaches the collector region, F_C = I_C / I_CN.

The avalanche multiplication factor M is used to describe the avalanche multiplication effect when the reverse voltage of the collector junction increases to close to the avalanche breakdown voltage. It is usually estimated with the following formula, M = 1 / [1 - (V / V_B)^n], n is determined by the material of the PN junction (silicon: n=1.5-4; germanium: n=2.5-8); V_B is the reverse breakdown voltage of the collector J₁; V is the voltage across the collector junction. When the absolute value of V tends to the absolute value of V_B, M tends to infinity, and avalanche breakdown will occur in the PN junction.

hFB(α) is usually used to express the common base DC current gain, hFB(α) = I_C / I_E, and its range is 0.95-0.99; hfb(α) is usually used to express the common base AC current gain, hfb(α) = ΔI_C / ΔI_E. Generally speaking, hfb(α) ≈ hFB(α).

2- Common Emitter Current Gain β

Common emitter current gain β (the full name is "hybrid parameter forward current gain, common emitter", HFE) is the ratio of collector current to base current, and its value is usually much larger than 1. The current amplification factor (or forward current gain) of the transistor usually refers to the common emitter current gain β. hFE(β) is usually used to express the common emitter DC current gain, hFE(β) = I_C / I_B, which can be measured directly by a multimeter. hfe(β) is usually used to express the common emitter AC current gain, hfe(β) = ΔI_C / ΔI_B.

3- Common Collector Current Gain γ

Common collector current gain γ (the full name is "hybrid parameter forward current gain, common collector", HFC) is the ratio of emitter current to base current. Generally, hFE(γ) is usually used to express the common collector DC current gain, hFC(γ) = I_E / I_B. hfc(γ) is usually used to express the common collector AC current gain, hfc(γ) = ΔI_E / ΔI_B. This parameter is rarely used in normal times.

4- Threshold Voltage U_TO

The threshold voltage U_TO is the voltage that triggers the conduction of the emitter junction of the transistor.

5- Characteristic Frequency fT

The characteristic frequency fT is also called the gain bandwidth product, which can be defined as the operating frequency of the transistor when β=1. If the operating frequency f₀ and the high-frequency current amplification factor β are known, the characteristic frequency fT can be obtained, fT = β * f₀. As the operating frequency increases, the current amplification factor will decrease. If the operating frequency of the transistor is equal to the characteristic frequency (f₀ = fT), the transistor will completely lose its current amplification ability; if the operating frequency of the transistor is greater than the characteristic frequency (f₀ > fT), the transistor will not work normally.

6- Maximum Operating Voltage U_CEM

U_CE is the voltage applied between the collector and emitter of the transistor, and its maximum allowable value is the maximum operating voltage U_CEM, that is, the rated voltage of the transistor. When U_CE > U_CEM, the transistor will be broken down.

7- Maximum Collector Allowable Current I_CM

I_C is the collector current, and its maximum allowable value is the maximum collector allowable current I_CM, that is, the rated current of the transistor. Empirically speaking, when the current gain β drops by half from its maximum, the corresponding collector current I_C is the I_CM of the transistor. In order to ensure the safety of use, twice the I_CM is generally used as the margin.

8- Maximum Collector Dissipation Power P_CM

The collector dissipation power P_C is the product of I_C and U_CE, that is, P_C = I_C * U_CE, and its maximum allowable value is the maximum collector power dissipated P_CM, which is the power that allows the transistor to reach its maximum junction temperature at 25°C room temperature. When P_C > P_CM, the PN junction structure inside the transistor will be permanently destroyed.

3.4 Basic Characteristics of Transistors

The parameters that have a slowly changing or fixed value during the stable operation of a transistor are static parameters of the transistor, and the relationship between them is known as the static characteristics of the transistor. The parameters that have a sharply changing value during the switching operation of a transistor are dynamic parameters of the transistor, and the relationship between them is known as the dynamic characteristics of the transistor. If there is only a DC signal in the input signal of the transistor, it is known as DC operation or static operation. If there is an AC signal in the input signal of the transistor, it is known as AC operation or dynamic operation.

We take the NPN transistor with common emitter connection method as an example, its input terminal is the base and its output terminal is the collector, so its input current is I_B, its input voltage is U_BE, its output resistance is R_C, its output current is I_C, and its output voltage is U_CE.

3.4.1 Static Characteristics of Transistors

The static characteristics of the transistor are divided into input characteristics (relationship between input current and input voltage), output characteristics (relationship between output current and output voltage), temperature characteristics (the influence of temperature on input characteristics and output characteristics) and safe operating area (stable operating conditions of the transistor).

1- Input Characteristics of Transistors

The input characteristic of the transistor is similar to the forward input characteristic of the power diode, as shown in Figure 15.

When U_CE is a fixed value and U_BE > U_TO, the base current I_B increases with the increase of U_BE.

When U_CE increases, U_TO increases and the input characteristics curve shifts to the right. This is because with the increase of U_CE, part of the carriers that should be injected into the base region from the emitter region transit the base region and are directly injected to the collector region, so that the carrier concentration in the base region is too low, resulting in the emitter junction cannot be turned on as its diffusion current is less than or equal to the drift current. Therefore, U_BE needs to be increased to allow more carriers to be injected from the emitter region into the base region to turn on the emitter junction, however, it will increase the threshold voltage, and move the input characteristic curves to the right. When the U_CE increases to a certain extent, most of the carriers that can be injected into the base region from the emitter region are collected by the collector region, so it is difficult to change the input characteristics of the transistor even if the U_CE continues to increase.

2- Output Characteristics of Transistors

Before introducing the output characteristics of transistor, it is necessary to introduce the concept of DC load line. The DC load line is the volt-ampere characteristic curve of the collector load R_C when the transistor is operating in a static state, I_C = (E_C - U_CE) / R_C. When the transistor is in the off state, it is equivalent to the collector circuit is in the off state. Therefore, U_CE = 0, and the voltage on R_C is equal to the power supply voltage E_C. When the transistor is in the on state, I_C = I_CM, that is, the current flowing through R_C rises to its maximum. Mark these two points in a rectangular coordinate system with I_C as the Y-axis and U_CE as the X-axis, and connect them into a line segment, that is, the DC load line, as shown in Figure 16. The intersection of the DC load line with the Y-axis is called the saturation point, and the intersection with the X-axis is called the cut-off point. The slope of the DC load line is the resistance value of R_C.

The intersection of the DC load line with the output characteristic curve of the transistor is called the quiescent operation point, or Q point. When the transistor works at the quiescent operation point, no matter how the AC signal in the input signal changes, the emitter junction will be forward biased and the collector junction will be reverse biased, that is, the transistor will work in a stable amplification state without nonlinear distortion. Therefore, when selecting the appropriate Q point, it should be kept as far away as possible from the saturation region of the transistor to avoid saturation distortion, and also away from the cut-off region of the transistor to avoid cut-off distortion, so that the transistor can exert its best amplification effect.

The power output of the transistor mainly depends on the reverse current of the collector junction J₁, so its output characteristic curve is very similar to the static characteristic curve under the reverse bias of the power diode, as shown in Figure 16. However, unlike power diodes, transistors have three operating states. In order to understand the relationship between the output current I_C and the input current I_B more intuitively, we can extend the X-axis to the left, and treat the left part of the X-axis as the positive half axis of the I_B, and then plot the input-output current characteristic curve based on the relationship between I_C and I_B. If we project this input-output current characteristic curve from the second quadrant to the DC load line of the first quadrant, then the corresponding projection point are the Q points of the transistor. Based on the slope change, the input-output current characteristic curve can be divided into Part 0, Part A, Part B, and Part C, which respectively correspond to the four operating regions of the transistor.

Cut-off Region (Part 0): When U_BE ≤ U_TO or I_B = 0, the emitter junction is in the off state. At this point, the transistor is in the off state even though the collector junction is reverse biased (U_BC < 0), however, strictly speaking, there is still a very small penetration current I_CEO in the transistor. In the same vein, if the collector junction is in the off state (I_C = 0), the transistor will not turn on even if the emitter junction is forward biased (U_BE > 0). Therefore, we can further conclude that the cut-off condition of the transistor is, I_C * I_B = 0.

Active Region (Part A): If the emitter junction is forward biased and greater than the threshold voltage (U_BE > U_TO > 0), then I_B > 0, in which case the transistor operates in amplification state if the collector junction is reverse biased (U_BC ≤ 0). At this point, the value of I_C is not related to U_CE, but only affected by I_B, and there is a linear relationship between I_B and I_C, I_C = β * I_B, which is the reason why the active region is also known as the amplification region.

Saturation Region (Part B and Part C): As the base current increases, the number of holes in the base region decreases, and the carriers injected into the base region from the emitter region also decreases, and the depletion layer of the base region widens. When the saturation boundary is reached, the amplification capability of the transistor begins to weaken (β＇ = ΔI_C / ΔI_B < β), and there is no longer a linear relationship between I_B and I_C, then the transistor enters the quasi-saturation state or the shallow saturation state (i.e., Part B). When the number of holes in the base region drops to a critical value, the potential of the base region is the same as that of the collector region, that is, the collector junction J₁ is zero-biased (U_BC = 0), so that the base current completely loses the amplification effect (β＇ = ΔI_C / ΔI_B = 0), then the transistor enters the full saturation state or the deep saturation state (i.e., Part C).

If the transistor is shallowly saturated, the base current I_B is small and the on-state voltage drop is large, that is, the equivalent resistance of the transistor is large, so it is easy to exit the saturation state. If the transistor is deeply saturated, the base current I_B is large and the on-state voltage drop is small, that is, the equivalent resistance of the transistor is small, then as I_B increases, the saturation degree of the transistor will continue to deepen, so it is difficult to exit the saturation state. In practice, when I_B(sat) = E_C / (β * I_CM), the transistor can be considered to have entered the deep saturation state, which is commonly known as the saturation state. Sometimes in order to accelerate the process of entering the deep saturation state of the transistor, we can apply a base current several times that of I_B(sat) to the transistor. It should be noted that the operating status of the transistor is also affected by the output resistance R_C, that is, the smaller the output resistance R_C, the greater the saturation current I_C, the greater the saturation voltage drop U_CE, and the greater the saturation trigger current I_B. As the output resistance RC decreases, the transistor is prone to burnout if the saturation current I_C is close to the I_CM. Furthermore, if the output resistance R_C is close to 0, the transistor will never enter the saturation state even if it burns out. From this it can be seen that in order to make it easier for the transistor to enter the saturation state, a large output resistance R_C is required.

3- Temperature Characteristics of Transistors

The increase in temperature will cause the intrinsic thermal excitation of the semiconductor, which will increase the carrier concentration inside the semiconductor and increase its conductivity. The increase of conductivity will cause many effects, such as an increase in leakage current, a decrease in the threshold voltage, and an increase in current gain, therefore, as the temperature increases, the input characteristic curve of the transistor shifts to the right, and the output characteristic curve of the transistor shifts upward. Besides, the increase of temperature will also increase the likelihood of thermal breakdown of the transistor, and as a consequence, in practical use, we should equip the transistor with sufficient heat dissipation conditions to make it work at a suitable temperature.

4- Safe Operating Area of Transistors

If we know the model number of a transistor, then we can get its P_CM parameter through its datasheet. Then we can plot the P_CM curve according to the formula P_CM = I_C * U_CE. The area enclosed by the I_CM, U_CEM, and P_CM curves is the safe operating area (SOA) of the transistor, and the transistor can operate stably and safely in this area without damage. Outside the safe operating area is the hazardous area, and if the transistor is operating in this area, its temperature will rise significantly and it will be highly susceptible to thermal breakdown. So, on the one hand, we should avoid the transistor operating in the hazardous area. On the other hand, we should choose a transistor with higher capacity to prevent unexpected situations that could cause the transistor to enter the hazardous area, such as inrush current, inrush voltage, induced current and induced voltage.

3.4.2 Dynamic Characteristics of Transistors

1- Turn-on process of Transistors

Once the turn-on condition (U_BE > U_TO) is met, the transistor will be turned on. The turn-on process of the transistor is divided into the delay time t_d, the rise time t_r, and the spread time t_s.

The delay time t_d is the time taken from 10% I_B1 to 10% I_C1. The barrier capacitor is charged during this time period.

The rise time t_r is the time taken for I_C to rise from 10% I_C1 to 90% I_C1. I_C rises sharply during this time period.

The spread time t_s is the time taken for I_C to rise from 90% I_C1 to 100% I_C1. The diffusion capacitor is charged during this time period.

To sum up, the calculation formula of the turn-on time is, t_on = t_d + t_r + t_s.

2- Turn-off process of Transistors

Once the cut-off condition (I_B = 0) is met, the transistor will be turned off. The turn-off process of the transistor is divided into the storage time t_s, the fall time t_f, and the tail time t_t.

The storage time t_s is the time taken from 90% I_B1 to 90% I_C1. During this time period, the carriers stored in the base region during saturated conduction process will be removed.

The fall time t_f is the time taken for I_C to fall from 90% I_C1 to 10% I_C1. I_C falls sharply during this time period.

The tail time t_t is the time taken for I_C to fall from 10% I_C1 to I_CEO. The remaining carriers recombine during this time period.

To sum up, the calculation formula of the turn-off time is, t_off = t_s + t_f + t_t.

§4. What is an Half-controlled Device?

4.1 Introduction to Half-controlled Devices

The half-controlled device (also known as thyristor, or Silicon Controlled Rectifier, SCR) is a bipolar device that can be turned on by a control signal but can not turned off by it. Half-controlled devices were invented in 1956 and were widely used in the 1960s and 1970s. However, with the invention of fully-controlled devices in the 1980s, the role of half-controlled devices was gradually replaced by them. Even so, the half-controlled devices still retain an important position in large-capacity applications due to their simple structure, reliable operation and ability to withstand very large voltages and currents. Similar to the transistor, the thyristor has three terminals, which are called anode A, cathode K, and gate G, but the thyristor has a four-layer structure (PNPN), which has one more PN junction than the transistor, as shown in Figure 20. We can turn on the thyristor by applying a control signal (also known as a gate trigger signal) to the gate of the thyristor, that is gate triggerring, but it should be noted that we cannot turn off the thyristor by gate triggering due to its special structure. Moreover, when the anode voltage rises to a very high value, it causes an avalanche effect that makes the reverse biased PN junction in middle to break down to turn on the thyristor. In addition, when the anode voltage rise rate dv/dt is too high, the thyristor can be turned on by the junction capacitance effect of the PN junction. Besides the above, thyristors can also be turned on by high junction temperature and light triggering, but in general, only the gate triggering is the most accurate, fastest, and most reliable of all control methods. According to the shape, thyristors can be divided into bolt type (the bolt is an anode and can be tightly connected to the heat sink for easy installation) and flat type (the flat thyristor is clamped in the middle by two heat sinks), however, with the development of semiconductor technology, modular thyristors are now very common (click to view more thyristor modules).

4.2 How does the Thyristor work?

4.2.1 Basic Structure of Thyristors

The thyristor can be divided into P+ region, P region, N+ region and N- region by the different doping degrees of the P-type semiconductor and N-type semiconductor inside the thyristor, as shown in Figure 21(a). The P+ and N+ regions have high doping concentrations and their functions are similar to the emitter region of the transistor. The P+ region is connected to the anode A of the thyristor, and the N+ region is connected to the cathode K of the thyristor. The P and N- regions have low doping concentrations and their functions are similar to the base region of the transistor, furthermore, there are two P+ regions with high doping concentration in the P base region, which are connected to the gate G of the thyristor. Functionally, the thyristor can be regarded as composed of PNP transistor V₁ and NPN transistor V₂, as shown in Figure 20. For V₁, the P+ region is its emitter region, the N- region is its base region, and the P region is its collector region; for V₂, the N+ region is its emitter region, the P region is its base region, and the N- region is its collector region. It should be noted that, similar to the transistor, the wiring of the thyristor is strictly regulated, the anode and cathode are connected to the positive and negative poles of the power supply respectively, and if the two are reversed during use, the thyristor will be burned out.

4.2.2 Working Principle of Thyristors

In the equivalent operating circuit, as shown in Figure 21(b), V₁ and V₂ are the equivalent transistors of the thyristor, futhermore, α₁ is the common base current gain of V1, and α₂ is the common base current gain of V2. The anode A and cathode K of the thyristor are connected to the output circuit, and the gate G of the thyristor is connected to the input circuit. E_A is the power supply of the output circuit, and E_G is the power supply of the input circuit. R is the output resistance. Moreover, IC1 is the collector current of V1 and IC2 is the collector current of V2. The current flowing through the anode A is the anode current I_A, the current flowing through the cathode K is the cathode current I_K, and the current flowing through the gate G is the gate current I_G. In general, the idea of turning on the thyristor is similar to that of turning on the transistor, that is, how to make the PN junction J₂ produce a sufficiently large reverse current, but unlike the transistor, the thyristor has only two operating states.

Cut-off State: When the forward bias voltage U_AK is applied to the anode A and cathode K of the thyristor, if no voltage is applied to the gate G, it is equivalent to the collector and base of V₁ and V₂ being open circuit, then the thyristor will be in the cut-off state. The depletion layers of J₁ and J₃ are narrowed and the depletion layer of J₂ is widened under the action of the forward bias voltage, therefore, even if the thyristor is in the cut-off state, there is a reverse saturation current I_CBO in J₂, which consists of two parts, one is the hole current I_CBO1 (i.e., the common base current of V₁) and the other part is the free electron current I_CBO2 (i.e., the common base current of V₂). These two currents will flow through J₁ and J₃ and become the leakage current of the thyristor. Because α₁ + α₂ is very small in the cut-off state, the leakage current value of the thyristor is slightly greater than the sum of the leakage current values of the equivalent transistors V₁ and V₂, as shown in Formula 14.

Conduction State: When a forward bias voltage is applied to the gate of the thyristor, the P+ region connected to the gate injects a large number of holes into the P base region, part of which enters the N+ region, making J₃ forward conduction, while at the same time, a large number of free electrons are injected from the N+ region into the P base region, which increases the minority carrier concentration in the P base region and thus increases I_CBO2, and another part of which enters the N- region, which increases the minority carrier concentration in the N- region and thus increases I_CBO1. Both leakage currents will reduce the minority carriers in the N- region and narrow the depletion layer of J₁. When a sufficiently large forward bias voltage is applied to the gate, the depletion layer of J₁ will be narrowed to a critical state so that the dynamic equilibrium of J₁ will be broken, and a large number of holes are injected into the N- region from the P+ region connected to the anode and then injected into the P base region, forming the current I_C1, and similarly, a large number of free electrons in the N+ region connected to the cathode are injected from the P base region into the N- region and then injected into the P+ region, forming the current I_C2. It is important to note that if I_C1 and I_C2 are not formed, then the conductivity of the base region will be very small, therefore α₁ + α₂ will be very small. However, once I_C1 and I_C2 are formed, the conductivity of the base regions of V₁ and V₂ will increase due to the conductance modulation effect, making I_C1 and I_C2 increase further. With this positive feedback, α₁ + α₂ rapidly increases to a value close to 1, causing the on-state voltage drop to drop dramatically and the anode current I_A to rise sharply, eventually causing the thyristor to be turned on.

* Calculation Formula of Thyristor

I_C1 = α₁ * I_A + I_CBO1, (10)

I_C2 = α₂ * I_K + I_CBO2, (11)

I_K = I_A + I_G, (12)

I_A = I_C1 + I_C2, (13)

I_A = (α₂ * I_G + I_CBO1 + I_CBO2) /[1 - (α₁ + α₂) ]. (14)

It can be seen from Formula 14:

If α₁ + α₂ approaches 0, I_A will tend to leakage current, that is, the thyristor is in the cut-off state.

If α₁ + α₂ approaches 1, I_A will tend to infinity, that is, the thyristor is in the conduction state.

4.3 Main Parameters of Thyristors

4.3.1 Static Parameters (Voltage)

1- Forward Non-repetitive Peak Voltage U_DSM / Reverse Non-repetitive Peak Voltage U_RSM

Under the conditions of the rated junction temperature and the open-circuited gate, the voltage that is not allowed to be applied to the anode and cathode of the thyristor repeatedly is called non-repetitive peak voltage, one of which is the forward non-repetitive peak voltage U_DSM (also known as the maximum off-state transient voltage) determined by the sharp bend point of the forward volt-ampere characteristic curve, and the other is the reverse non-repetitive peak voltage U_RSM (also known as the maximum reverse transient voltage) determined by the sharp bend point of the reverse volt-ampere characteristic curve.

2- Forward Turning Voltage U_BO

Under the conditions of the rated junction temperature and the open-circuited gate, if a forward sinusoidal half-wave voltage applied between the anode and cathode of the thyristor is capable of switching the thyristor directly from the off-state to the on-state, then this voltage is known as the forward turning voltage U_BO.

3- Reverse Breakdown Voltage U_BR

Under the condition of the rated junction temperature, if a reverse sinusoidal half-wave voltage applied between the anode and cathode of the thyristor is capable of causing a sharp increase in the reverse leakage current of the thyristor, then this voltage is known as the reverse breakdown voltage U_BR.

4- Forward Repetitive Peak Off-state Voltage U_DRM / Reverse Repetitive Peak Off-state Voltage U_RRM

Under the conditions of the rated junction temperature and the open-circuited gate, the forward repetitive peak off-state voltage U_DRM (also known as the repetitive peak off-state voltage) is the forward peak voltage allowed to be repeatedly applied to the anode and cathode of the thyristor. The repetition frequency of U_DRM is 50 times per second, with each duration not exceeding 10 ms. The voltage value of the U_DRM is generally specified as 90% of the voltage value of the U_DSM. Also, for safety reasons, the voltage value of the U_DRM should be about 100V less than the voltage value of the U_BO.

Under the conditions of the rated junction temperature and the open-circuited gate, the reverse repetitive peak off-state voltage U_RRM (also known as the reverse repetitive peak voltage) is the reverse peak voltage allowed to be repeatedly applied to the anode and cathode of the thyristor. The repetition frequency of U_RRM is 50 times per second, with each duration not exceeding 10 ms. The voltage value of the U_RRM is generally specified as 90% of the voltage value of the U_RSM. Also, for safety reasons, the voltage value of the U_RRM should be less than the voltage value of the U_BR.

It is important to note that the values of U_DRM and U_RRM are not fixed, and they decrease as the temperature increases. Therefore, the temperature of the thyristor should be strictly controlled by various heat dissipation methods during testing and use in order to keep the U_DRM and U_RRM within a reasonable range. In addition, we usually choose the smaller one of U_DRM and U_RRM as the rated voltage of the thyristor to ensure that the thyristor can operate safely and efficiently.

5- Gate Trigger Voltage U_GT

Under the conditions of the specified ambient temperature and the forward bias of the thyristor, the minimum gate DC voltage required to be able to switch the thyristor from the off-state to the on-state is known as the gate trigger voltage U_GT. Typically, the U_GT of a thyristor is 1-1.5V, but it is common to apply 4-10V to the gate as a trigger voltage in order for the thyristor to be reliably turned on.

6- Forward Average Voltage Drop U_F

Under the conditions of the specified ambient temperature and the specified heat dissipation method, if the on-state current of the thyristor is equal to the rated current, the average value of the voltage drop between the anode and cathode of a thyristor is known as the forward average voltage drop U_F (or the on-state average voltage, and the on-state voltage drop). Typically, the U_F of a thyristor is 0.4-1.2V.

7- Peak On-State Voltage U_T

Under the conditions of the specified ambient temperature and the specified heat dissipation method, if the on-state current of the thyristor is equal to the peak on-state current, the peak transient voltage between the anode and cathode of the thyristor is called the peak on-state voltage U_T (or the peak on-state voltage drop), which is typically 2V.

4.3.2 Static Parameters (Current)

1- Rated On-state Current I_T

Under the conditions of the specified ambient temperature and the specified heat dissipation method, if the load of the thyristor is a resistor and the conduction angle of the thyristor is not less than 170°, the maximum power frequency sinusoidal half-wave current allowed to flow through the thyristor is known as the rated on-state current I_T. Furthermore, if the current waveform is not a power frequency sinusoidal half-wave, even though the thyristor is a semiconductor that does not have the same volt-ampere characteristic curve as the resistor, we can use the equivalent resistor, which has the same heating effect as the thyristor, as a reference to determine the rated on-state current of the thyristor, which is generally 1.5-2 times the value of the current flowing through the equivalent resistor. The rated on-state current can be used in a variety of ways, for example, the average value of the rated on-state current I_T(AV) can be used as the rated current for a unidirectional thyristor, the effective value of the rated on-state current I_T(RMS) can be used as the rated current for a bi-directional thyristor, and n times the rated on-state current (usually n=3) can be used as the peak on-state current I_TM of the thyristor.

2- Off-state Leakage Current I_DRM / Reverse Leakage Current I_RRM

I_DRM is the leakage current corresponding to U_DRM, and I_RRM is the leakage current corresponding to U_RRM, and both of them are generally less than 100μA.

3- Gate Trigger Current I_GT

Under the conditions of the specified ambient temperature and the forward bias of the thyristor, the minimum gate DC current required to be able to switch the thyristor from the off-state to the on-state is known as the gate trigger current I_GT. The I_GT of an ordinary thyristor is typically a few milliamps, and the I_GT of a high-sensitivity thyristor is typically a few microamps.

4- Holding Current I_H

The holding current I_H is the minimum current required to maintain the on-state of the thyristor, which is typically tens to hundreds of milliamps. The gate of the thyristor does not have the ability to turn off the thyristor, so once the thyristor is turned on, it will still remain in the on-state even if the gate trigger signal is removed. However, as long as the anode current of the thyristor is reduced below I_H, the thyristor can be turned off. Moreover, I_H is susceptible to the junction temperature of the thyristor, the higher the junction temperature, the smaller the I_H, and the less likely the thyristor is to be turned off. Therefore, in order to better control the operating state of the thyristor, a good heat dissipation method is necessary.

5- Latching Current I_L

The latching effect, or self-locking effect, refers to the phenomenon that the thyristor still remain in the on-state even if the gate trigger signal is removed due to the positive feedback process in the thyristor. When the thyristor has just been turned on and the gate trigger signal is immediately removed, the minimum current that can keep the thyristor in the on-state is known as the latching current I_L, which is generally 2-4 times that of I_H.

6- Inrush Current I_TSM

In the half-cycle of the power frequency sinusoidal wave, the non-repetitive maximum forward overload current caused by the circuit abnormality is known as the inrush current I_TSM, which will cause the junction temperature of the thyristor to exceed its rated junction temperature. Typically, during one positive cycle of the power frequency sinusoidal wave, the thyristor can withstand an inrush current of up to 6 times its rated current. However, there is a limit to the number of times a thyristor can withstand the inrush current throughout its service life, and if it is exceeded, the thyristor may be permanently damaged, so when choosing a thyristor, it is necessary to fully consider the inrush current that may exist in the circuit.

7- Forward Turning Current I_BO

Under the conditions of the rated junction temperature and the open-circuited gate, the anode current that can switch the thyristor directly from the off-state to the on-state is called the forward turning current I_BO.

4.3.3 Dynamic Parameters

1- Turn-on time t_gt

The turn-on time t_gt is the time it takes for the thyristor to switch from the off-state to the on-state after a gate trigger signal is applied. During the turn-on process, the output voltage U_AK of the thyristor will gradually decrease to the on-state voltage drop U_F, and its anode current I_A will gradually increase to the rated on-state current I_T.

2- Turn-off time t_q

The turn-off time t_q is the time it takes for the anode current I_A of the thyristor to decrease from the rated on-state current I_T to 0 until the thyristor begins to withstand the specified off-state voltage. The t_q is not only related to the internal structure of the thyristor, but also to temperature, dv/dt and di/dt. The t_q of an ordinary thyristor is about a few hundred milliseconds, and it can be further reduced by increasing the reverse voltage. During t_q, the thyristor is not completely turned off, so if the anode voltage is reapplied, the thyristor can be turned on again, but after t_q, the thyristor will not be turned on no matter how the anode voltage is increased on the premise that the thyristor is not reverse broken down.

The t_gt and t_q together determine the operating frequency of the thyristor. Generally, if the t_q of a thyristor is small, the t_gt will be smaller than it, so this parameter makes it possible to distinguish between normal and fast thyristors. In summary, if we want to choose a fast thyristor for a high-frequency switching circuit, then simply choose the one with a small t_q.

3- Critical Off-state Voltage Rise Rate dv/dt

Under the conditions of the rated junction temperature and the open-circuited gate, the maximum rise rate of the output voltage of the thyristor from the off-state to the on-state is known as the critical off-state voltage rise rate dv/dt. The dv/dt affects the safe and stable operation of the thyristor, for example, for a thyristor with a large dv/dt, if the charging current of its junction capacitor is large, then the thyristor is prone to be mis-conducted. The dv/dt of small-current thyristors (50-100A) is typically 225V/μs, and the dv/dt of high-current thyristors (200A or more) is typically 50V/μs.

4- Critical On-state Current Rise Rate di/dt

Under the conditions of the rated junction temperature and the close-circuited gate, the maximum rise rate of the on-state current that a thyristor can withstand is called the critical on-state current rise rate di/dt. The thyristor generates a large power loss at the turn-on instant, which is always concentrated in the cathode region near the gate due to its limited conduction expansion capability. Therefore, if di/dt is too large, then even if the on-state current is not too large, the thyristor can be permanently damaged by localized overheating of the gate. Furthermore, the larger the rated current of the thyristor, the more obvious this problem.

4.4 Basic Characteristics of Thyristors

4.4.1 Static Characteristics of Thyristors

The static characteristics of a thyristor is the volt-ampere characteristics of its output current and output voltage. The static characteristics curve of the thyristor is shown in Figure 22, where I_G is the gate trigger current, I_A is the anode current (i.e., output current), and U_AK is the voltage applied to the anode and cathode of the thyristor (i.e., output voltage).

1- Forward static characteristics

Forward Blocking State: When I_G = 0, α₁ + α₂ is very small, and even if U_AK > 0, there is only a small forward leakage current. This operating state of the thyristor is known as the forward blocking state. However, when U_AK ≥ U_BO or I_A ≥ I_BO, α₁ + α₂ approaches 1 and the thyristor begins to enter the forward conduction state.

Forward Conduction State: When I_G > 0 and U_AK ≥ U_GT, the conductivity of the base region of the thyristor increases significantly until α₁ + α₂ is close to 1, the anode current I_A will tend to infinity, and the thyristor enters into the saturated conduction state, that is, the forward conduction state. Once the thyristor is in forward conduction state, the gate loses its ability to control the thyristor, then the thyristor can only be turned off when I_A decreases close to 0. It should be noted that although the theoretical value of I_A tends to infinity when α₁ + α₂ is close to 1, its actual value is determined by the external circuit. In addition, under the same external conditions, the larger the I_G, the smaller the U_GT.

2- Reverse Static Characteristics

The reverse static characteristics of the thyristor is similar to those of the power diode. When a reverse voltage is applied to the thyristor, no matter whether there is a gate trigger current or not, the thyristor will not be turned on, but there will be a small reverse leakage current. This operating state of the thyristor is known as the reverse blocking state. However, the reverse blocking state does not mean that the thyristor can completely block the reverse voltage, if the reverse voltage reaches the reverse breakdown voltage U_BR, then it will cause the avalanche breakdown, so that the thyristor will lose the reverse blocking ability due to the internal short circuit.

4.4.2 Dynamic Characteristics of Thyristors

1- Turn-on Process

When U_AK1 is applied to the output of the thyristor, since the thyristor is not turned on, the output voltage U_AK of the thyristor is 100% of U_AK1. When U_G ≥ U_GT, the thyristor enters the conduction state after a series of turn-on processes. After the thyristor is turned on, U_AK will remain at a very small value, that is, the on-state voltage drop U_F.

The turn-on process of the thyristor is divided into the delay time t_d, the rise time t_r, the and the spread time t_s.

The delay time t_d is the time it takes for I_A to rise from the off-state leakage current I_DRM to 10% of I_A1 and for U_AK to fall from 100% of U_AK1 to 90% of U_AK1. The delay time t_d is typically 0.5-1.5 μs. In addition, the delay time t_d decreases as the gate current increases.

The rise time t_r is the time it takes for I_A to rise from 10% of I_A1 to 90% of I_A1 and for U_AK to fall from 90% of U_AK1 to 10% of U_AK1. The rise time t_r is typically 0.5-3 μs. The rise time t_r is affected by the characteristics of the thyristor itself, external circuit impedance, temperature, and anode voltage, furthermore, the delay time t_d and rise time t_r can be significantly shortened by increasing I_A.

The spread time t_s is the time it takes for I_A to rise from 90% of I_A1 to 100% of I_A1 and for U_AK to fall from 10% of U_AK1 to on-state voltage drop U_F. In addition, the spread time t_s depends on the cross-sectional area of the cathode.

In general, the thyristor can be considered to have been turned on when I_A reaches 90% of I_A1. Therefore, the calculation formula for the turn-on time is: t_gt = t_d + t_r.

2- Turn-off Process

If the U_AK of the thyristor is reduced to 0 or a large enough reverse voltage U_AK2 is applied to the thyristor so that I_A gradually decreases to 0, then the thyristor can be switched from the on-state to the off-state. The turn-off process of the thyristor is divided into the reverse blocking recovery time t_rr, and the forward blocking recovery time t_gr.

During the turn-off process, the inductance of the external circuit generates a reverse current (or reverse recovery current) I_R in the thyristor. The reverse voltage U_R of the thyristor increases with the increase of I_R, and when the I_R reaches its peak I_RP, the U_R also reaches its peak U_RP, and then both I_R and U_R decrease rapidly. The reverse blocking recovery time t_rr is the time it takes for the I_A to go from 10% of I_A1 to the reverse leakage current, and it is also the time it takes for the U_AK to go from 10% of U_AK1 to U_AK2.

The forward blocking recovery time t_gr (or gate recovery time) is the time it takes for a thyristor to fully regain its forward blocking capability from the end of its reverse recovery process. Since there are still a small number of carriers remaining on the PN junction near the gate during t_gr, the positive feedback mechanism of the thyristor is still active. At this point, if a forward bias voltage is applied to the thyristor, it will enter the forward conduction state again. This trigger mode does not require any gate trigger signal.

Therefore, the calculation formula for the turn-off time is: t_q = t_rr + t_gr.

4.5 Series and Parallel Connection of Thyristor

1- Series Connection of Thyristor

Similar to the series connection of resistors, connecting multiple thyristors in series can increase their overall voltage capability. However, unlike resistors, the voltage is not evenly distributed across the thyristor in series. Due to the different generation mechanisms, this kind of uneven voltage can be divided into static uneven voltage and dynamic uneven voltage. Therefore, the total withstand voltage value of thyristors in series cannot be calculated simply by multiplying the withstand voltage value of one individual thyristor by their number, but should be calculated by adding up the actual voltage withstood by each thyristor.

Static Uneven Voltage: Even though the leakage currents flowing through thyristors in series are the same, the static voltage applied to each thyristor is different because their static volt-ampere characteristics are inconsistent and dispersed. Even in some extreme cases, one thyristor will withstand almost all the voltage, while the others will only withstand a very small voltage. In order to effectively reduce the negative impact of static uneven voltage on the operating efficiency and service life of thyristors, the specifications and static volt-ampere characteristics of thyristors in series should be as consistent as possible. In addition, the unevenness of the static voltage can be reduced by the resistor equalization method, that is, treating the thyristor as a high-resistance resistor (about 1 megohm), then giving each thyristor in the series circuit a low-resistance resistor in parallel, and the static voltage applied to each thyristor can be evenly distributed by the means of adjusting the equivalent resistance value of each thyristor by adjusting the resistance value of the low-resistance resistor.

Dynamic Uneven Voltage: During the switching process of thyristors, the dynamic voltages applied to each thyristor in the series circuit are different due to their inconsistent volt-ampere characteristics. In order to effectively reduce the negative impact of dynamic uneven voltage on the operating efficiency and service life of thyristors, the specifications and dynamic volt-ampere characteristics of thyristors in series should be as consistent as possible. In addition, by applying a sufficiently large gate trigger signal, the on-time difference between thyristors in series can be significantly reduced, which also contributes to the reduction of the dynamic uneven voltage. Of course, by connecting a RC circuit in parallel to each thyristor in the series circuit to absorb the overvoltage, the dynamic voltage applied to each thyristor can be evenly distributed.

2- Parallel Connection of Thyristor

Similar to the parallel connection of resistors, connecting multiple thyristors in parallel can increase their overall current capability. Due to differences in parameters and volt-ampere characteristics, the current flowing through thyristors in parallel is not evenly distributed. Therefore, by selecting thyristors with as consistent parameters and volt-ampere characteristics as possible, the dynamic and static uneven currents of thyristors in parallel can be effectively reduced. Besides, the dynamic uneven current can be effectively reduced by equipping with a current sharing reactor, whose current loss is less than that of the resistor. Moreover, the on-time difference between thyristors in parallel can be significantly reduced by applying a sufficiently large gate trigger signal, so that each thyristor can be effectively triggered in a short period of time to achieve the dynamic current balancing. However, with the development of semiconductor technology, the current capacity of the thyristor is getting larger and larger, so in practical use, there is no need to connect thyristors in parallel.

3- Series-Parallel Connection of Thyristor

In some cases, we need to connect multiple thyristors in series and parallel into a single circuit at the same time, but since the on-state voltage drop of the thyristors is small, it is recommended to connect the thyristors in series first to reduce the difference in the on-state voltage drop between the thyristors, and then connect the thyristors in parallel to achieve current balancing.

4.6 Main Types of Thyristor

1- Fast Switching Thyristor

The fast switching thyristor (FST) has excellent dynamic characteristics. Compared with ordinary thyristors, FST has the advantages of short turn-on time (typically 4-8 μs), short turn-off time (typically 10-60 μs), and large tolerances for dv/dt and di/dt. In addition, ordinary thyristors can only operate at a frequency of 50Hz, while FST can operate at a higher frequency (above 400Hz). There is also a kind of fast-switching thyristor called high-frequency thyristor (HFT), which has a shorter switching time and faster switching speed, which is more suitable for working in high-frequency circuits (above 10kHz). Because of the high operating frequency of FST and HFT, the thermal effect of switching losses should not be ignored, so their rated voltage and rated current are usually not high enough to avoid burnout.

2- Bidirectional Thyristor

The bidirectional thyristor (also known as a triode AC semiconductor switch or TRIAC) can be thought of as a pair of unidirectional thyristors (also known as a silicon controlled rectifier or SCR) connected in anti-parallel. Bidirectional thyristors are core components in AC solid state relays and AC modules click to view more AC solid state relays). The bidirectional thyristor has the same forward characteristics as the unidirectional thyristor, but its reverse characteristics are different from those of the unidirectional thyristor. This is because a bidirectional thyristor does not have reverse blocking capability, but instead has a capability similar to forward conduction. Therefore, the volt-ampere characteristic curve of the bidirectional thyristor is center-symmetrical on the coordinate axis. The bidirectional thyristor has a T1 pole (the main electrode connected to the P-type semiconductor material), a T2 pole (the main electrode connected to the N-type semiconductor material), and a G pole (Gate pole). Since the bi-directional thyristor is a AC switch, its current rating is the effective value of the rated on-state current, i.e., I_T(RMS). Moreover, there is no forward peak voltage and reverse peak voltage in the parameters of bi-directional thyristors, only the maximum peak voltage. Other than these, most of the parameters of a bidirectional thyristor are the same as those of a unidirectional thyristor.

3- Reverse Conducting Thyristor

Similar to the design idea of bidirectional thyristors, a reverse conducting thyristor (RCT) can be obtained by connecting a unidirectional thyristor and a freewheeling power diode in reverse parallel. This special structure of the reverse conducting thyristor puts the emitter junctions at both the anode and cathode in a short-circuit state, so that the reverse conducting thyristor has a shorter turn-off time (a few microseconds) and a higher operating frequency (tens of kilohertzes) than that of the fast switching thyristor. Because the reverse conducting thyristor has the advantages of low on-state voltage drop, short turn-off time, high rated junction temperature, high voltage resistance and high temperature resistance, applying the reverse conducting thyristor to applications such as switching power supplies and UPSs helps to simplify the circuit design.

4- Light Triggered Thyristor

The light-triggered thyristor (LTT, also known as light-controlled thyristor) is a type of thyristor that uses optical signals as a trigger method. The gate region of the light triggered thyristor integrates a photoelectric power diode as a optical trigger, and in terms of its working principle, it is a gate trigger that simply replaces the gate trigger current with the strength of the optical signal. This trigger method helps to ensure electrical insulation between the main circuit and the control circuit and also avoids the effects of electromagnetic interference. In addition, in order to improve the trigger sensitivity of the light triggered thyristor, the gate region is usually an amplification gate structure or a double amplification gate structure. In practice, low-power light triggered thyristors are often used for electrical isolation, or to provide trigger signals for high-power thyristors. High-power light triggered thyristors are often used to ensure good electrical insulation between the control device and the high-voltage power equipment.

§5. What is a Fully-controlled Device?

5.1 Introduction to Fully-controlled Devices

The fully-controlled device was invented in the 1980s, ushering in a new era of power electronics technology. Unlike the half-controlled device, the fully-controlled device is a self-shut-off device that can be completely turned off by the control signal. There are many types of fully-controlled devices, such as gate turn-off thyristors (e.g., GTO), bipolar junction transistors (e.g., GTR), field effect transistors (e.g., MOSFET, JFET), and composite devices (e.g., IGBT, MCT, SIT, SITH, IGCT). Most of them have unique composite structures with excellent performance and a wide range of applications. Below we will give a brief introduction to them.

5.2 Gate Turn-off Thyristor

5.2.1 Introduction to GTO

Shortly after the advent of the thyristor, the gate turn-off thyristor (GTO) appeared as a derivative of the thyristor. However, even so, the two work on completely different principles. For example, the SCR is a half-controlled device that once turned on, the gate signal no longer has any effect, whereas GTO is a fully-controlled device that can be turned off by applying a negative pulse gate signal even after it is turned on.

5.2.2 How does the GTO work?

5.2.2.1 Basic Structure of GTO

Since GTO is a derivative of SCR, its structure is very similar to that of SCR, however, GTO has one more N+ buffer region than SCR, which results in a shorter turn-on time and weaker reverse blocking ability than those of SCR. Of course, in terms of actual construction, a GTO tube is a multi-unit power device that contains dozens or even hundreds of GTO units. The cathode and gate of these GTO units are connected in parallel inside the GTO tube and share a common anode. Since the cathode region of each GTO unit is small, which dramatically shortens the distance between the cathode and the gate, making the lateral resistance of the P2 base region small, and thus allowing greater current to be drawn from the gate region. In this case, as long as the reverse current of the gate is large enough, the collector of V₁ is prone to be turned off, and thus the GTO is also prone to be turned off. This special structure of the GTO tube gives it a higher capacity density and a shorter turn-on time than those of SCR, but its reverse blocking ability is weaker than that of SCR, only 20-30V. Moreover, the GTO tube has a stronger di/dt withstand capacity and overload withstand capacity, for example, if the di/dt is too large, the SCR will burn out once it is locally overloaded and overheated, because it has only one unit. However, for GTO tube, the overload and overheating caused by di/dt will be distributed to its internal GTO units, then even if some of the GTO units are burned out, the other can continue to work, and thus GTO is widely used for high power applications above the megawatt level. Besides, GTO does not require a commutation circuit, so it can be used in devices above 1kHz, while SCR can only be used in devices below 1kHz.

In order to optimize the switching characteristics of the GTO, a power diode is typically connected in anti-parallel to it. For ease of design and use, this power diode will be integrated directly into the GTO and form a reverse conducting GTO, which is somewhat similar in structure to a reverse conducting thyristor (RCT). It should be noted that the reverse conducting GTO no longer has the ability to withstand reverse voltages. If the reverse conducting GTO needs to withstand a large reverse voltage, a power diode needs to be connected in series to it.

5.2.2.2 Working Principle of GTO

The turn-on principle of the GTO is very similar to that of SCR, and both are triggered by the positive feedback mechanism of the gate signal. However, the GTO has an additional N+ region, which reduces its conductivity and helps to speed up this positive feedback process. The α₁ of the GTO is designed to be very small, e.g., α₁ + α₂ ≥ 1.15 when the SCR is turned on, whereas α₁ + α₂ ≈ 1.05 when the GTO is turned on, and thus the saturation depth of the GTO is shallower than that of the SCR, and closer to the critical saturation state. This design makes the equivalent transistor V₂ more sensitive to the gate control signal and easier to be turned on and off, but as a result, the on-state voltage drop of the GTO will be higher than that of the SCR. If the GTO is to be turned off, only a reverse current large enough to be given to the gate to drain the holes in the P base region and inject a large number of free electrons into the N- base region, then as the holes injected into the base region from the P+ region gradually decrease, I_C1 and I_C2 will also decrease. After a series of positive feedback processes, until the anode current I_A is less than the holding current I_H, the GTO will be completely turned off because of exiting the saturation state.

* Calculation Formula of GTO

The calculation formula of the GTO is the same as that of the SCR.

I_C1 = α₁ * I_A + I_CBO1, (10)

I_C2 = α₂ * I_K + I_CBO2, (11)

I_K = I_A + I_G, (12)

I_A = I_C1 + I_C2, (13)

I_A = (α₂ * I_G + I_CBO1 + I_CBO2) / [1 - (α₁ + α₂)]. (14)

|I_GRP| > (α₁ + α₂-1) * I_ATO / α₂, (15)

β_off = I_ATO / |I_GRP|. (16)

As can be seen in Equation (14):

When α₁ + α₂ approaches 0, I_A will tend to leak current;

When α₁ + α₂ approaches 1, I_A will tend to infinity.

5.2.3 Main Parameters of GTO

Most of the parameters of the GTO are the same as those of the SCR.

1- Turn-on time t_on

The turn-on time t_on is the sum of the delay time t_d and the rise time t_r.

2- Turn-off time t_off

The turn-off time t_off is the sum of the storage time t_s and the fall time t_f.

3- Maximum Turn-off Anode Current I_ATO

The maximum turn-off anode current I_ATO, also known as the maximum controllable anode current, is the rated current of the GTO. If the anode current I_A is greater than I_ATO, α₁ + α₂ cannot satisfy the controllable condition of slightly greater than 1, so the saturation depth of the GTO will deepen, making it impossible to be turned off normally.

4- Turn-off Current Gain β_off

The turn-off current gain β_off is the ratio of the maximum turn-off anode current I_ATO to the peak gate reverse pulse current I_GRP, that is β_off = I_ATO / I_GRP. However, since the β_off of the GTO is so small (typically 3-8), a large gate reverse pulse current is required to cut the GTO off. For example, assuming β_off = 3, the gate reverse pulse current should be 300A if a GTO rated at 1000A is to be turned off. From the perspective of working principle, the process of turning off the GTO is equivalent to controlling a strong output current with a strong input current, which is disappointing both in terms of power consumption and safety, and thus this major shortcoming limits the application fields of the GTO.

5.2.4 Basic Characteristics of GTO

5.2.4.1 Static Characteristics of GTO

The static characteristics of the GTO are the same as those of the SCR except that the latching current I_L of the GTO (normally 2A) is greater than that of the SCR (normally 100-500mA).

5.2.4.2 Dynamic Characteristics of GTO

1- Turn-on Process

Similar to the conduction process of the SCR, when U_AK = 100% U_AK1 and U_G ≥ U_GT, the GTO will enter the conduction state and a small on-state voltage drop will be generated at its output. Unlike SCR, the GTO requires a larger gate trigger current I_GT due to its multi-unit construction. The delay time t_d of the GTO is about 1-2 μs. The rise time t_r of the GTO increases with the increase of the on-state anode current.

2- Turn-off Process

When a reverse pulse voltage is applied to the gate to provide a sufficiently large reverse pulse current, the GTO will enter the turn-off process. The turn-off process of the GTO is divided into the storage time t_s, the fall time t_f and the tail time t_t.

The storage time t_s is the time it takes for I_A to decrease from 100% of I_A1 to 90% of I_A1. When the gate reverse pulse voltage U_G2 is applied to the gate of the GTO, the gate reverse pulse current generated at the gate extracts the carriers stored in the P base region during saturated conduction, thus causing the equivalent transistor V₂ to exit saturation state. Then, the gate reverse pulse current rises rapidly from 0 to I_GRP. The rise rate of the gate reverse pulse current di/dt depends on the circuit inductance and the anode voltage. During the storage time, U_AK and I_A remain unchanged because GTO has not yet fully exited saturation state, and can be turned on again under the right conditions.

The fall time t_f is the time it takes for I_A to decrease from 90% of I_A1 to 10% of I_A1. When the reverse pulse current reaches the I_GRP, the anode current I_A starts to fall rapidly, and the anode voltage U_AK starts to rise. As the conductivity of the base region decreases, resulting in α₁ + α₂ ≤ 1, the GTO begins to enter the turn-off state. Since the fall time t_f is very short (about 2μs) and the fall rate of I_A is large, a spike voltage appears at the output of the GTO.

The tail time t_t is the time it takes for I_A to fall from 10% of I_A1 to 0. Under the action of the gate reverse pulse voltage, the residual carriers in the P base region will be further recombined, causing U_AK to gradually rise to U_AK1 and I_A to gradually decrease to 0. During this period, a transient overshoot occurs at the output of the GTO due to the snubber circuit and a spike voltage is generated. It should be noted that the rise rate of U_AK should not be too large, otherwise the GTO may be turned on again. In addition, the tail time t_t can be effectively shortened by maintaining an appropriate gate reverse pulse voltage.

In general, the fall time t_f is less than the storage time t_s, and the storage time t_s is less than the tail time t_t, that is, t_f < t_s < t_t. When I_A falls to 10% of I_A1 the GTO can be considered to have been turned off, so the calculation formula for the turn-off time is: t_off = t_s + t_f.

5.3 Giant Transistor

5.3.1 Introduction to GTR

The Giant Transistor (GTR) is a bipolar junction transistor that was invented in the 1970s. Since the switching time of the GTR is very short (typically a few microseconds), the GTR has a higher operating frequency (typically 1-20kHz) than that of the GTO (typically a few hundred hertz). The power capacity, rated voltage and rated current of the GTR is also very high, such as 1800V/800A/2kHz, 1400V/600A/5kHz and 600V/3A/100kHz, so it is also known as Power Transistor, or Power BJT. In addition, the GTR also has the advantages of low saturation voltage, good switching characteristics, wide safe operating area, and strong self-shut-off ability, therefore, the GTR gradually replaces the GTO in the medium capacity and medium frequency fields, such as the power supply, the motor control, and the general-purpose inverter. However, because the GTR requires very high driving power to control its switching state, its driving circuit is designed to be very complex. Besides, the GTR has poor resistance to inrush current, and is susceptible to damage due to secondary breakdown. With the development of power electronics technology, these shortcomings of the GTR have led to its gradual replacement by the power MOSFET and the IGBT.

According to the different structures, the GTR can be divided into the NPN type GTR and PNP type GTR, and the following is illustrated with the NPN type GTR as an example. Depending on the construction method, the GTR can be further divided into the single-tube GTR, the composite-tube GTR（or Darlington GTR）and the GTR module. The single-tube GTR only has one unit with the characteristics of low saturation voltage drop, fast switching speed, small current capacity, and high driving power. Since its small current gain β (usually around 10), the single-tube GTR is generally used in small-capacity inverter circuits. The composite-tube GTR consists of several GTR units that are connected in parallel by integrated circuit technology, each unit consisting of a Darlington tube. The composite-tube GTR has the advantages of large current gain β (usually up to tens or hundreds), large current capacity, and low driving power. However, due to its high saturation voltage drop and slow turn-off speed, the composite-tube GTR is generally used in medium and large-capacity inverter circuits. The GTR module is a combination of two or more single-tube GTRs or composite-tube GTRs encapsulated in an insulating resin housing. By changing the connection of the internal units, the GTR module can be used as a single bridge arm, single-phase bridge, three-phase bridge, and three-phase bridge with a bleeder circuit. With the advantages of easy installation, excellent performance and strong adaptability, the GTR module has gradually replaced the single-tube GTR and the composite-tube GTR, and is widely used in various inverter circuits.

* Darlington tube

The Darlington tube was invented by Sidney Darlington in the 1950s to solve the problem of low values and large deviations of the current gain β of early silicon transistors. As shown in Figure 33, the Darlington tube is a common emitter amplifier consisting of a cascade of a transistor on the left and a transistor on the right. The transistor on the left acts as a current amplifier and is responsible for amplifying the input signal and outputting it to the input of the transistor on the right, while the transistor on the right acts as a voltage amplifier and is responsible for converting the input driving current signal into the output voltage signal. The Darlington tube can be considered as an equivalent transistor of the same type as the transistor on the left, for example, if the transistor on the left is an NPN transistor, then the Darlington tube is also an NPN transistor. If both the transistor on the left and the transistor on the right operate in the amplification region, then the current gain β of the Darlington tube is product of the current gain β of the two transistors.

5.3.2 How does the GTR work?

5.3.2.1 Basic Structure of GTR

The structure of the GTR is similar to that of BJT, taking the NPN type GTR as an example, it is divided into three regions by the collector junction J₁ and the emitter junction J₂, namely the emitter region, the base region and the collector region. The emitter region has the characteristics of small area and high doping concentration, and the base region has the characteristics of thin thickness (5-20 μm) and low doping concentration, while the collector region can be divided into two parts, one is the N- collector drift region with large area and low doping concentration, and the other is the N+ substrate region with small area and high doping concentration.

5.3.2.2 Working Principle of GTR

The working principle and calculation formula of GTR are the same as those of BJT. However, it should be noted that during the turn-on process of the GTR, the N+ substrate region will inject a large number of free electrons into the N- collector drift region to increase the reverse current of J₁.

5.3.3 Main Parameters of GTR

Most of the parameters of GTR are the same as the main parameters of BJT.

1- Breakdown Voltage BV

The breakdown voltage BV is the output voltage of the GTR when the breakdown occurs, and the GTR will not be broken down as long as the output voltage U_CE of the GTR is lower than that of BV. The breakdown voltage BV can be divided into the following types, depending on how the GTR is connected to the external circuit.

● BV_CBO is the reverse breakdown voltage between the collector and the base when the emitter is open.

● BV_CEO is the breakdown voltage between the collector and the emitter when the base is open.

● BV_CER is the breakdown voltage between the collector and the emitter when the emitter and the base are connected by a resistor.

● BV_CES is the breakdown voltage between the collector and the emitter when the emitter and the base are short-circuited.

● BV_CEX is the breakdown voltage between the collector and the emitter when the emitter junction is reverse biased.

As shown in Figure 35, the numerical relationship of the above breakdown voltages is as follows:

BV_CBO > BV_CEX > BV_CES > BV_CER > BV_CEO

It can be concluded that no matter how the GTR is connected, as long as its output voltage U_CE is lower than BV_CEO, it can be effectively prevented from breakdown.

2- Maximum Collector-emitter Voltage U_CEM

The maximum collector-emitter voltage U_CEM is the maximum output voltage of the GTR during normal operation, that is, the rated voltage of the GTR. In practice, for the sake of stable operation of the equipment and the safety of the operator, it is generally recommended that the U_CEM of the GTR be lower than the BV_CEO.

3- Secondary Breakdown Power P_SB

The power at which the secondary breakdown of the GTR occurs is the P_SB, which is the product of the secondary breakdown voltage U_SB and the secondary breakdown current I_SB, furthermore, neither U_SB nor I_SB are constants and will vary with I_B.

5.3.4 Basic Characteristics of GTR

5.3.4.1 Static Characteristics of GTR

The static characteristics of GTR are similar to those of BJT. However, unlike BJT, GTR only operates in the saturation and cut-off regions, so that it has no amplification state. This is because if the GTR is operated in the amplification state, its current and power consumption will be very high, which not only increases the unnecessary cost of use, but also makes the GTR burn out due to high junction temperature. Of course, the GTR must pass through the amplification region during the switching process between on and off state. Therefore, in order to avoid damage to the GTR, it is important to quickly switch the operating state of the GTR to minimize the time it spends in the amplification region.

5.3.4.2 Dynamic Characteristics of GTR

The dynamic characteristics of GTR are similar to those of BJT.

1- Turn-on Process of GTR

The GTR can be turned on by applying a forward base current I_B1. The turn-on process of GTR is divided into the delay time t_d and the rise time t_r.

The calculation formula of turn-on time: t_on = t_d + t_r

2- Turn-off Process of GTR

The GTR can be turned off by cutting the base current I_B. In addition, if a reverse current is applied, the turn-off process of the GTR can be accelerated. The turn-off process is divided into the storage time t_s and the fall time t_f.

The calculation formula of the turn-off time: t_off = t_s + t_f

* How to accelerate the Turn-on Process of GTR?

● Add an acceleration capacitor. Since the capacitor voltage does not change abruptly at the commutation instant, the switching characteristics of the GTR can be improved by connecting a capacitor in parallel to the base resistor. Of course, a similar effect can be also achieved if a fast-switching GTR tube with a small junction capacitance is selected.

● Increase the forward driving current. During the turn-on process, t_d and t_r can be reduced to shorten the turn-on time t_on if a forward driving current with large amplitude and steep front edge is applied. However, the driving current can not be too large, otherwise it will cause the diffusion time t_s of the GTR to increase due to supersaturation.

* How to accelerate the Turn-off Process of GTR?

● Reduce the saturation depth. By reducing the saturation depth of GTR during the turn-on process, the carriers stored at the base will be reduced, which helps to shorten the storage time t_s of the GTR.

● Applying reverse driving current I_B2. If an reverse overshoot current is applied to the base when turning off the GTR, the carriers in the base can be extracted more quickly, which helps to shorten the storage time t_s of the GTR.

● Apply reverse base voltage U_B2. If the reverse base voltage is increased when turning off the GTR, the dissipation rate of the stored charge will be accelerated, which also helps to shorten the storage time t_s of the GTR. However, it should be noted that the reverse base voltage should not be too large to avoid breakdown of the emitter junction.

3- Secondary Breakdown of GTR

Primary Breakdown: When U_CE exceeds BV_CEO, IC increases rapidly and U_CE will remain at the sustaining voltage BVsus, that is, the primary breakdown of GTR. In this case, the GTR will not be damaged as long as the current to the GTR is limited by the external circuit. That is to say, if the U_CE is reduced below the BV_CEO, the GTR will return to normal without affecting its characteristics. Thus, the primary breakdown is not permanent, but reversible.

Secondary Breakdown: If current limiting measures are not taken immediately after the primary breakdown, then once the I_C reaches the I_SB, the lattice defects present inside the GTR will lead to an increase in the local current density in the GTR, causing an increase in the local power consumption and leading to a localized overheating inside the GTR. This causes a sharp increase in the concentration of intrinsic carriers, which further increases the local current. After a series of positive feedback, although the surface temperature of the GTR is not high, the localized region inside it melts into a filamentary short-circuit region due to the high temperature in a very short period of time, which results in the formation of a low-resistance channel between the collector and the emitter, that is also known as the negative resistance phenomenon. In this case, the I_C increases sharply and the U_CE decreases sharply, that is, the secondary breakdown of the GTR. The secondary breakdown will cause irreversible and permanent damage to the GTR and can significantly degrade its performance or even render it unusable.

4- Safe Operating Area of GTR

If we want the GTR to operate safely and stably, then we need to make sure that it can neither be subjected to primary nor secondary breakdown. Therefore, the area enclosed by the I_CM, U_CEM, P_CM, and P_SB is the safe operating area of the GTR, as shown in Figure 39.

5.4 Power MOSFET

5.4.1 Introduction to MOSFET

The Field Effect Transistor (FET) is a unipolar device controlled by voltage. The field effect transistor works on the principle that a conductive channel is formed in it under the action of an electric field to control its conductivity, thus realizing the turning-on and off of the field effect transistor. Depending on the structure and material, the field effect transistor can be divided into the Junction Field Effect Transistor (JFET, also known as static inductive transistor, SIT) and the Insulated Gate Field Effect Transistor.

The Metal Oxide Semiconductor Field Effect Transistor (MOSFET) is a kind of insulated gate field effect transistor, which is mainly composed of metals, oxides and semiconductors, and it is a very important power electronic device in modern electronic technology. MOSFET has many advantages, such as high input impedance, low noise, low power consumption, large dynamic range, good temperature characteristics, no secondary breakdown, low driving power, fast switching speed, high operating frequency, good thermal stability (better than GTR), wide safe operating area, simple driving circuit, and easy to integrate. In addition, the gate bias of the MOSFET can be positive, negative as well as zero. Compared to other power electronic devices, MOSFET has a small current capacity and low withstand voltage, but by using a multi-unit integrated structure, we can obtain a power MOSFET with high-power and high-capacity, that is, the double-diffused MOSFET (DMOS). Overall, MOSFET is suitable for high-frequency power electronic devices below 1kW, such as signal amplifiers, impedance converters, variable resistors, constant current sources, and DC solid state relays.

5.4.2 How does the MOSFET work?

5.4.2.1 Basic Structure of MOSFET

The main component of a MOSFET is the body (B), which also known as the substrate or bulk, and its main material is silicon. Depending on the material of the substrate, MOSFET can be divided into N-MOSFET (NMOS) that uses a P-type semiconductor as a substrate (P-Sub) and P-MOSFET (PMOS) that uses a N-type semiconductor as a substrate (N-Sub). The source (S) and drain (D) of a MOSFET are two highly doped semiconductors embedded in the substrate, such as two N + regions embedded in NMOS and two P + regions embedded in PMOS. The source and drain of a MOSFET are completely symmetrical, which means that the source and drain are interchangeable. The carriers are free electrons for NMOS and holes for PMOS, and the conduction speed of free electrons is twice as fast as that of holes, so compared with PMOS, NMOS is fast, with low on-resistance, high driving capability, small threshold voltage, and high 1/f noise (a low-frequency noise whose noise power is inversely proportional to frequency). It should be noted that NMOS will turn on when a high voltage is applied to the gate and turn off when a low voltage is applied, so the source of NMOS must be connected to the lowest potential of the circuit. And PMOS will turn off when a high voltage is applied to the gate and turn on when a low voltage is applied, so the source of PMOS must be connected to the highest potential of the circuit. The gate (G) of a MOSFET is an insulating layer on the substrate，and its commonly used material is alumina (aluminum gate), but it is gradually being replaced by polysilicon (polysilicon gate). The internal resistance of the gate is very high (up to several hundred megohms), so the gate does not conduct with the source, drain, or substrate.

In practice, to avoid transient reverse currents in the circuit to breakdown the MOSFET, we can connect a power diode in parallel between the drain and source of the MOSFET. And for ease of use, this power diode is integrated directly into the MOSFET.

The branch types and derivative devices of MOSFET are very diverse. For example, when the gate voltage is zero, if there is a conductive channel between the source and drain of a MOSFET, the MOSFET is a Depletion mode MOSFET, and vice versa, an enhancement mode MOSFET. If the source and drain of a MOSFET are in the same plane, then the MOSFET is a lateral MOSFET (LMOS), and vice versa, a vertical MOSFET (VMOS). Although MOSFET has four terminals, the source and base are usually grounded, so common MOSFET is three-terminal.

1- Lateral Double-diffused MOSFET (LDMOSFET)

Taking the N-channel LDMOSFET as an example, shown in Figure 41, whose substrate is a low-doped P-Sub region and connected to the body (B) through a high-doped P+ region. The source and drain regions of the N-channel LDMOSFET are two high-doped N+ regions connected to the source (S) and drain (D), which are identical and theoretically interchangeable. The gate (G) of the LDMOSFET is on top of the insulating layer and does not have any contact with the substrate, forming a equivalent capacitor. Free electrons in the LDMOSFET need to be conducted with the help of a conductive channel located between the source and drain regions. When the gate voltage is 0, there is no channel between the source and drain of the enhancement mode N-channel LDMOSFET, as shown in Figure 41(a), whereas a channel exists between the source and drain of a Depletion mode N-channel LDMOSFET, as shown in Figure 41(b).

* Equivalent Capacitor of MOSFET

The equivalent capacitor of MOSFET takes the gate and substrate as the plate of the capacitor, and the insulating layer as the dielectric, the thickness of the insulating layer T_ox is the distance between the plates, as shown in Figure 41(c). Then we can get the calculation formula of its capacitance is C = ε * A / T_ox, where ε is the dielectric constant and A is the plate area. Therefore, we can see that the greater T_ox, the smaller C. Besides, according to the electric charge quantity calculation formula, Q = C * V, if V is determined, the smaller the C, the smaller the Q, the smaller the electron concentration in the conductive channel, the smaller the current (the quantity of electric charge passing through any cross section per unit time), that is, the larger the channel resistance, or the larger the on-resistance of the MOSFET.

* Conductive Channel

The distance between the source and drain regions is called the channel length (L), as shown in Figure 41(a), which is calculated as L_drawn = L_eff + 2 * L_D, where L_drawn is the total channel length, L_eff is the effective channel length, and L_D is the lateral diffusion length. By reducing the channel length, the channel resistance can be reduced, which not only reduces the power consumption of the MOSFET, but also increases the number of MOSFETs per unit area. However, due to the limitations of the manufacturing process, the channel length is currently only available down to the nanometer scale.

The distance perpendicular to the horizontal direction of the conductive channel is called the channel width (W), as shown in Figure 41(d). whose increase will increases the plate area of the MOSFET's equivalent capacitance, increases the equivalent capacitance, and thus decreases the channel resistance. However, as the equivalent capacitance increases, the charging time of the equivalent capacitor also increases, which decreases the switching speed of the MOSFET, so that when a certain limit value is reached, the increase in channel width cannot continue to reduce the channel resistance.

The distance from the bottom of the insulating layer to the bottom of the conductive channel is called the channel thickness (T), as shown in Figure 41(b). By increasing the gate voltage, the channel thickness can be increased, which reduces the channel resistance, but on the one hand, the body effect of the MOSFET reduces the effective channel thickness T_eff, and on the other hand, the gate voltage can not be increased indefinitely, and if the gate voltage exceeds the maximum driving voltage, the MOSFET will be broken down.

* Body Effect

The P-Sub region and the N+ region form a PN junction, then a depletion layer will be formed at the PN junction, which decreases the effective thickness T_eff of the conductive channel, and results in the need for a higher driving voltage to turn on the MOSFET. This phenomenon of threshold voltage shift is the body effect of the MOSFET, which is noticeable when the source and drain regions are in the same plane. Usually, a highly doped P+ region is added to the P-Sub region as the body (B) and grounded together with the source (S), which can effectively reduce the depletion layer and reduce the negative effects caused by the body effect.

2- Complementary MOS (CMOS)

Since NMOS and PMOS are complementary in physical characteristics, that is, NMOS operates at high potentials (similar to logic 1) and PMOS operates at low potentials (similar to logic 0). So if NMOS and PMOS can be integrated together, complex logic circuits can be formed (e.g., NOT, AND, NAND, OR, NOR, XOR, XNOR). This design process of fabricating integrated circuits containing both NMOS and PMOS on silicon wafer templates is called the Complementary MOS (CMOS). Usually PMOS is embedded directly in NMOS, so the substrate of PMOS is also called N-Well. Because CMOS combines the features of NMOS and PMOS, it has the advantages of high speed, low noise, high impedance, and wide operating voltage. Besides, CMOS consumes energy only when switching process, thus it also has the advantages of low power consumption, and low heat generation.

3- VDMOSFET

Unlike the LDMOSFET, the source and drain of the vertical double-diffused MOSFET (VDMOSFET) are not in the same plane, as shown in Figure 43, thus there is no need for a separate lead to the Body (B). The horizontal area of VDMOSFET is smaller than that of LDMOSFET, but its vertical area is larger than that of LDMOSFET. This is because VDMOSFET has an extra N- epitaxial region with low doping concentration, and this N- epitaxial region is equivalent to the channel of JFET in terms of working principle, so VDMOSFET can be regarded as a combination of MOSFET and JFET. Although the N- epitaxial region increases the on-resistance of the VDMOSFET so that it can withstand higher voltages, the equivalent resistance of the N- epitaxial region decreases as the current flowing through the PN junction increases due to the conductance modulation effect. Even so, the voltage withstand capability of the VDMOSFET is always stronger than that of the LDMOSFET. And this PN junction can form a parasitic power diode, which can effectively protect the VDMOSFET from reverse voltage breakdown. If the N+ region of the VDMOSFET is changed to the P region, we get a vertically conducting IGBT with two carriers involved in the conduction process, resulting in low on-resistance and low switching speed.

In summary, VDMOSFET has the advantages of the BJT and the ordinary MOSFET, so it has the advantages of almost infinite static input impedance characteristics, very fast switching speed, constant-like transconductance, high dV/dt, small switching loss, small driving power, and good frequency characteristics. In addition, VDMOSFET also has a negative temperature coefficient, large safe operating area, and no secondary breakdown problem. Furthermore, VDMOSFET is the ideal power device for both switching and linear applications, and are therefore widely used in motor speed control, inverters, uninterruptible power supplies (UPS), switching power supplies, electronic switches, hi-fi, automotive appliances and electronic ballasts. According to the shape of the groove gate, VDMOSFET can be divided into V-groove VDMOSFET (VVDMOS, or VMOS) and U-groove VDMOSFET (UVDMOS, or UMOS). Even though UMOS has a better voltage withstand capability than VMOS, but due to the very high proportion of VMOS in the field of power electronics, so that the power MOSFET generally refers to VMOS.

5.4.2.3 Working Principle of MOSFET

Through the above introduction of the MOSFET structure, we can understand that the external electric field formed by the gate voltage can change the conductivity of the MOSFET. When there is no conductive channel inside the MOSFET, its conductivity is very small, corresponding to the off-state of the switch. When there is a conductive channel inside the MOSFET, its conductivity is very large, corresponding to the on-state of the switch. Then we can utilize this feature of the MOSFET to make it an electronic switch with low driving power. The MOSFET can be divided into enhancement mode MOSFET and Depletion mode MOSFET based on whether there is a conductive channel inside the MOSFET when no gate voltage applied. Since NMOS has good characteristics, and the internal working of LDMOSFET is easy to observe, we take N-channel LDMOSFET as an example to discuss the working principle of MOSFET.

1- Enhancement Mode MOSFET

The enhancement mode MOSFET is a normally closed (NC) device. When no gate voltage is applied (U_GS = 0), there is no conductive channel between the source and drain regions of the enhancement mode MOSFET. Only after a certain gate voltage is applied, a conductive channel is formed inside the enhancement mode MOSFET. According to the formation process of this conductive channel, we can divide the working process of enhancement mode MOSFET into four stages.

Accumulation Layer Stage: When U_GS < 0, an electric field perpendicular to the semiconductor surface will be generated at the gate, attracting holes in the P- region to the bottom of the insulating layer and repelling free electrons in the P- region, and thus forming an accumulation layer. This process is similar to the charging of a capacitor, that is, under the action of the electric field force, positive charges will accumulate to the positive electrode of the capacitor, and negative charges will accumulate to the negative electrode of the capacitor, and thus this equivalent capacitor of the gate can be considered as a parasitic capacitor C_GS of the MOSFET. Then in this case, even if a forward voltage is applied between the source and drain, the current cannot flow directly from the drain to the source (without considering the leakage current), so we can consider that the MOSFET is in the off-state.

Depletion Layer Stage: When 0 < U_GS < U_T, an electric field perpendicular to the semiconductor surface will be generated at the gate, attracting free electrons in the P- region to the bottom of the insulating layer and repelling holes in the P- region, and thus forming a depletion layer. The working process of the depletion layer stage and the accumulation layer stage is very similar, except that the flow direction of the carriers is different. Besides, the MOSFET remains off-state at the depletion layer stage because too few free electrons are accumulated at the bottom of the insulating layer to form a conductive channel.

Inversion Layer Stage: When U_GS ≥ U_T, the free electrons accumulated at the bottom of the insulating layer are sufficient to form an N-type narrow layer that connects the drain region to the source region, which is knwon as the N-type conductive channel, and is also known as the inversion layer because its type is opposite to that of the P- region. At this stage, the current inside the MOSFET is free to flow from the drain region to the source region, so the MOSFET can also be considered to be in the on-state. The electric field formed by the gate is enhanced with the increase of the gate voltage U_GS, which results in more and more free electrons being attracted to the bottom of the insulating layer, resulting in a wider conductive channel, a smaller on-resistance, and, of course, an increase in the drain current I_D of the MOSFET. When the MOSFET is at the inversion layer stage, there is a linear relationship between its drain current ID and gate voltage UGS, that is G_FS = I_D / (U_GS - U_T), which is the forward transconductance that we will discuss later. However, it is important to note that the cross-section of the conductive channel is not a rectangle shape, but an approximate right-angle trapezoidal shape or a triangle shape. This is mainly caused by the parasitic capacitance C_DS consisting of the P- region, source region and drain region, so that holes accumulate toward the drain and free electrons toward the source. Then under the combined effect of C_GS and C_DS, the closer to the drain region, the narrower the conductive channel, and conversely, the closer to the source region, the wider the conductive channel.

Pinch-off Region Stage: When U_GS > U_T, a conductive channel is formed in the MOSFET, making the MOSFET on. At this time, because the voltage between the drain and the source is no longer 0, the condition for maintaining the conductive channel becomes U_GS - U_DS > U_T. If the U_DS rises so that U_GS - U_DS = U_T, the conductive channel changes from a right-angled trapezoid to right triangle with its vertex touching the edge of the drain region. And this vertex is the equilibrium point between the gate and drain electric fields where the width of the conductive channel is 0, so we call it the pinch-off point with the value U_GD = U_GS - U_DS. If the UDS continues to rise, then the drain electric field will gradually replace the gate electric field, which results in a pinch-off point that will gradually move away from the drain region and closer to the source region, then the large depletion layer formed between the pinch-off point and the drain region is called the pinch-off region. Due to the presence of the pinch-off region, there is almost no contact between the conductive channel and the drain region, then a strong inversion channel cannot be formed in the pinch-off region, which is called channel pinch-off. Since the pinch-off region is composed of holes attracted by the drain electric field, the closer the drain region, the higher the concentration of holes, so that the free electrons from the source region can still pass through the pinch-off region into the drain region, so that the MOSFET is still on, which is also known as the pinch-off inertia of the MOSFET. However, as the U_DS continues to rise, the pinch-off region becomes larger and larger, and its hole concentration becomes lower and lower, which makes it more and more difficult for free electrons to pass through the pinch-off region, and therefore the MOSFET is in a saturation state, which means that even if the drain voltage continues to rise, the drain current remains at a constant value, that is, the saturated drain current ID (sat). When MOSFET is saturated, its saturated drain current ID(sat) is not affected by U_DS but is determined by U_GS, which is to say, ID(sat) is related to the square of UGS, which is known as the square law transfer characteristic of the MOSFET.

2- Depletion Mode MOSFET

The Depletion mode MOSFET is a normally open (NO) device. When no gate voltage is applied (U_GS = 0), a conductive channel exists between the source and drain regions of the Depletion mode MOSFET, that is the N channel. The conductive channel widens if a positive gate voltage is applied, and narrows if a reverse gate voltage is applied. When the reverse gate voltage reaches a certain value, the conductive channel disappears. Therefore, we can control the turn-on and turn-off of the Depletion mode MOSFET by applying forward and reverse gate voltages.

Conduction State: When U_GS ≥ 0, the electric field generated on the gate will attract more free electrons to the N channel and repel the holes in the N channel. As U_GS increases, the channel becomes wider, the channel resistance becomes smaller, and the saturated drain current increases.

Cut-off State: When U_GS < 0, the electric field generated on the gate will attract more holes to the N channel and repel the free electrons in the N channel, making the channel narrower and the channel resistance larger. When U_GS reaches the pinch-off voltage U_PO, the free electrons in the N channel are depleted, the conductive channel disappears, and the saturated drain current tends to zero.

5.4.3 Main Parameters of MOSFET

Most of the parameters of MOSFET are the same as those of BJT.

1- Forward Transconductance G_FS

Generally, the ratio of the amount of current change to the amount of current change is called the current amplification factor, and the ratio of the amount of current change to the amount of voltage change is called the transconductance, which is also known as the Guide Mutual or GM. Unlike the BJT, the drain current I_D is not affected by the gate current I_GS but is affected by the gate voltage U_GS due to the working principal of the MOSFET. When the MOSFET is on, the ratio of the amount of drain current change to the amount of gate voltage change is known as the forward transconductance G_FS, GFS = d(l_D)/d(U_GS). In some amplifiers and switching circuits, the forward transconductance is a very important parameter, the larger the forward transconductance, the greater the ability of the MOSFET to amplify the input signal, that is to say, the more sensitive the response of the drain current to changes in the gate voltage.

2- MOSFET Capacitance

There are three internal parasitic capacitors inside the MOSFET, namely the gate-source parasitic capacitor C_GS, the gate-drain parasitic capacitor C_GD (also known as Miller capacitor), and the drain-source parasitic capacitor C_DS. The capacitance of these parasitic capacitors affects the dynamic characteristics of the MOSFET. If the capacitance is small, the switching current and driving power will be small, and the switching speed will be fast, and vice versa. The capacitance parameters usually given by MOSFET manufacturers satisfy the following calculation formula:

Input capacitance of MOSFET: C_iss = C_GS + C_GD

Output capacitance of MOSFET: C_oss = C_DS + C_GD

Reverse transfer capacitance of MOSFET: C_rss = C_GD

3- Drain-source On-resistance R_DS(on)

The drain-source on-resistance R_DS(on) refers to the on-resistance between the drain and source when MOSFET is on. R_DS(on) is determined by parasitic resistors inside the MOSFET, R_DS(on) = R_CS + R_N+ + R_CH + R_A + R_JFET + R_D + R_SUB + R_CD. R_CS is the source contact resistance between the N+ source region and the source electrode. R_N+ is the N+ resistance between the N+ source region and the conductive channel. R_CH is the channel resistance of the conductive channel. R_A is the accumulation resistance of the accumulation layer. R_JFET is the JFET resistance of the equivalent JFET. R_D is the drift resistance of the drift region. R_SUB is the substrate resistance of the substrate region. R_CD is the drain contact resistance between the substrate region and the drain electrode. The parasitic resistor of the MOSFET is only an equivalent resistor, and its composition is still semiconductor, so its resistance value is affected by external conditions, such as temperature and voltage. The higher the junction temperature of the MOSFET, the greater the R_DS(on) and vice versa. The higher the U_GS, the smaller the R_DS(on) and vice versa.

4- Maximum Drain Current I_DM

The maximum drain current I_DM is the rated current of the MOSFET. The maximum drain current is the drain current that enables the MOSFET to reach its maximum junction temperature when the case temperature is at a certain value. The maximum drain current is not only related to the structure of the MOSFET, but also to the way the MOSFET is packaged and the ambient temperature.

5- Maximum Drain-Source Voltage BV_DSS

The maximum drain-source voltage BV_DSS is the rated voltage of the MOSFET. The maximum drain-source voltage is the maximum voltage at which the drain and source of a MOSFET do not experience avalanche breakdown at 25℃ ambient temperature. In practical applications, the maximum drain-source voltage is usually the voltage between the drain and the source measured when the drain current is 250 μA.

6- Maximum Gate-Source Voltage U_GSM

In order to improve the control ability of the gate signal, the insulating layer of the MOSFET is usually not very thick, so it can be broken down if a sufficient voltage is applied. The maximum gate-source voltage, also known as the maximum driving voltage, is the maximum input voltage that can cause permanent damage to the gate insulating layer of the MOSFET in a very short time, and it is generally recommended not to exceed ±20V.

7- Switching Frequency

The switching frequency of MOSFET is the highest in common power electronic devices, up to 100 kHz or even up to a few MHz. Although high switching frequency helps to increase the efficiency of the MOSFET, it also increases the dynamic power consumption of the MOSFET due to the charging and discharging of the MOSFET's parasitic capacitors that can lead to driving power consumption during dynamic operation (switching process). However, the MOSFET requires almost no driving current during static operation (on or off state), so its static power consumption will be very low.

5.4.4 Basic Characteristics of MOSFET



To better understand the basic characteristics of the MOSFET, we put it into a working circuit as shown in Figure 48. The MOSFET in this figure is an enhancement mode N-MOSFET (simply known as enhanced N-MOS), whose output power source is V_DD and driving signal source is U_P. U_GS is the voltage drop between the gate and the source, and U_DS is the voltage drop between the drain and the source. R_S is the internal resistance of the driving circuit, R_G is the internal resistance of the gate, R_L is the load resistance, and R_F is the detection resistance used to detect the drain current. I_D is the drain current.

5.4.4.1 Static Characteristics of MOSFET

1- Input Characteristics of MOSFET

Due to the working principle of the MOSFET, U_GS does not affect the current between gate and source, but it does affect the drain current I_D. When U_GS < U_T, I_D is a leakage current close to zero. When U_GS ≥ U_T, I_D is approximately linear with respect to U_GS, and its slope is the forward transconductance G_FS of the MOSFET.

2- Output Characteristics of MOSFET

The static output characteristic curve of MOSFET is similar to that of GTR. We can divide it into the cut-off region, the saturation region and the non-saturation region. Normally, MOSFET only works in the switching state, that is, quickly switch back and forth between the cut-off region and the non-saturation region, to prevent it from burning out due to excessive power consumption when working in the saturation region.

Cut-off Region: The cut-off region of the MOSFET is similar to that of the BJT. Even though the U_DS is very high, the drain current is a leakage current that tends to 0, and I_D ≈ 0.

Saturation Region: The saturation region of the MOSFET is similar to the active region of the BJT. I_D is not affected by U_DS, but increases with U_GS. Therefore the MOSFET is subjected to high voltage and current in the saturation region, which results in very high power consumption.

Non-saturation Region: The non-saturation region of the MOSFET is similar to the saturation region of BJT. I_D increases with the increase of U_DS. This variation of current with voltage is equivalent to a voltage-controlled resistor, and hence the non-saturation region is also known as ohmic region or variable resistor region. In this region, although the MOSFET is subjected to a large current, the voltage it is subjected to is very small, so the power consumption of the MOSFET in non-saturation state is small.

5.4.4.2 Dynamic Characteristics of MOSFET

The working principle of MOSFET is essentially the charging and discharging of its parasitic capacitors, so during the switching process of MOSFET, the charging and discharging process of parasitic capacitors must be considered.

1- Turn-on Process of MOSFET

In order to turn on the MOSFET, we need to apply a stable high voltage U_P1 to the gate of the MOSFET. Due to the existence of the driving ciruit resistor R_S, the gate parasitic resistor RG and the gate-source parasitic capacitor C_GS, the gate voltage U_GS of the MOSFET cannot form a pulse square wave like U_P1, but rises at a certain slope. When the charging voltage of C_GS reaches U_T, the drain current I_D begins to increase with the increase of U_GS, that is, the MOSFET enters the saturation region from the cut-off region. After the charging voltage of the C_GS reaches the maximum value, the U_GS remains at U_GS1 and the I_D remains at I_D1. At this time, due to U_GS - U_DS < U_T, the conductive channel in the MOSFET is still in the pinch-off state, but due to the existence of pinch-off inertia, the drain-source parasitic capacitor C_DS begins to discharge and the U_DS begins to decrease. When U_DS reaches the minimum value, due to U_GS - U_DS > U_T, the conductive channel is completely formed, the gate-drain parasitic capacitor C_GD starts to charge, and U_GS rises again. At this time, the MOSFET enters the non-saturation state, and the I_D no longer increases with the increase of U_GS. When the C_GD charging voltage reaches the maximum value, the U_GS remains at U_GS2, which indicates that the MOSFET turn-on process has completed. The time period from 10% U_GS2 to 10% I_D1 is called the turn-on delay time t_d(on). The time it takes for I_D to rise from 10% I_D1 to 90% I_D1 is called the rise time t_r. In general, the turn-on time t_on of a MOSFET is the sum of the turn-on delay time t_d(on) and the rise time t_r.

The calculation formula of the turn-on time is, t_on = t_d(on) + t_r.

2- Turn-off Process of MOSFET

After U_P1 is removed, C_GD starts to discharge through R_G, and U_GS decreases with a certain slope. After the C_GD is discharged, the U_GS remains at U_GS1. At this time, the conductive channel is pinch off, the C_DS starts to charge, and the U_DS starts to rise. After the charging voltage of the C_DS reaches the maximum value, the U_DS remains at U_DS1. At this time, the MOSFET enters the saturation region, the C_GS starts to discharge through the R_G, the U_GS decreases again, and the I_D decreases with the decrease of the U_GS. When the U_GS drops below U_T, the MOSFET enters the cut-off state and the I_D drops to the minimum value. Since MOSFET does not have a minority carrier storage effect, its turn-off process is very fast, approximately tens of nanoseconds. The time period from 90% U_GS2 to 90% I_D1 is called the turn-off delay time t_d(off). The time it takes for I_D to fall from 90% I_D1 to 10% I_D1 is called the fall time t_f. In general, the turn-off time t_off of a MOSFET is the sum of the turn-off delay time t_d(off) and the fall time t_f.

The calculation formula of the turn-off time is, t_off = t_d(off) + t_f.

* How to speed up the Switching Process of MOSFET?

● By reducing the resistance and inductance of the driving circuit, the charging and discharging time of the parasitic capacitor of the MOSFET can be effectively reduced to speed up the switching speed of the MSOFET and reduce its switching loss.

● By applying a short-time overcharge voltage to the gate, the charging of the MOSFET parasitic capacitor can be effectively accelerated, speeding up the turn-on process of the MOSFET.

5.4.5 Series and Parallel Connection of MOSFET

MOSFETs are not suitable for series connection. Due to the raw materials, design and process, the performance of all MOSFETs can not be completely consistent, so their operating frequency and voltage withstand capacity are different. And because MOSFETs are high-frequency devices, they generate a lot of heat during the switching process. Therefore, if they are connected in series, then it is easy to cause the MOSFET with the lowest voltage withstand capacity to be burned out.

On the contrary, MOSFETs are well suited for parallel connection. This is because their on-resistance has a positive temperature coefficient, that is, the higher the temperature, the higher the on-resistance. When the current in one of the MOSFETs is too large, its on-resistance becomes large, then the current starts to flow into the other MOSFETs with smaller on-resistance, which effectively reduces the dynamic uneven current of the MOSFETs. If a small inductor is connected as a current current sharing reactor in the source circuit, it can effectively reduce the dynamic uneven current of the MOSFETs. Of course, the parallel connection of MOSFETs can also increase their current capacity, for example, dozens of MOSFETs are connected in parallel in the output circuit of an inverter welder to increase its output current.

5.5 Insulated-Gate Bipolar Transistor

5.5.1 Introduction to IGBT

The insulated-gate bipolar transistor (IGBT, or IGT) is a composite Bi-MOS device with the high input impedance of Power MOSFET and the high current capacity of BJT. IGBT is mainly used in areas where the withstand voltage is 600V or more, the current is 10A or more, and the frequency is 1kHz or more, such as the converter, inverter, pulse width modulation system (PWM), uninterruptible power supply (UPS), switching mode power supply (SMPS), resonant converter, industrial motor, and new energy vehicle.

IGBT has lots of advantages, such as high input impedance, low noise, fast switching speed, simple driving circuit, low driving power, low on-state voltage drop, low switching loss, wide safe operating area, small size, high current density, high current capacity, high withstand voltage, strong resistance to pulse current impact, and no secondary breakdown. However, compared to the power MOSFET, IGBT has disadvantages such as slow switching speed and susceptibility to latch-up.

5.5.2 How does the IGBT work?

5.5.2.1 Basic Structure of IGBT

The structure of an IGBT is very similar to that of a VMOS, with its gate (G) corresponding to the gate (G) of the VMOS, its emitter (E) corresponding to the source (S) of the VMOS, and its collector (C) corresponding to the drain (D) of the VMOS. However, IGBT has a P+ injection layer in the substrate region, which makes the IGBT become the PNPN structure like the thyristor.

IGBT can be categorized into Punch-Through (PT) and Non-Punch-Through (NPT) types based on whether it contains an N+ buffer layer or not. The forward breakdown voltage of PT type IGBT is higher than its reverse breakdown voltage, so it is more suitable for the DC circuit. The forward breakdown voltage of NPT type IGBT is the same as its reverse breakdown voltage, so it is more suitable for the AC circuit. PT type IGBT has the advantages of low switching loss, low on-state loss, and large current capacity, but its temperature characteristic is not as good as that of NPT, so it is not suitable for parallel connection.

IGBT can be also divided into N-channel and P-channel types, which is similar to MOSFET. The switching speed of the P-channel IGBT is 2-3 times slower than that of the N-channel IGBT, so the safe operating area of the P-channel IGBT is smaller than that of the N-channel IGBT. In addition, the cost of the P-channel IGBT is higher than that of the N-channel IGBT, so the P-channel IGBT is rare in practical use. Therefore, let's take PT type N-channel IGBT as an example to introduce the IGBT.

By understanding the function of each PN junction, we can roughly think of a PT-type N-channel IGBT as an equivalent circuit consisting of a parasitic enhanced N-MOSFET, a parasitic JFET, a parasitic PNP transistor V1 (P+ N- P), a parasitic NPN transistor V₂ (N+ P N-), and an equivalent extension resistor R₂, as shown in Figure 54(a). In this circuit, V₁ is the main output channel, V₂ is formed with the formation of the MOSFT, the JFET is formed by the N- drift region, and R₂ is formed by the equivalent resistor of the P- base region of V₂. Of course we can simplify this equivalent circuit diagram in order to visualize the operation of the IGBT more intuitively. First, we can simplify the parasitic JFET to the equivalent modulation resistor R₁, which is mainly formed by the equivalent resistor in the N- drift region, as shown in Figure 54(b). Then, we can further consider the PNPN structure of the IGBT as a parasitic thyristor SCR, as shown in Figure 54 (c). Finally, we can think of the IGBT as a MOSFET with high current switching capability, so the P+ and N+ regions can be regarded as a power diode VD₁, as shown in Figure 54 (d).

5.5.2.2 Working Principle of IGBT

From the simplified equivalent circuit diagram, we can see that the IGBT works not only like a voltage-driven SCR, but also like a MOSFET with reverse bloacking ability. Therefore, the IGBT has the forward and reverse blocking states in addition to the conduction state and cut-off state.

Forward Blocking State: When a forward voltage is applied to the IGBT output, as well as the gate and emitter are short-circuited, the IGBT enters the forward blocking state. At this time, the PN junctions J₁ and J₃ are forward biased, and the PN junction J₂ is reverse bias. The reverse voltage makes the depletion layer on both sides of J₂ extend to the P base region and the N- drift region.

Reverse Blocking State: When a reverse voltage is applied to the IGBT output, the PN junction J₁ is reverse biased, and the reverse voltage makes the depletion layer of J₁ extend to the N+ buffer region, that is the IGBT enters the reverse blocking state. By increasing the width of the N+ buffer region, the reverse blocking capability of the IGBT will be improved, but it will also increase the forward voltage drop of the IGBT. However, the reverse withstand voltage of an IGBT is usually only a few tens of volts, so to prevent the IGBT from operating in a reverse blocking state, a fast recovery epitaxial diode (FRED) is usually connected in reverse parallel to the output of the IGBT. Of course, sometimes the IGBT and the FRED will be encapsulated together to form an reverse-conducting type IGBT module for convenience.

Conduction State: When a forward voltage is applied to the IGBT output, as well as a certain voltage is applied to the gate, the P base region will form an N sub-channel region, allowing electrons to transfer from the N+ buffer region to the N- drift region. This electrons flow will reduce the potential of the N base region and provide the base current I_B1 for V₁. If the voltage drop generated by this electrons flow is about 0.7V, then the PN junction J₁ will be forward biased, that is the IGBT enters the conduction state. Since the N- drift region of IGBT is very wide and its doping concentration is low, the conductivity of the N- drift region is very low. When the IGBT is operated at high currents, the conductivity increases as the carrier concentration in the N base region increases due to the conductance modulation effect, which reduces the saturation voltage at the IGBT output and the total on-state power consumption of the IGBT. If there are an electron current I_N and a hole current I_P, it means that the IGBT is fully turned on. It is important to note that when the IGBT is in the conduction state, if the collector current is not limited, the IGBT will undergo a latch-up effect.

Cut-off State: When the IGBT is in the conduction state, if the gate voltage of the IGBT is reduced below the threshold voltage, or a reverse bias voltage is applied to the gate of the IGBT, then the N sub-channel in the IGBT will disappear, and the base current in the IGBT will be cut off, so that I_N and I_P will disappear, and finally the IGBT will enter the cut-off state. However, due to the minority carrier effect, the IGBT output current does not immediately drop to zero, but instead generates a tail current like BJT, whose characteristics are related to U_CE, I_C, and T_C. It should be noted that the minority carrier effect increases the switching time and switching loss of the IGBT.

* Latch-up Effect

The latch-up effect, also known as the latching effect, can be divided into static latch-up effect and dynamic latch-up effect. The static latch-up effect is caused by the excessive collector current of the IGBT. From the circuit in Figure 56, we can see that the V₁ and V₂ can be considered as the equivalent SCR (P+ N- P N+) in Figure 54(C) if the forward bias voltage on R₂ can supply enough voltage to trigger V₂. Once the equivalent SCR is triggered, the positive feedback mechanism of the SCR takes away control of the IGBT gate, so the IGBT cannot be turned off. The dynamic latch-up effect is mainly due to the large displacement current caused by the large di/dt and dv/dt at the high speed switching of the IGBT, which generates a forward bias voltage on R2 that is sufficient to trigger V₂, resulting in the equivalent SCR triggering.

The latch-up effect is detrimental to the IGBT. On the one hand, the latch-up effect increases the collector current of the IGBT, which significantly increases the power consumption of the IGBT. On the other hand, the electronic protection circuit will not be able to turn off the IGBT in time if there is an overcurrent in the circuit, so the IGBT will burn out due to the overcurrent. Therefore, the following measures are usually taken to avoid the latch-up effect.

● By adjusting the internal structure of IGBT, the resistance of R₂ is lowered so that the voltage across it is much less than the gate trigger voltage of V₂ to prevent it from triggering.

● By optimizing the N+ buffer layer, the hFE (β) of V₁ is reduced, so as to reduce the base current of V₂ and prevent it from triggering.

5.5.3 Main Parameters of IGBT

Most of the parameters of IGBT are the same as those of MOSFET.

1- Latching Current I_L

The latching current I_L refers to the value of the collector current that will cause the latch-up effect of the IGBT. The latching current I_L is usually more than 5 times of the I_CM. In the past, I_L was one of the main reasons limiting the current capacity of the IGBT. However, with the development of technology, the internal structure of the IGBT can effectively prevent the static latching effect, which is the reason why the capacity of the IGBT is now getting larger and larger.

5.5.4 Basic Characteristics of IGBT

5.5.4.1 Static Characteristics of IGBT

1- Input Characteristics of IGBT

The static input characteristic curve of IGBT is similar to that of MOSFET.

2- Output Characteristics of IGBT

The static output characteristics of IGBT can be divided into forward blocking region, active region, saturation region, and reverse blocking region. Normally, the IGBT only works in the switching state, that is, it quickly switch back and forth between the forward blocking region and the saturation region, to prevent the IGBT from burning out due to excessive power consumption when operating in the active region.

Forward Blocking Region: This region is similar to the cut-off region of BJT. When U_GE < U_T, the parasitic MOSFET of the IGBT is turned off and a leakage current ICEO occurs between the collector and emitter.

Active Region: This region is similar to the active region of BJT. When U_GE ≥ U_T and U_CE > U_GE - U_T, the IGBT operates in the active region and generates a on-state voltage drop of 0.7V. In the active region, the electron current I_N flowing into the N base region of the IGBT is controlled by the gate voltage U_GE, which limits the base current I_B1 of V₁ and further limits the hole current I_P, so that the collector current I_C of the IGBT behaves in a saturation state similar to that of a MOSFET. The IGBT is subjected to a very high voltage and current in the active region, so the IGBT should pass through this region as soon as possible to avoid damage due to excessive power consumption.

Saturation Region: This region is similar to the saturation region of BJT. When U_GE ≥ U_T, and U_CE ≤ U_GE - U_T, the collector current I_C is no longer controlled by the gate voltage U_GE, but determined by the external circuit. Of course, this region is also similar to the non-saturation region of the MOSFET, so the saturation region of the IGBT is also known as the ohmic region or the variable resistor region.

Reverse Blocking Region: This region is similar to the reverse blocking region of power diode.

* Differences between saturation regions of IGBT and MOSFET

According to the working principle of the IGBT, the saturation voltage drop after its full turn-on is mainly dependent on the conductance modulation effect, so the saturation region of the IGBT is voltage saturation.

According to the working principle of the MOSFET, the saturation voltage drop after its full turn-on is mainly dependent on the drain current, so it behaves as a resistive characteristic, that is, the saturation region of the MOSFET is current saturation.

5.4.2 Dynamic Characteristics of IGBT

The dynamic characteristics of the IGBT are similar to those of a combination of the MOSFET and the BJT.

1- Turn-on Process of IGBT

The turn-on process of IGBT is similar to that of MOSFET. It should be noted that the falling process of U_CE is divided into two steps, namely, t_fv1 when the equivalent MOSFET works alone, and t_fv2 when the equivalent MOSFET and the equivalent BJT work together. The time period from 10% U_GS1 to 10% I_C1 is called the turn-on delay time t_d(on). The time it takes for I_C to rise from 10% I_C1 to 90% I_C1 is called the rise time tr. Generally, the turn-on time t_on of the IGBT is the sum of t_d(on) and t_r.

The calculation formula of the turn-on time is, t_on = t_d(on) + t_r.

2- Turn-off Process of IGBT

The turn-off process of IGBT is similar to that of MOSFET. It should be noted that the falling process of I_C is divided into two steps, namely, t_fi1, when the equivalent MOSFET works alone, and t_fi2, when the equivalent MOSFET and the equivalent BJT work together. The time period from 90% U_GS1 to 90% I_C1 is called the turn-off delay time t_d(off). The time it takes for I_C to fall from 90% I_C1 to 10% I_C1 is called the fall time t_f. The tail time t_t is the time it takes for the tail current to disappear, which we will discuss briefly below. Generally, the turn-off time t_off of the IGBT is the sum of t_d(off) and t_f.

The calculation formula of the turn-off time is, t_off = t_d(off) + t_f = t_d(off) + t_fi1 + t_fi2.

* Tail Current

When the IGBT is turned on, there will be the hole current I_P injected from the P+ emitter region to the N- drift region. On the one hand, these holes are recombined with the free electrons in the N- region, forming a space charge region, and on the other hand, they remain in the N- region as a minority carriers, forming a carrier storage region. When the IGBT is turned off, the minority carriers in the carrier storage region are gradually cleared to the external circuit. During this process, as UCE rises, the space charge region becomes larger, which strengthens the drift motion, so that the minority carriers that are too late to get to the external circuit will generate a composite current inside the IGBT, that is, the tail current.

Although the tail current will gradually disappear during the turn-off process of the IGBT, in the case of high-frequency operation, the tail current will not only prolong the turn-off time of the IGBT, but also increase the switching loss of the IGBT. Of course, the tail current can be reduced by some means, such as increasing the collector resistance to reduce the amount of holes injected from the collector, or increasing the turn-off gate resistance to reduce the minority carrier lifetime.

5.5.4.3 Safe Operating Area of IGBT

Forward bias safe operating area (FBSOA): Determined by I_CM, U_CEM and P_CM.

Reverse bias safe operating area (RBSOA): Determined by I_CM, U_CEM and dU_CE/dt.

5.5.5 Series and Parallel Connection of IGBT

The IGBT is not suitable for series connection for similar reasons as the MOSFET.

Whether the IGBT is suitable for parallel connection depends on the collector current I_C, mainly because the temperature characteristics of the IGBT on-resistance R_ON are influenced by the I_C. When the IC ≤ 1/3 I_CM, the R_ON exhibits a negative temperature coefficient, so the IGBT is not suitable for parallel connection. When the IC > 1/3 I_CM, the R_ON exhibits a positive temperature coefficient, so the IGBT is suitable for parallel connection, just like the MOSFET. In addition, when the load current is small, the uneven current has little effect on the IGBT. Therefore, on the whole, the IGBT is very suitable for parallel connection.

5.6 Other Fully-controlled devices

Through the working principle of IGBT, we find that different types of power devices can be integrated together to form a completely new device through the composite structure, which can not only play to their advantages, but also make up for their disadvantages. At present, the common new fully-controlled composite devices on the market are MCT, SIT, SITH and IGCT, with the development of technology, these new devices will gradually replace the old ones, and other new devices will continue to appear in the future.

1- MOS Controlled Thyristor

The design idea of MOS controlled thyristor (MCT) is similar to that of IGBT, which is to control the operating state of PNP thyristor through MOSFET. The structure of the MCT is similar to that of the GTO, that is, the MCT consists of tens of thousands of MCT units, with each unit consisting of a PNP thyristor and a MOSFET. MCT combines the advantages of the MOSFET and the thyristor, which has very high di/dt and dv/dt tolerances, high switching speed, low on-state voltage, low switching loss, high voltage capacity and high current capacity. However, the voltage and current capacities of MCT are far from expected, and its cost is higher than that of the IGBT, so its market share is not high.

2- Static Induction Transistor

The static induction transistor (SIT) is a JFET with majority carriers involved in the conduction process. On the basis of ordinary JFET, SIT adds a majority carrier barrier in the channel to prevent free electrons from flowing from the source to the drain. Therefore, this barrier height can be changed by varying U_GS and U_SD, thereby controlling the number of majority carriers from the source region to control the channel current, that is, controlling the internal potential distribution of the channel by static means. SIT has a higher operating frequency and power capacity than those of the power MOSFET, making it suitable for high-frequency and high-power applications such as radar communication equipments, electronic ballasts, pulsed power amplifiers and high-frequency induction heating systems. In addition, SIT has characteristics such as negative temperature coefficient and radiation resistance, which can meet the stringent requirements of aerospace and military equipment. However, SIT has a high on-resistance and high on-state loss, and as a normally open device, so it cannot be used as widely as the power MOSFET.

3- Static Induction Thyristor

The static induction thyristor (SITH), also known as a field controlled thyristor (FCT), is a bipolar device in which two carriers are involved in the conduction process. On the basis of the SIT structure, SITH adds a PN junction with the function of injecting minority carriers, so that a parasitic thyristor composed of two parasitic transistors is formed inside it. The working principle of SITH is similar to that of SIT. Therefore, SITH has the advantages of good conductance modulation effect, low on-state voltage drop and large current capacity, as well as its switching speed is higher than GTO, turn-off current gain is smaller than GTO, and the driving circuit is simpler than GTO. Of course, SITH also has the same disadvantages as SIT, for example, their driving power is high, and they are all normally open devices.

4- Integrated Gate-Commutated Thyristor

The integrated gate-commutated thyristor (IGCT, or GCT) is a power device that integrates a GTO chip with an anti-parallel diode and a gate driving circuit. Compared with GTO, although the power capacity of IGCT is almost the same as that of GTO, its switching speed is 10 times faster than that of GTO, and it does not need a complex driving circuit as that of GTO. In addition, IGCT has the advantages of high reliability, compact structure, low loss, low manufacturing cost and high yield rate. Therefore, even though the driving power of IGCT is still large, IGCT can replace GTO as the preferred solution for high-power low-frequency applications until MCT technology matures.

§6. How to choose Power Electronic Devices?

6.1 Semiconductor Material

Before selecting power electronic devices, it is necessary to understand their semiconductor materials. With the development of materials technology, semiconductor materials can be roughly divided into three generations. The first generation of semiconductor materials, represented by germanium (Ge) and silicon (Si), are the most common semiconductor materials and are mainly used for low-voltage, low-frequency, medium-power power electronic devices, as well as in photodetectors. The second generation of semiconductor materials, represented by gallium arsenide (GaAs), are mainly used in microwave, millimeter-wave devices and light-emitting devices. The third generation of semiconductor materials, represented by silicon carbide (SiC), gallium nitride (GaN), aluminum nitride (AiN), and zinc oxide (ZnO), are mainly used in high-temperature, high-frequency, and high-power radiation-resistant power electronic devices, as well as in semiconductor lasers.

6.2 Discrete Power Electronic Devices

The following is a brief summary of the content of chapters 1 to 5 of this article, which will help to quickly clarify the differences between various power electronic devices. We can first reduce the selection range by analyzing our own requirements, and then consider the advantages and disadvantages of each power electronic device to identify the most suitable one.

6.2.1 Types of Discrete Power Electronic Devices

By understanding the classification methods of power electronic devices, such as the degree of control, the driving circuit signal, the effective trigger signal and the carrier involved, it is helpful to find the required devices quickly. We need to clarify what kind of components are required for the circuit we are designing, and roughly understand what role it plays in the circuit. For example, if we need a voltage-driven fully-controlled device to drive a high-frequency circuit, we can consider a MOSFET or IGBT.

1- Degree of Control

Uncontrollable device: Power Diode

Half-controlled device: SCR

Fully-controlled device: GTO, GTR, MOSFET, IGBT

2- Driving Circuit Signal

Voltage-driven device: IGBT, MOSFET, SIT, SITH

Current-driven device: SCR, GTO, GTR

3- Effective Trigger Signal

Pulse triggering device: SCR, GTO

Level triggering device: GTR, MOSFET, IGBT

4- Carrier involved

Unipolar device: MOSFET, SIT

Bipolar device: Power Diode, SCR, GTO, GTR

Composite device: MCT, IGBT, SITH, IGCT

6.2.2 Advantages and Disadvantages of Discrete Power Electronic Devices

We need to realize that there is no such thing as a perfect power electronics device that can meet the needs of every application. If a power electronic device has a large current capacity and voltage capacity, then its operating frequency is usually not very high. Besides, even if there exists a power electronic device with a large current and voltage capacities and a high operating frequency, then it may be very expensive. Therefore, when selecting power electronic devices, it is necessary to consider their advantages and disadvantages, as well as the overall cost.

1- Uncontrollable device

Uncontrollable devices are mainly used in rectifier circuits in industrial and power systems. The power diode, as the simplest power electronic device in terms of structure and principle, is low cost and stable in operation. However, since the power diode is an uncontrollable device and cannot be turned off, an additional turn-off circuit is required, which not only complicates the circuit design but also increases the overall cost of the entire circuit. And the switching frequency of the power diode is very low, so it is difficult to be used in high-frequency rectifier circuits.

General Purpose Diode (GPD) has the advantages of high peak reverse voltage, low forward voltage drop and strong rectifier capability. But GPD has long reverse recovery time and low operating frequency, therefore, FRD and SBD are generally recommended for higher operating frequency requirements.

Fast recovery diode (FRD) has the advantages of short reverse recovery time, high operating frequency, low forward voltage drop and high peak reverse voltage, but its rectifying ability is weak.

Schottky barrier diode (SBD) has the advantages of very short reverse recovery time, high operating frequency, and low forward voltage drop, but its has the disadvantaged of low reverse peak voltage, high temperature sensitivity, and high leakage current. SBD is used for high frequency rectification of low voltage and high current circuits, and can also be used as the reverse protector and the IC protector.

2- Half-controlled device

Half-controlled devices are mainly used as electronic switches in industrial and power systems. Thyristor (also known as SCR) has high current and voltage withstand capabilities, but its operating frequency is lower than that of MOSFET, and its current and voltage capacities are lower than those of GTO.

Fast thyristor (FST) has the advantages of short switching time, high operating frequency, high dv/dt and di/dt tolerance, and low rated voltage and low rated current.

Bidirectional thyristor (also known as TRIAC) has the same reverse characteristics as the forward characteristics, so it can work in the AC circuit, but this also results in its lack of the reverse blocking capability.

Reverse conduction thyristor (RCT) has the advantages of low on-state voltage drop, short turn-off time, high rated junction temperature, and high voltage withstand capability, as well as, its operating frequency is significantly higher than FST. The RCT integrates a power diode internally, which helps to simplify the circuit design, but also increases its overall cost.

Light-controlled thyristor (LTT) is the half-controlled device with the largest power capacity at present. Light triggering can ensure good insulation between the control circuit and the main circuit, which makes the LTT highly resistant to electromagnetic interference and allows its current and voltage capacities to be designed to be very large, but also results in its low operating frequency.

3- Fully-controlled device

Fully-controlled devices are mainly used in computers, communications, consumer electronics, automotive electronics and other fields.

Gate turn-off thyristor (GTO) is the first choice for high power, high voltage and low switching frequency applications. GTO has the advantages of high voltage capacity, high current capacity, high power (megawatt class), small turn-off current gain and good thermal stability, as well as it has the conductance modulation effect. However, GTO has the disadvantages of low switching speed, large gate negative pulse turn-off current and high driving power, and the design of its driving circuit is very complicated.

Giant Transistor (GTR) is suitable for medium power applications. GTR has the advantages of large voltage capacity, large current capacity, good switching characteristics and low saturation voltage drop, but it also has the disadvantages of low switching speed, large driving power, complex driving circuit, risk of secondary breakdown, as well as being a current-driven device. Further, GTR is more expensive than IGBT, so it is gradually being phased out as IGBT power capacity increases.

Metal oxide semiconductor field effect transistor (MOSFET) is the first choice for medium and low power, low voltage and high switching frequency applications. MOSFET has the advantages of fast switching speed, high input impedance, good thermal stability, low driving power, simple driving circuit, high operating frequency, and no risk of secondary breakdown. However, MOSFET has a small current capacity and low withstand voltage, so it is generally only suitable for applications with a power of up to 10kW.

Insulated gate bipolar transistor (IGBT) is the first choice for medium power, high voltage, and low switching frequency applications. IGBT has the advantages of fast switching speed, low switching loss, high input impedance, low on-state voltage drop, low driving power, and no risk of secondary breakdown. Although the switching speed of IGBT is lower than that of MOSFET, and its voltage and current capacities are not as high as those of GTO, IGBT will gradually replace GTO as its capacities increase.

MOS-controlled thyristor (MCT) has the advantages of extremely fast switching speed, low on-state voltage drop, low switching losses, high voltage capacity, high current capacity, and the ability to withstand very high di/dt and dv/dt. However, MCT has smaller voltage and current capacities and higher costs than IGBT, so its application areas are very limited.

Static induction transistor (SIT) has a greater operating frequency and power capacity than MOSFET, but it also has a greater on-resistance and on-state losses than MOSFET.

Static induction thyristor (SITH) has the advantages of conductance modulation effect, low on-state voltage drop and large current capacity, and it has higher switching speed and smaller turn-off current gain than GTO. Both SIT and SITH are normally open devices, which are safety hazards during use and therefore require more attention when designing circuits.

Integrated gate commutated thyristor (IGCT) has the advantages of high voltage and current withstand capabilities, high reverse blocking capability, low turn-off losses, high resistance to inrush current and electrical stress, and no need for complex snuber circuits. In addition, IGBT has the same power capacity as GTO, and its switching speed is 10 times faster than that of GTO, so it is widely used in high-current and high-power applications, such as power transmission systems, motor drives and AC drives.

6.3 Integrated power electronic device

In the 21st century, the integrated circuit technology has grown tremendously, so the power integrated circuit (power IC) that the integrate power electronic device and its auxiliary circuits is extremely common in consumer electronics. Power IC has the advantages of small size, light weight, long service life, high reliability, good performance, low cost, few leads and solder joints, thus significantly simplifying circuit design. Of course, the power IC has a small surface area and poor heat dissipation due to its small size, which results in low operating power.

The power module solves this problem very well, mainly due to the fact that it is equipped with a metal base plate with a large surface area and high heat dissipation capability. In addition to high power and good heat dissipation, the power module also offers many other advantages, such as its outer packaging is an sturdy insulating resin, which has excellent corrosion resistance, impact resistance, and moisture resistance, as well as, the internal integrated circuit surface of the power module is covered with a protective layer of silicone gel, which can enhance its vibration resistance and anti-interference ability. Of course, similar to power IC, the power module reduce the number of leads in the circuit and simplify the circuit design, so the inductance of the entire circuit is greatly reduced. This helps to reduce the need for protection circuits and snuber circuits, so the circuit design can be further simplified, which leads to lower total power consumption and more reliable wiring. In conclusion, power modules reduce the manufacturing cost of high-power circuits and are therefore widely used in power systems and electronic products. Common power modules include solid-state thyristor modules, solid-state power diode rectifier modules, solid-state fully-controlled bridge rectifier modules and solid-state half-controlled bridge rectifier modules. You can click the product page to get more information about the power module.

§7. How to use Power Electronic Devices?

7.1 Introduction to Power Electronic System

Power electronic devices cannot operate independently, and they need to work in a power electronic system (PES), which consists of a main circuit, a control circuit, a driving circuit, a detection circuit and a protection circuit. Below, we will introduce them separately.

Main Circuit: It is used to realize the change or control of electric energy. Power electronic devices are the core components of the main circuit.

Control Circuit: It is used to provide control signals to the driving circuit.

Driving Circuit: It is used to convert the control signal of the control circuit into a gate signal for the main circuit.

Detection Circuit: It is used to detect the working status of the main circuit and feed it back to the control circuit.

Protection Circuit: It is used to protect the control circuit and the main circuit to ensure the reliable operation of the entire system.

Electrical Isolation: It is used to isolate the small current control circuit from the large current main circuit.

As the most common power electronic device, the internal structure of a solid state relay is a basic power electronic system (Details see the working principle of Zero-Crossing AC Solid-State Relays).

7.2 How to drive Power Electronic Devices?

The driving circuit is the interface between the main circuit and the control circuit, and is used to convert the control signal of the control circuit into a turn-on signal or a turn-off signal for the main circuit. For half-controlled devices, the driving circuit only needs to provide a turn-on signal. For fully-controlled devices, the driving circuit should provide both a turn-on signal and a turn-off signal. Therefore, there are many factors that need to be considered when designing a driving circuit so that the power electronics can operate in an ideal state. It is well known that a good driving circuit can effectively reduce the switching time and switching power consumption of power electronic devices, while ensuring the efficiency, safety and reliability of power electronic devices.

According to the type of driving signal, the driving circuit can be divided into the current-driven circuit and the voltage-driven circuit. The current-driven circuit can provide a current-driven signal and a threshold voltage for the current-driven device, and the voltage-driven circuit can provide a voltage-driven signal for the voltage-driven device. Generally speaking, voltage-driven devices are preferred because voltage-driven circuits are easier to design and manufacture than current-driven circuits. The driving circuit can also be divided into the discrete driving circuit and the integrated driving circuit. Currently, the discrete driving circuit is the mainstream, but it needed to be designed separately for different power electronic devices, and it also needed to consider its parameter matching, electromagnetic compatibility, and other issues. Therefore, in order to solve these issues, the device manufacturer usually develop integrated driving circuit specifically designed to keep the power electronic device in optimal operating condition.

Obviously, the driving circuit alone cannot make the main circuit work properly. This is because the power of the main circuit is usually very high, and the driving circuit will be subject to burnout if electrical isolation measures, such as optical and transformer isolation, are not added between the driving circuit and the main circuit.

1- Optical Isolation

Optical isolation is a electrical isolation technology that transmits driving signals through optical signals. The optocoupler (OPT) is the most common optical isolation device, as shown in Figure 63(a). The interior of the optocoupler contains a Light Emitting Diode (LED) and a phototransistor, as shown in Figure 63(b). When the optocoupler is in operation, the Light Emitting Diode emits infrared light, which is received by the phototransistor and converted into an output signal to control the main circuit. Due to the dynamic characteristics of the phototransistor, the output waveform of the optocoupler is similar to an isosceles trapezoid, as shown in Figure 63(c). The input and output characteristics of the optocoupler are similar to those of the BJT, with an input current I_D equal to the base current of BJT and an output current I_C equal to the collector current of BJT. But unlike BJT, ordinary optocouplers usually have a current gain of less than 1, that is I_D/I_C < 1. Of course, the current gain can be increased by a Darlington structure, such as a high-transmission-ratio optocoupler, but at the cost of a weakened voltage withstand capability, which will typically be within 2000V. If there is a demand for voltage withstand capability, we can consider using a light-triggered thyristor (LTT), whose characteristics are the same as those of the ordinary thyristor except for the trigger mode. In addition, since the light-triggered thyristor transmits the signal through the optical fiber, there is no need to add additional optical isolation methods, and since it is a half-controlled device, there is no need to consider how to turn it off.

2- Transformer Isolation

Transformer isolation is an electrical isolation technology that uses the pulse transformer (PTR) in Figure 64(a) to transmit driving signals. The core of the pulse transformer isolates the primary winding from the secondary winding, as shown in Figure 64(b) and the pulse transformer transmits the control signal from the input circuit to the output circuit by the magnetic saturation characteristics of the core, so the transformer isolation is also known as magnetic isolation. The high-frequency signal can obtain a complete pulse output waveform through the pulse transformer, while the low-frequency signal cannot, as shown in Figure 64 (c), so the pulse transformer can only be used for high-frequency signals. Considering the nonlinear distortion characteristics of the transformer, the frequency of the pulse transformer is usually very high, so its heat generation and power loss will inevitably be large. In order to ensure that a pulse transformer can output a high frequency pulse waveform, it must operate at the initial permeability of the core, so its shape is designed as a toroidal structure and its size is much larger than that of other transformers.

7.2.1 Driving Circuit for Half-controlled Device

1- Turn-on Requirements of Half-controlled Device

● If we want the thyristor to turn on reliably, then it is necessary to make sure that the control signal has a certain pulse width, that is, the pulse sustaining time is long enough to establish a stable internal positive feedback mechanism.

● In order to protect the operation of the thyristor from burr interference, it also need to ensure a certain pulse flat-top amplitude to provide sufficient driving current.

● Of course, it is necessary to make sure that the gate voltage, gate current and gate power are all within the rated triggering range to avoid permanent damage to the thyristor.

● It is also essential to provide some necessary protection measures such as electrical isolation, temperature control, anti-interference, etc.

2- Common Driving Circuit of Half-controlled Device

● The pulse transformer can be used as an electrical isolation device for high-power load, as shown in Figure 65(a). Sometimes a transistor amplifier (TRA) can be added to the input side of the pulse transformer to obtain gate pulse currents with greater amplitude, longer duration, and shorter current rise time.

● The optocoupler can be used as an electrical isolation device for low-power load, as shown in Figure 65(b). If the load circuit is an AC circuit, an optocoupler with a built-in phototriac can be used.

7.2.2 Driving Circuit for Current-Driven Fully-controlled Device

Driving circuits for fully-controlled power electronic devices will vary with their parameters. Typically, the design of the current-driven circuit will be much more complex to that of the voltage-driven one. In order to utilize the proper performance of the power electronic devices, it is generally recommended to directly use the integrated driving circuit provided by the package manufacturer.

7.2.2.1 Driving Circuit for GTO

1- Turn-on Requirements of GTO

● Same as those of thyristor.

2- Turn-off Requirements of GTO

● It is necessary to provide a much larger turn-off current than the turn-on current through the gate reverse bias circuit to accelerate the turn-off process of the GTO.

3- Common Driving Circuit of GTO

Normally, the driving circuit of GTO includes turn-on driving circuit, turn-off driving circuit and gate reverse bias circuit. According to the coupling method, the driving circuit of GTO can be divided into pulse transformer coupling type driving cirucit and direct coupling type driving cirucit. Each of these two drive circuits has its own characteristics, as shown in Figure 66.

● The pulse transformer coupling type driving circuit is shown in Figure 66(a), where the pulse transformer is used not only as a signal source for the GTO, but also as an electrical isolation measure for the entire driving circuit. PTR1 provides the turn-on signal, PTR2 provides the turn-off signal, and their secondary windings are connected to the input of the GTO. In addition, since the output signal of the pulse transformer is an AC signal and the leakage inductance of the pulse transformer will interfere with the normal operation of the GTO, an RD circuit consisting of a resistor and a diode should be connected in series with the output of the pulse transformer to rectify the AC signal and absorb the leakage inductance. In short, although the design of the pulse transformer coupling type driving circuit is simple, its anti-interference ability is not strong, and its overall cost is high due to the high price of the pulse transformer.

● The direct coupling type driving circuit is shown in Figure 66(b), where the control signal directly controls the turn-on and turn-off of the GTO through three MOSFETs. Transformer T is only used as the voltage source of the entire circuit, and outputs a 5V AC voltage and a 15V AC voltage according to the voltage dividing principle of the transformer. VD₁ and C₁ supply the +5V voltage to the output of V2, and VD₄ and C₄ supply the -15V voltage to the output of V3. And the voltage amplifier consisted of VD₁, VD₂, VD₃, C₁, C₂ and C₃ can amplify the +5V voltage into +15V voltage, and supply it to the output of V1. In addition, resistors, capacitors and diodes can form RCD circuits to reduce mutual interference and parasitic oscillations within the circuit. To turn on the GTO, we should first turn on V₁ for a short period of time to provide a steep front part of the positive pulse to shorten the turn-on time of the GTO, and then we should quickly turn on V₂ and turn off V₁ to provide a flat top part of the positive pulse, so that the GTO is maintained at a critical saturation state that can be easily turned off. To turn off the GTO, we should turn off V₂ and turn on V₃ to provide a strong negative pulse to inhibit the positive feedback process inside the GTO to make it exit the saturation state, and then turn off V₃ after turning off the GTO so that R₃ and R₄ can provide a negative gate bias voltage to maintain the GTO in the turn-off state. In short, although the direct coupling type driving circuit has the disadvantages like complex design, high power consumption, low efficiency, due to its overall cost and strong anti-interference ability, it is still widely used in many fields.

7.2.2.2 Driving Circuit for GTR

1- Turn-on Requirements of GTR

● Same as those of thyristor.

2- Turn-off Requirements of GTR

● It is necessary to apply a certain base current, which helps to reduce the turn-off time and turn-off losses.

● It is also necessary to apply a negative bias voltage between the base and emitter, which will ensure a more reliable turn-off process of the GTR.

● It is needed to add protection measures such as anti-saturation, desaturation detection, overcurrent, overvoltage, over-temperature and pulse width limiting, to avoid damage to the GTR.

3- Common Driving Circuit of GTR

Figure 67 shows a common GTR driving circuit, which can not only maintain the GTR in a quasi-saturation state and provide automatic protection, but also improve the switching characteristics of the GTR, such as shortening the switching time, reducing the driving power, and improving the driving efficiency. The signal isolation circuit, consisting of the OPT, can electrically isolate the control circuit from the driving circuit. The desaturation detection circuit, consisting of VD₆ and voltage comparator A₁, outputs an overload protection signal from A₁ if the GTR exits saturation state. The adaptive output driving circuit, consisting of V₃, V₄, V₅, VD₇, VD₈, VD₉, C₂ and TVS, can not only output forward bias and reverse bias to the base of the GTR, but also has other functions like anti-saturation.

When the OPT is turned on, the voltage at point D U_D is zero, the current flows out of the circuit through the OPT, so VD₁ is turned off and the supply voltage charges C₁ via R₈ and VD₃. According to the characteristics that the voltage on both sides of the capacitor cannot change suddenly, the base voltage of V₂ is equal to U_D, so it also becomes zero, resulting in V₂ turned off, that is, the voltage at point E U_E is not zero, so that V₃ and V₄ are turned on, and V₅ is turned off. V₃ and V₄ form a Darlington structure that provides an overcharge drive current to the base of the GTR, allowing it to rapidly enter the saturated conduction state, as well as charging C₂. However, when the load is light, this overcharge driving current will supersaturate the GTR, prolonging the desaturation time when it is turned off. Of course, the Baker clamp circuit (also known as the anti-saturation circuit) consisting of VD₇ and VD₈ can solve this problem well. When the GTR is in a supersaturation state, its collector potential is lower than the base potential, which will cause the VD₇ to be automatically turned on and the excess driving current will flow into the collector, maintaining the U_BC to be zero, which means that when the GTR is turned on, it will always be in a critical saturation state.

It's time to explain how the automatic protection circuit works. When the GTR is turned on, the U_CE decreases, causing VD₆ to turn on, which makes U_B = U_CE < U_A, so that U_C is high and turns on V₁. The turn-on of V₁ not only maintains the turn-off state of V₂ and the turn-on state of V₃ and V₄, but also provides a discharge circuit for C₁, causing the voltage across the C₁ to drop to zero to prepare for the next work. However, if the GTR exits the saturation state due to overload and other reasons, the U_CE will continue to rise, which results in the turn-off of VD₆, so U_B is no longer equal to 0, but U_B > U_A, and therefore U_C is low, causing V₁ to turn off, V₂ to turn on, and ultimately the GTR to be turned off.

When OPT is turned off, VD₁ is turned on, VD₂ is turned off, the voltage at point B U_B is V_CC, and the voltage at point A U_A is the voltage dividing voltage of R₃ and R₄, then V_B > V_A, which makes the voltage at point C U_C low, so V₁ is turned off, V₂ is turned on, that is, the voltage at point E U_E is zero, so V₃ and V₄ are turned off, and V₅ is turned on. After V₅ is turned on, C₂ begins to discharge, and its output path is left side of C₂ → output of V₅ → TVS → VD₉ → right side of C₂, so that the GTR is quickly turned off due to the reverse bias of the base. TVS is a power diode capable of suppressing unidirectional transient voltages. Its forward characteristics are the same as those of ordinary voltage regulator diodes, but its reverse characteristics are similar to those of typical PN junction avalanche devices. Therefore, when the reverse biased transient overvoltage pulse generated during the C₂ discharge is applied to both ends of the TVS, the impedance of the TVS changes from high impedance to low impedance in a few picoseconds to absorb surge power up to several kilowatts and clamp the voltage across it to a preset value, which greatly accelerates the shutdown process of the GTR.

7.2.3 Driving Circuit for Voltage-Driven Fully-controlled Device

1- Turn-on Requirements of Voltage-Driven Fully-controlled Device

● It is generally believed that the conduction of voltage-driven devices does not require a driving current, but requires a stable and reliable driving voltage, such as 10-15V for MOSFET in general and 15-20V for IGBT in general.

● The output resistance of the driving circuit needs to be small enough to quickly establish the driving voltage.

● The construction of an IGBT is equivalent to a MOSFET and an SCR, so the working principle of the MOSFET and the IGBT is essentially the charging process for the parasitic capacitor of the MOSFET. Therefore, the driving circuit need to provide a large enough charging current to the gate-source capacitor of the MOSFET to turn it on quickly and reduce parasitic oscillations.

● The design of the driving circuit should be simple enough to ensure that its output resistance is small enough, which not only speeds up the turn-on process of the MOSFET, but also reduces the turn-on loss of the MOSFET.

2- Turn-off Requirements of Voltage-Driven Fully-controlled Device

● The driving circuit needs to provide a low impedance circuit when it is turned off to quickly discharge the gate-source capacitor, thereby shortening the turn-off time of the MOSFET and reducing the turn-off loss of the MOSFET.

● If possible, adding a reverse bias voltage during turn-off can speed up the turn-off process of the MOSFET and improve its turn-off reliability.

3- Common Driving Circuit of Voltage-Driven Fully-controlled Device

Figure 68 shows a common MOSFET driving circuit diagram, and we can clearly feel that its construction is much simpler than that of the GTO and GTR. This is mainly due to the fact that it is much easier to get a stable driving voltage than a stable driving current, as well as the fact that the turn-off process of a voltage-driven device is much simpler and faster than that of a current-driven device. Through the MOSFET structure in Figure 68, we can find that there are parasitic capacitors between the various terminals of the MOSFET, so the turn-on and turn-off process of the MOSFET is actually the charging and discharging process of these parasitic capacitors. V₁ and V₂ constitute a push-pull circuit, which functions to enhance the ability of the driving circuit to provide current, and quickly complete the charging process of the MOSFET gate-source capacitor, so that the MOSFET can be turned on quickly, which not only reduces the high-frequency oscillation of the rising edge, but also helps to reduce the switching loss of the MOSFET.

* Parasitic Oscillation

Parasitic oscillations are oscillations generated by internal parasitic parameters of power electronic devices that are independent of or outside the operating frequency range. Parasitic oscillations can be divided into low-frequency parasitic oscillations below the operating frequency and high-frequency parasitic oscillations above the operating frequency. Parasitic oscillations are independent of the operating frequency of the power electronic device, that is, even if the input of the power electronic device is short-circuited, there will still be an oscillation signal at its output. Of course, parasitic oscillations can be reduced or even eliminated completely if the driving circuit is properly optimized. In circuit design or practical applications, we can detect parasitic oscillations by the following characteristics:

● Most parasitic oscillations are high-frequency oscillations with large amplitudes, except for low-frequency oscillations caused by poor power supply decoupling.

● The frequency and amplitude of parasitic oscillations can vary with the internal parameters of power electronic devices, sometimes even causing abnormal turning on and off of power electronic devices.

● The period and waveform of parasitic oscillations are generally regular.

7.3 How to protect Power Electronic Devices?

Through the learning of the PN junction, we know that semiconductor devices are very afraid of overcurrent and overvoltage, which will cause their junction temperature to rise sharply and damage their internal structure. Therefore, it is necessary to equip power electronic devices with suitable protection circuits to eliminate the risk of overcurrent and overvoltage and provide a safe and reliable operating environment. Of course, power electronic devices are theoretically very difficult to damage as long as they are equipped with a sufficient number of protection circuits and a strong enough heat dissipation method. However, in practical applications, we also have to consider the design cost and power consumption of the whole circuit. Therefore, the protection circuit needs to be appropriately streamlined under the premise of ensuring safety margins.

7.3.1 Overcurrent Protection

1- Common Overcurrent Sources

The overcurrent source of power electronic devices is usually external circuit equipment failure or operation error, such as overload, short circuit, ground fault, and phase fault. The overload current is more than 20% greater than the rated current, and the short-circuit current is several or even tens of times greater than the rated current. If the equipment is not effectively grounded, there will be a abnormal zero sequence current. If there is a short circuit or poor contact between phases, there will be a abnormal phase current. All of these abnormal overcurrents can burn out power electronics, so they need to be avoided if possible.

2- Common Overcurrent Protection Measures

Figure 69 shows the common overcurrent protection measures in practical applications, and we can choose one or more of them according to our actual needs.

The current transformer CT is a specialized transformer for measurement purposes, which converts the large current of the main circuit into a small current that can be detect. The current transformer consists of a closed iron core column and two coils insulated from each other. The primary side of the current transformer is connected to the main circuit, and the secondary one is connected to the detection circuit, and the two are electrically isolated. However, when using a current transformer, it is necessary to avoid an open circuit on the secondary side of the current transformer. This is because once the secondary side is open, the current transformer loses the demagnetizing effect of the secondary winding, resulting in an excitation current being formed in the primary winding, which will increase the magnetic flux in the core, oversaturate the iron, and ultimately lead to heating and damage to the current transformer. In addition, if the number of secondary winding turns is high, a high voltage will be generated, endangering the safety of operators and equipment.

The overcurrent electronic protection circuit consists of an detection circuit, an actuator circuit and a trigger circuit. The detection circuit receives the current from the secondary side of the current transformer and compares it with its preset value. If the detection circuit believes that there is too much current in the main circuit, it will notify the actuating circuit to shut down the main circuit. After receiving the command of the actuator circuit, the trigger circuit will switch the power electronic device from the on state to the off state in an instant. Due to the PN junction, the power electronic device in the off state is a high-impedance device that can reduce the current of the main circuit to a very small leakage current, thus protecting other devices in the circuit very well. Of course, since the half-controlled devices cannot be turned off by the driving signal, the overcurrent electronic protection circuit is only suitable for fully-controlled devices, such GTO, GTR, MOSFET, and IGBT. The overcurrent electronic protection circuit is simple and inexpensive, requiring only the power electronic device in the circuit to be used as the actuator without the need to purchase a separate one, and is therefore usually integrated directly into the driving circuit of the power electronic device. However, due to the presence of the leakage current, the power electronic do not completely cut off the main circuit, and its ability to block the overcurrent depends on its ability to withstand the current in the off state.

The overcurrent protection relay KA has a built-in current transformer, detection circuit and actuation circuit, so its function is similar to that of an overcurrent electronic protection circuit. When the detection circuit inside the overcurrent protector detects that the current of the main circuit exceeds the preset value, it will control the AC circuit breaker QF1 to cut off the main circuit. Compared with the overcurrent electronic protection circuit, the overcurrent protection relay can completely cut off the main circuit, but once it has operated, a manual or automatic reset is required to restore it, which ensures that it can only be put back into operation after the fault has been removed. In addition, since the shutdown speed of the AC circuit breaker is not as fast as that of the power electronic device, both solutions are usually used together.

The DC fast circuit breaker QF2 has an internal electromagnetic circuit, which generates a huge repulsive force when there is a short-circuit current and causes the contacts to open in a very short time (2-3 milliseconds), to cut off the circuit. Compared to QF1, QF2 requires no additional control circuit and works much faster.

The fast fuse FU is a passive protection method because it cuts off the circuit by fusing itself rather than active operation. Fast fuses perform less quickly than high-frequency devices, so they are not suitable for MOSFETs, IGBTs, but very suitable for thyristors. On the one hand, the current withstand capacity of thyristors and fast fuses is related to their withstand time. On the other hand, the thyristor has a high current withstand capacity and is able to withstand a certain amount of overcurrent before the fast fuse blows. Therefore, the safety of the thyristor and subsequent circuits can be ensured as long as the fast fuse can be blown before the thyristor's withstand current reaches the maximum value. It is important to note that a fast fuse is a disposable device and once blown, it must be replaced with a fast fuse of the same size.

7.3.2 Overvoltage Protection

1- Common Overvoltage Sources

The overvoltage in power electronic devices is caused by overvoltage sources, which can commonly be divided into external overvoltage sources and internal overvoltage sources.

External overvoltage sources are mainly the lightning strike and the external circuit. The voltage amplitude of a lightning strike can be as high as 100 million volts, and there are virtually no power electronic devices that can withstand this voltage, so the usual solution is to direct it toward the earth through lightning rods. If the external circuits of power electronic devices have capacitive or inductive loads, such as transformers and capacitors, they will generate overvoltage during operation.

Internal overvoltage is generally caused by parasitic inductors and capacitors of power electronic devices. When the power electronic device is turned off, its parasitic inductors will generate the overvoltage at its output due to the rapid decrease of the forward current. This overvoltage can usually be reduced by connecting a reactor or filter in series with the output of the power electronic devices. When the power electronic device is turned off, its parasitic capacitors will discharge in an instant and generate the overvoltage. This overvoltage can usually be reduced by connecting a resistor in parallel with the output of the power electronic devices.

2- Common Overvoltage Protection Measures

Figure 70 shows the common overvoltage protection measures in practical applications, and we can choose one or more of them according to our actual needs. Of course, overvoltage is often accompanied by overcurrent, so it is sometimes necessary to use it with overcurrent protection devices.

The arresters F and the transformer electrostatic shielding layer D, as the most basic protection measures, can handle very high power overvoltage. Due to their ability to protect all electrical equipment throughout the building, they are placed farthest away from electrical equipment. The arrester has the same characteristics as the varistor because they are made of zinc oxide. Therefore, when a lightning strike occurs, the lightning overvoltage is immediately conducted to the ground without causing any damage to all components in the circuit. The transformer electrostatic shielding layer functions somewhat similarly to the arrester, and is able to conduct the high voltage static electricity generated by the transformer in the power grid when it is in operation to the earth.

The varistor RV and the suppression capacitor C are used to protect the individual electrical equipment and typically placed in switching power supplies. The varistor RV is a resistor with nonlinear volt-ampere characteristics, whose resistance decreases rapidly as the voltage increases. When an overvoltage occurs, the varistor becomes zero resistance for an instant and conducts this overvoltage to the earth. The suppression capacitor C is capable of absorbing overvoltages, but during its discharge, these absorbed voltages generate large discharge currents. Of course, this discharge current can be absorbed by connecting a resistor with a large resistance value to the suppression capacitor in series, although it will reduce the ability of the suppression capacitor to absorb the voltage.

The snubber circuit is capable of handling low-power overvoltages, so they are usually placed closest to power electronic devices. The RC1 circuit is the simplest snubber circuit, consisting of a capacitor and a resistor connected in series and connected in parallel to the output of the power electronic device. The RC2 circuit is based on the RC1 circuit, and a resistor is connected in parallel to the capacitor, so that the capacitor has two discharge circuits. The advantage of this design is that if the DC overvoltage generated at the output of the power electronic device, it can be quickly absorbed through these two resistors. In addition, during the discharge of the capacitor, the two discharge circuits can quickly absorb the discharge current, so that the capacitor can be quickly reset to the uncharged state to cope with the next overvoltage. The RCD circuit is based on the RC1 circuit, and a diode is connected in parallel to the resistor, which can effectively increase the charging speed of the capacitor.

* Snubber Circuit

The snubber circuit, also known as the absorption circuit or the buffer circuit, is the essential circuit in electronics. The snubber circuit can effectively absorb overvoltage and overcurrent, so that power electronic devices can work in a safe operating area, thus improving its stability and reliability. In addition, the snubber circuit can also reduce di/dt and dv/dt, improve EMI, and implement soft switching of power electronics to reduce their power loss.

The RLCD snubber circuit consists of a resistor R, an inductor L, a capacitor C, and a diode VD, as shown in Figure 71(a). The resistor can decrease the current value by converting electrical energy into heat. The inductor suppresses the current rise rate (di/dt) of the power electronic device according to its characteristic that the current on both sides does not change suddenly. The capacitor suppresses the voltage rise rate (dv/dt) of the power electronic device according to its characteristic that the voltage on both sides does not change suddenly. The diode restricts the direction of current according to the unidirectional conductivity of the PN junction. When the power electronic device is turned on, the inductor will suppress its di/dt, and when the power electronic device is turned off, the capacitor will suppress its di/dt. For the power electronic device with parasitic capacitors, the overvoltage when the device is turned off is absorbed by the capacitor, and when the device is turned on again, the energy stored on the capacitor is consumed by the resistor and waits for the next charge.

It is important to note that in order to reduce the power consumption of high-frequency power devices, the inductance is usually removed, so the RLCD snubber circuit can be simplified to an RCD snubber circuit, as shown in Figure 71(b). If there is a diode reverse recovery current in the circuit, it is likely to be exacerbated unless the RCD snubber circuit is further simplified to the RC snubber circuit in Figure 70.