Why are Power Transistors Necessary?
Power Transistors are capable of providing high currents and high blocking voltages and therefore, high power. They can be classified into BJT (Bipolar Junction Transistors), MOSFET (Metal Oxide Semiconductor Field Effect Transistors), and devices such as IGBT (Insulated Gate Bipolar Transistors) that combine bipolar and MOS technologies.
The principle of operation behind high power transistors is conceptually the same as bipolar or MOS transistors. The main difference is that the active area of the power devices is distinctly higher, resulting in a much higher current handling capacity. For this reason, they have large packages.
Three Major Device Technologies
The above figure depicts the typical structure of BJT, MOSFET and IGBT devices.
For a BJT to maintain conduction, a high continuous current through the base region is required. This imposes the necessity of high power drive circuits.
MOSFETs and IGBTs are voltage-controlled devices. The IGBT has one more junction than the MOSFET, which allows for a higher blocking voltage but limits the switching frequency. In IGBTs, during conduction, the holes from the collector p+ region are injected into the n- region. The accumulated charge reduces IGBT's on-resistance and thus the collector-to-emitter voltage drop is also reduced.
Maximum Power Dissipation
High currents and voltages in power devices produce very high internal power loss. This loss occurs in the form of heat that must be dissipated; otherwise, the device can be destroyed as a result of overheating.
The maximum power dissipation Pmax indicates a device's maximum capability to transfer and conduct this power loss without overheating.
The above figure compares the power dissipation of different devices. As a device type is selected, the circular area changes to show the device's maximum power dissipation. As illustrated in the figure, larger metal cases are more efficient in removing heat from the transistor.
Thermal resistance Rthj-a indicates the ability of the device to conduct and transfer heat from the junction (with temperature Tj) to ambient (with temperature Ta). More resistance means that less heat will dissipate.
The junction-to-ambient thermal resistance Rthj-a consists of two resistances: Rthj-c and Rthc-a in series. The Rthj-c is the junction-to-case thermal resistance. The Rthc-a is the case-to-ambient thermal resistance. The Rthc-a is many times higher than Rthj-c.
The above figure shows the values of Rthj-a for different packages. The particular value of each resistance depends on the type of materials and on the case size. Smaller values are typical for large metal cases.
Maximum Power Dissipation vs. Temperature
The maximum power dissipation Pmax of the transistor depends on the highest junction temperature that will not destroy the device Tjmax, the ambient temperature Ta, and the thermal resistance Rthj-a according to the equation shown in the above figure.
If the ambient temperature is less than or equal to 25°C the device reaches its maximum specified power rating. When the ambient temperature increases, the power rating decreases. If the ambient temperature Ta reaches the maximum junction temperature Tjmax, maximum power Pmax becomes zero.
One way to increase the power rating of the device is to diminish the thermal resistance Rthj-a. A heat sink, which is usually a metal construction with a large surface area, is used to allow heat to dissipate to ambient more easily.
When a heat sink is present, the global thermal resistance Rthj-a decreases because there are more paths available for heat dissipation. The case-to-heat-sink thermal resistance Rthc-s and the heat-sink-to-ambient thermal resistance Rths-a both facilitate heat dissipation. As a result, the power rating increases as illustrated in the above figure.
Safe Operating Area for BJT
The safe operating area (SOA) defines the current and voltage limitations of power devices.
The above figure shows the typical SOA of a power bipolar transistor. It can be partitioned into four regions. The maximum current limit (section a-b) and the maximum voltage limit (d-e) are determined by the technological features and construction of the particular device. The maximum power dissipation limits the product of the transistor's currents and voltages (section b-c). A secondary breakdown (c-d) occurs when high voltages and high currents appear simultaneously when the device is turned off. When this happens, a hot spot is formed and the device fails due to thermal runaway.
SOA for MOSFET & IGBT
The above figure shows the safe operating area (SOA) of MOSFET transistors. This area is bounded by three limits: current limit (section a-b), maximum power dissipation limit (b-c), and the voltage limits (c-d). The SOA of an IGBT is identical to that of the MOSFET SOA.
Since the drain current decreases when the temperature increases in MOSFET transistors, the possibility of secondary breakdown is almost nonexistent. If local heating occurs, the drain current - and consequently the power dissipation - both diminish. This avoids the creation of local hot spots that can cause thermal runaway.
SOA in Pulse Mode
The above figure demonstrates how the SOA of a device increases when the device is operating in pulse mode.
When the device is operating in DC mode the safe operating area is at its smallest. The SOA grows when pulse mode is used. The shorter the pulse signal, the higher the SOA.
Static Mode Operation
A power device is working in static mode when dc currents and voltages are applied to its terminals.
The above figure shows the schematic symbols, the terminal names, and the specific currents and voltages for BJT, MOSFET and IGBT.
Bipolar transistors are current controlled devices. In contrast, MOSFETs and IGBTs are voltage-driven devices. The performance characteristics of MOSFETs are generally superior to those of bipolar transistors: significantly faster switching times, higher input impedance, the absence of second breakdown, the ability to be paralleled, simpler drive circuitry.
Bipolar-based power devices include high-power bipolar transistors and Darlington transistors.
The current gain β of power bipolar transistors depends on the collector current IC. When the current increases, the current gain decreases as shown in the above figure.
To raise the current gain, Darlington transistors are used. Darlington transistors consist of two transistors connected in an emitter-follower configuration that share a common collector. The key advantage of the Darlington configuration is that the total current gain of the circuit equals the product of the current gain of both devices.
Linear Voltage Regulator
The above figure shows a linear voltage regulator. It produces a constant output voltage Uout that is smaller than unregulated input voltage Uin.
In this circuit, the power transistor operates as a common emitter amplifier and its collector-emitter voltage UCE compensates for the difference between input and output voltages. To this aim, the control circuit follows the output voltage and generates an appropriate base current that regulates the internal resistance of the transistor. Because the entire load current passes continuously through the transistor, its power dissipation has high values as is shown.
MOS Transistors as Switches
The behavior of enhancement-mode transistors depends on the applied gate-source voltage UGS. The n-MOS transistor is turned-off when UGS is smaller than the threshold voltage UT. There is no current and the device is replaced with an open switch. The inverse diode is inherent to the structure of power MOSFET and ensures a reverse-voltage blocking capability of the device.
When UGS > UT, the transistor is now turned-on. The key to MOSFET operation is the creation of the inversion channel beneath the gate; this allows current to flow. The region between the drain and source can be thought of as a resistor RON, although it does not behave in the same linear manner as a conventional resistor.
RON of MOS Transistors
The above figure illustrates the relationship of resistance RON to gate-source voltage UGS and drain current ID.
Usually, the threshold voltage of n-MOS power transistors is about 4-5 V. The higher the applied gate-source voltage, the wider the channel, and therefore the value of resistance RON decreases.
When the drain current of a turned-on MOS transistor with fixed value of UGS increases, the transistor channel narrows and resistance RON increases.
On-Resistance vs. Temperature
The above figure shows the values of the normalized drain-source on-resistance RON versus drain current ID with parameter junction temperatures Tj. The reference values for RON(ref) are: Tj = 25 oC and ID =2A.
The on-resistance RON increases with increases in temperature, thus avoiding thermal runaway. If unwanted heating occurs, RON increases, causing ID to decrease. This will diminish the power dissipation and thus the device temperature.
MOS transistors can easily be connected in parallel. Any imbalance between transistors will lead to an increase in the resistance RON in the transistor with a higher current. This will cause the higher current to be partially diverted to the other parallel device, thus avoiding thermal runaway.
The above figure shows the equivalent circuit and static output I-V characteristics of an IGBT.
An IGBT is in a turn-off state if the gate-emitter voltage UGE is smaller than the threshold voltage UT. When UGE > UT the MOS transistor goes to turn-on. Then, the collector current from the IGBT flows through the series connected emitter-base junction of the PNP transistor, and subsequently through the drain-source channel of the MOS transistor. The output characteristics of IGBT devices are similar to those of MOSFET devices, but due to the series connected junction they are shifted approximately 0.7 V to the right. The output characteristics have two regions – the ohmic region (for small UCE values), and the saturation region (for high UCE values).
IGBT Collector to Emitter Voltage
In power circuits, the IGBT is used as a switch. When it is set to turn-off, the switch is open. When it is set to turn-on, the switch is closed. IGBTs combine the advantages of bipolar and MOS transistors. They have higher collector to emitter voltages than BJT devices.
The above figure shows the typical relationship between UCE(ON) and junction temperature. At lower levels of collector current, UCE(ON) decreases with increasing temperature. At higher collector current UCE(ON) increases with increasing temperature.
The transfer characteristic is the relationship between collector current and gate voltage, at specified collector voltage. It is recomended the value of UGE to be up to 15 V.
Switching Mode of Operation
The above figure illustrates linear and switching operating modes of transistors. In linear mode, the transistor operates around point C; thus half of the supply voltage and half of the maximum collector current IC = UCC/RC will be dissipated, even if there is no ac input voltage. In switching mode, if the base current is high, the transistor turns on (point A). In this state, IC reaches max, and the voltage drop UCE across the switch is very low. There is close to zero power dissipation (also referred to as conduction losses). If the base current is zero, the transistor turns off (point B) and its losses are negligible. When in switching mode, if the circuit is rapidly switched between operating at A and operating at B, the average point of operation will still be at C, but now with nearly zero losses.
The above figure shows the ideal behavior of a transistor switch. When the input voltage Uin is zero, the transistor is turned off. No output current flows Iout through the transistor. The output voltage Uout equals the supply voltage UCC. When a positive input signal is applied, the device turns on. The output current jumps to its maximal value. The output voltage falls to zero. In this case, the global losses are zero.
In reality, switching characteristics are not so clear cut. They depend on the load type, the device characteristics, the type of driving circuit, etc.
BJT Switching - Resistive Load
The above figure shows the real switching characteristics of a BJT with a resistive load. When the input signal Uin is zero, the transistor is cut off and the collector current IC is near zero. If the positive input signal Uin is applied, IB increases to IB1 and after delay time td the collector current reaches 10% of its on-state value. The current IC continues to rise and after rise time tr it reaches 90% of its on-state value. The sum of both times is called turn-on time ton.
When the input signal Uin falls to zero, IC continues to flow for storage time ts. After that, IC falls from 90% to 10% of its maximal value for fall time tf. The sum of both times is called turn-off time toff. Usually the turn-off time toff is longer than the turn-on time ton.
Turn-On Switching with Inductive Load
If the input signal Uin is zero, the transistor is cut off, the collector current IC is zero and the collector-emitter voltage UCE equals the supply voltage UCC. Due to inductive load, the current IL is unchangeable and flows through the diode.
When the input voltage Uin is applied, the transistor starts conducting. The collector current increases slowly to its on-state level ICsat within time interval t1 - t2. During this period the free-wheeling diode is still on. As a result, the collector-emitter voltage UCE stays unchanged. After t2, the transistor goes to its saturation point and thus UCE decreases to UCEsat.
Turn-Off Switching with Inductive Load
If the input signal Uin is positive, the transistor is on, and the collector current IC is in its on-state level. Collector-emitter voltage UCE equals the saturation value UCEsat.
When the input voltage Uin falls to zero (at t1), the collector current remains constant within time interval t1 - t2. During the time interval t2 - t3, the collector current IC decreases to zero. At the time t2 UCE jumps to the supply voltage UCC.
The lag between the decrease in IC and the increase in UCE can provoke the appearance of very high values for the power dissipation within time interval t2 - t3.
Use the appropriate blocks to construct a switched-mode transistor circuit for dc motor speed control.
MOSFET with Resistive Load
The above figure illustrates the switching behavior of a MOS transistor with resistive load.
When the input signal rises up to a positive level, the transistor turns on and its drain current ID starts increasing. After the rise time ton, the current ID has reached its on-state level. This level depends on supply voltage UDD and resistor R. The UDS voltage is determined by the transistor's RON value.
When the input voltage is removed, the transistor turns off after fall time toff. The drain current drops to zero and UDS voltage increases to UDD.
MOS with Clamped Inductive Load
The above figure demonstrates the switching behavior of a MOS transistor with a clamped inductive load.When the input level is smaller than the threshold voltage, the transistor is turned off, drain current ID is zero, and UDS voltage is equal to the supply voltage UDD.
When the input jumps to positive, the transistor turns on and UDS voltage falls to near zero. The current ID flows through the inductor and its value increases. When the input falls to zero, the transistor turns off; at this point the magnetic field around the coil collapses, inducing a large voltage of opposite polarity across the coil. This raises the potential of a drain and so the diode turns on, increasing the UDS voltage to the battery voltage UCL.
Step-Down Switching Regulator
The above figure shows a step-down switching regulator that directly converts a fixed voltage dc supply to a variable output dc voltage.
The rectangular pulses on the base of the power transistor alternately saturate and cut off the pass transistor during each cycle. When the transistor is off, the voltage drop UCE is high, but the current IT is near zero. When the transistor is on, the current is high with negligible voltage drop. This way, the transistor power dissipation is always near to zero. LC circuit blocks the ac component of the rectangular voltage at the diode and forms dc at the output. Changing the ton/T ratio of the control pulses can regulate the output voltage value.
MOSFET with Capacitive Load
The above figure demonstrates the switching behavior of a MOS transistor with a capacitive load. In the initial state, the input level is near to zero, the transistor is turned off, drain current ID is zero, UDS voltage is equal to the supply voltage UDD, and the capacitor C is charged.
When the input jumps to positive, the transistor turns on and the capacitor starts discharging. The current ID through the transistor is at this point the sum of the current through resistor IR and the discharge current of the capacitor C. After time interval ton, the capacitor discharges; subsequently, the voltage UDS drops to near zero, and the current ID becomes determined only by the current running through resistor R. When the input changes to zero, the transistor turns off, the capacitor starts charging, and the switch resets to the initial state.
IGBT Switching Times
The above figure illustrates the switching waveforms of an IGBT. The switching times are:
Turn-on delay time td(on): the time it takes for the collector current to reach 10% of IC on-state value from the time a pulse is injected into the gate.
Rise time tr: the time it takes for the collector current to reach 90% of IC from 10%.
Turn-off delay time td(off): the time it takes for the collector current to reach 90% of IC on-state value from the time a pulse is removed from the gate.
Fall time tf: the time it takes for the collector current to reach from 90% of IC to 10%.
These times generally rise when the current and the temperature of the system increase.
Drive Circuit Requirements
Drive circuits ensure the safe switching of power transistors. The type of drive circuit necessary depends on various requirements that different devices impose on the input control. Drive circuits are divided into two basic groups - circuits for current control and circuits for voltage control.
Current control circuits are used to drive bipolar power transistors. They have to ensure a high-value base current in saturation mode (ON switch position) and reliable setting of the transistor in cut-off mode (OFF switch position). Voltage control circuits are used to drive MOS transistors and IGBTs. They have to guarantee input voltage values higher than the threshold when the transistor is turned on, and values lower than the threshold when the transistor is turned off.
Simple Drive Circuit for BJT
The above figure shows a typical drive circuit for a power BJT.
A positive voltage level is applied to the drive circuit for setting the transistor to its saturation point. The voltage is transformed via current IB = IB1 through the resistor R. This current should be much higher than base saturation current IBsat to ensure fast switching. The higher the base current in saturation, the shorter the turn-on switching time, but the larger the turn-off time.
When a negative input voltage is applied, the transistor cuts off after toff time. To minimize this time, the base current IB2 should be high.
Speed-up Drive Circuit for BJT
To reduce the turn-on time of a BJT without increasing the turn-off time, a capacitor C is added in parallel to the resistor.
When a high control voltage is applied, the base current I+max is much higher than IBsat.; which ensures fast transistor switching. The capacitor charges and the input current goes to a value a little bit bigger than IBsat.
At zero input voltage, the base appears inversely biased due to the charged capacitor. During capacitor discharging, the base current changes its direction thus ensuring a reliable and fast turn-off of the transistor.
Simple Drive Circuit for MOS & IGBT
The above figure shows a simple drive circuit for MOS and IGBTs implemented with a CMOS inverter. The gate resistor R is used to avoid undesired parasitic oscillations in the power transistor input.
If the drive voltage UDR increases, the input parasitic capacitor CGS charges through resistor R. The transistor turns on when the capacitor voltage reaches the threshold voltage UTH. After that the output current increases to its maximum value. If the drive voltage UDR falls, the input capacitor discharges. When the capacitor voltage decreases under threshold voltage UTH, the transistor turns off.
Speed-up Drive Circuit for MOS & IGBT
The above figure shows a fast drive circuit for a MOS and an IGBT.
For fast charging and discharging of the input capacitance, the value of R should be as small as possible. A small gate resistance will overload the inverter output. To prevent this, parallel connected inverters are used. This increases the maximum allowed output current, thus allowing resistance R to be decreased.
When the drive voltage UDR increases, the input capacitor rapidly charges through resistor R.
When the drive voltage UDR decreases, the input capacitor discharges through the diode. This accelerates the switching process.
Switching transients cause over-current, over-voltage, or rates of current (di/dt) and voltages (du/dt) in excess of critical rates to appear in power devices. Switching transients expand power losses, thus degrading device performances and sometimes even destroying the transistor.
The above figure illustrates the behavior of a typical inductive load switch. The power dissipation Pdis increases in both switching periods (t1 - t2) and (t3 - t4). The power losses increase when the switching times increase or switching frequency increases.
The trade-off between switching speed and power losses is a consideration in the normal operation of the circuits.
Snubber circuits are used to avoid the simultaneous occurrence of peak voltage and peak current, or to diminish the rates of change of currents and voltages.
The above figure shows the voltage-current switching locus of a typical power BJT switch. When the snubber is missing, the transition from on to off state and vice versa passes through the region with high voltage and current; this is called hard switching.
Snubber circuits ensure an alternate path for current and voltage, which significantly reduces power losses; this is called soft switching. Turn on snubber limits the rate of the rise of the current while the turn off snubber limits the rate of the rise of the voltage.
RCD Snubber Example
The above figure shows the turn-off switching of a power BJT with an inductive load. Without a snubber, peak voltage and peak current occur simultaneously and there is a very high amount of power dissipation.
With a RCD snubber, during turn off IC decreases, the capacitor charges and the collector-emitter voltage UCE increases. This avoids a simultaneous occurrence of high voltage and current in the output; it also avoids exceeding the critical rate of change of UCE voltage. The diode prevents a parasitic oscillation of the circuit.
The capacitor discharges through the snubber resistor when the device turns on.
Power Transistor Applications
The above figure shows some of the more widespread applications of power transistors in linear and switching modes.
In linear applications, continuous currents and voltages are applied between the terminals of power devices. In this operating mode, a large mount of power is dissipated.
When the power transistors operate in switching mode, the pulse currents and voltages are applied to its terminals. In this mode, a low average amount of power is dissipated. Power dissipation is high only during the small time interval when the transistor switches.
When the load of an amplifier has a high resistance (such as in a set of headphones) the necessary output power is very low (just a fraction of a watt). In this case, amplification can be accomplished with just a simple integrated amplifier.
If the load is larger (such as for a loudspeaker) more power is needed. This can be ensured with serially connected power bipolar transistors.
When high power sound systems must be supplied, parallel connected MOS power transistors are used to guarantee high amplification without thermal runaway problems.
DC Motor Control
The above figure shows a circuit for dc motor control using a half bridge chopper. It is composed from two power MOSFETs controlled by pulse signals S1 and S2. The same structure can be implemented with BJTs and IGBTs.
The motor mode of the dc machine is controlled by the upper transistor and the diode D2. The motor speed can be regulated by changing the turn on transistor switching time.
If the lower transistor and the diode D1 are operated, the dc machine works in regenerative braking mode. Then S2 switches at regular intervals and this way the stopping process is controlled.
Full Bridge Switching
The full bridge configuration of the single-phase voltage source inverter is shown.
Alternatively fired IGBT switches S1 and S4 determine the voltage UA (at point A). Switches S2 and S3 determine UB (at point B). The firings of the right leg (S2, S3) is shifted by an angle φ° with respect to the left leg (S1, S4) of the inverter. This produces resulting load voltage UAB with pulse width of φ°.
When an inductive load is connected, sinusoidal current appears through it. The amplitude of this current can be regulated by changing the angle φ°.
Optimum Operating Area
The figure shows the recommended operating ranges of power transistors depending on switched voltages and currents for different frequencies. Some additional factors as on-state losses, current density, reliability, prices, etc. must also be taken into account.
A BJT (including a Darlington connection) has the widest range of application utilizing switched power, but its frequency of operation is the lowest. It can be used in linear and switched mode voltage regulators, TV and monitors, motor control, etc. Typical MOSFET applications include switched mode power supplies, battery charging devices, power amplifiers. An IGBT is used in motor control, uninterruptible power supply (UPS), and low frequency lighting.