Gate driver
In electronics , especially power electronics , a gate driver ( MOSFET driver , IGBT driver or half-bridge driver ) is a discrete or integrated electronic circuit that controls power switches such as MOSFETs or IGBTs .
Furthermore, a simple gate driver can be seen as a combination of level shifter and amplifier .
motivation
It is often said that transistors with an insulated gate electrode such as B. MOSFETs can be driven without power or at least without current, but this is not correct. In principle, in contrast to, for example, bipolar transistors, due to their functional principle, such transistors only do not require a constantly flowing control current as long as their switching state is not to be changed. However, the insulated gate electrode in the transistor forming a capacitor (gate capacitor), which at each switching operation must be reloaded of the transistor. Since a transistor needs a certain voltage at the gate in order to switch through, this capacitor must be charged to at least this voltage. Conversely, when the transistor is switched off, this voltage must be reduced again, i.e. the capacitor must be discharged.
When a transistor is switched, it does not suddenly change from the non-conductive to the conductive state (or vice versa), but rather goes through a certain resistance range depending on the charging voltage of the gate capacitance . As a result, a more or less large amount of power is converted in the transistor during the switchover under current flow , which heats it and in the worst case can even destroy it. It therefore makes sense to make the switching process of the transistor as short as possible in order to keep the switching losses as low as possible.
For a capacitor with capacitance C, the following relationship applies between charging current (I), change in charging voltage (dU) and recharging time (dt):
Since the necessary voltage swing (change in the gate voltage between switching on and off, dU) and the gate capacitance C are specified by the transistor, the switching time dt of the transistor is the smaller, the greater the current I with which the Gate is controlled (reloaded). The level of this charge reversal is limited by the resistance and inductance of the gate current path. The source of the switching signals must be able to deliver these recharging currents. The gate capacitance itself depends, within certain limits, on the gate voltage applied. For a specific transistor, the manufacturer therefore usually gives the product of gate capacitance C and gate voltage U. This is the gate charge Q, which has to be transported into or out of the gate in the course of a switching process. Typical values of this gate charge for power MOSFETs are in the order of 100 nC (nanocoulombs). Because of the already mentioned thermal losses in the transistor in the course of the switching process, switching times in the order of magnitude of microseconds and less are aimed for, particularly in the case of periodic operation (rapid switching on and off). Correspondingly, the charge reversal currents that are required to control the gate under these conditions certainly reach values in the range of a few hundred milliamps up to the order of magnitude of amperes. With the typical gate voltages of around 10 to 15 volts, outputs of a few watts are easily required. If large currents have to be switched at high frequencies (e.g. in DC choppers for large electric motors), several transistors are often switched in parallel; the charging currents and switching capacities then multiply according to the number of transistors.
The switching signals for transistors are usually generated by logic circuits or microcontrollers , which make the signal available at standard logic outputs. Since these can usually only withstand currents in the double-digit milliampere range, directly connected power transistors are switched relatively slowly. The losses occurring during switchover are correspondingly high. At the same time, the gate capacitance of the transistor for the driving logic output forms an electrical short circuit at the moment of switching. Without protective measures, this can lead to an overload on the current side of the driver module, which can destroy it due to the heating due to ohmic losses in the current path. To counteract this, driver circuits (gate drivers) adapted to the intended use are used between the logic outputs and the power transistors.
Driver circuits
Control of individual transistors
In order to be able to switch over individual power transistors quickly, discrete electronic circuits or ready-made driver ICs are ideal. The following discrete driver circuits relate to driving n-channel transistors. Similarly, the driver circuits can also be used for p-channel transistors by changing the reference potentials .
Simple driver circuit
The simplest form of a driver circuit consists of a bipolar transistor T1 with a collector resistor R2 as a pull-up resistor . If no voltage is applied to the control input IN , the bipolar transistor T1 blocks and the gate of the power transistor Q1 is pulled through the resistor R2 to the operating voltage VT of the driver circuit. The gate capacitance is thus charged via this resistor R2 and the power transistor Q1 begins to conduct. If a voltage is now applied to the control input IN , then the bipolar transistor T1 short-circuits the gate of the power transistor Q1 , whereby the gate capacitance is discharged and the transistor Q1 begins to block.
The power transistor Q1 is thus switched on via a resistor R2 and switched off by short-circuiting the gate voltage. Since the discharge current through the bipolar transistor T1 is significantly higher in this case than the charging current through the resistor R2 , the transistor Q1 is switched off faster than it is switched on. Under certain circumstances, this behavior can even be desired, since switching on the power transistor too quickly results in high electromagnetic emissions .
Logic gate driver circuit
As already mentioned, a logic output only supplies low output currents. By connecting a plurality of logic gates U1B-U1F in parallel , their output currents can be added, so that in total an output current that is suitable for driving power transistors Q1 can flow. When connecting logic gates in parallel, it is important to have a steep signal edge at the input in order to achieve a nearly simultaneous overturning of all gates. In order to reliably generate such a signal, one of these gates U1A can be connected upstream in order to shape the input signal. The logic gate can be of type HC4069, for example, in order to achieve a drive voltage higher than 5V.
Push-pull driver circuit
In order to be able to deliver even higher output currents, the driver circuit can be designed as a push-pull output stage . A resistor R2 is inserted between the gate of the power transistor Q1 and the push-pull output stage so that the switching times are not too short and thus the electromagnetic emission is too long. By connecting a resistor-diode combination R3 + D1 in parallel , which reduces the total resistance for the switch-on process, it can be achieved that the power transistor Q1 switches on faster than it switches off. Fast switching on reduces switching losses, slower switching off reduces voltage peaks due to parasitic inductances .
Control of a half bridge
For certain applications it is necessary to switch a load not with just one power transistor, but with a half bridge . The simplest form of a half bridge consists of a combination of an n-channel transistor and a p-channel transistor. A driver circuit can now be used for each transistor in order to control the transistors in opposite directions.
Since p-channel transistors generally have poorer properties than n-channel transistors, only half-bridges with n-channel transistors are built in power electronics. For the control of the high-side transistor, however, problems arise from the fact that the potential of the control voltage at the gate does not change to the positive potential of the supply voltage, as is the case when using a p-channel transistor at this point, but to that at times variable potential of the midpoint of the half bridge relates. In particular, the transistor would not have to be fully controlled if the gate potential could only be raised to that of the supply. It requires its own driver circuit.
Driver circuit with bootstrapping
In order to be able to switch through the upper transistor Q1 in an n-channel half bridge, a voltage must be applied between the output of the half bridge (connection point of the two power transistors Q1, Q2 ) and the gate. This can be done with the help of a bootstrapping circuit .
If a voltage is applied to the control input IN , the lower power transistor Q2 is switched through (slowly). At the same time, the gate voltage of the upper power transistor Q1 is short-circuited via the bipolar transistor T1 and the lower power transistor Q2 . Ground potential is thus present at the output of the half-bridge , as a result of which the capacitor C1 is charged via the diode D1 . If the control input IN is now connected to ground, then not only the lower power transistor Q2 , but also the bipolar transistor T1 , which causes the gate capacitance of the upper power transistor Q1 to charge through the resistor R1 , initially from the supply voltage VP . When the output voltage rises, through an inductive load or because the upper power transistor Q1 begins to conduct, this voltage swing propagates through the capacitor C1 proceeds, the diode D1 blocks and the potential for the supply of the gate rises as desired via which the supply voltage VP at . Without the bootstrapping circuit , consisting of the diode D1 and the capacitor C1 , the output of the half bridge would assume a maximum of the voltage potential VP minus the minimum threshold voltage of the power transistor Q1 (VP - V GS ).
Since the supply voltage VP is short-circuited if both power transistors Q1, Q2 are conducting at the same time, it is important that one power transistor Q1 / Q2 is blocked before the other Q2 / Q1 is conducting. In this circuit, this is achieved by unequal switch-on and switch-off times of the power transistors Q1, Q2 .
It is not possible with this driver circuit to statically switch on the upper power transistor Q1 , since the bootstrap capacitor C1 loses its charge due to leakage currents . Before the upper power transistor Q1 comes out of saturation, the lower power transistor Q2 must be switched on again.
Driver circuit with isolated supply voltage
Another possibility of being able to switch through the upper power transistor of an n-channel half-bridge and even to switch it on statically is to supply the driver stage with an electrically isolated supply, for example by means of a switching converter or a charge pump . There are driver ICs for this, most of which have already integrated the necessary circuitry.
Other driver circuits
With switching converters , it may be necessary to control the power transistors in an electrically isolated manner in order to maintain the electrical isolation of the switching converter. The control current required to control the power transistor can be transmitted via transformers ( pulse transformers ) with suitable switching. It is therefore not necessary for the driver voltage to be generated specifically on the secondary side.
In general, there is a suitable integrated solution for every application. With half-bridge driver chips in particular, there is a clear advantage over a discrete solution. Some driver chips generate a dead time (locking time) so that the supply voltage is not short-circuited when the half-bridge is switched over. This ensures that both transistors never conduct.
practice
Especially with high currents, voltage drops and voltage peaks occur at the ground connections. These voltage differences lead to potential differences between the driver circuit and the power transistor, as a result of which the control voltage at the gate of the power transistor can be significantly higher than the supply voltage of the driver circuit. Excessive voltages at the gate of a power transistor can destroy them. It is therefore important to ensure good grounding in order to minimize these effects.
literature
- Ulrich Tietze, Christoph Schenk: Semiconductor circuit technology 12th edition, Springer, Berlin, Heidelberg, New York, 2002, ISBN 3-540-42849-6
- Ulrich Schlienz: Switching power supplies and their periphery 3rd edition, Vieweg & Sohn Verlag | GWV Fachverlage GmbH, Wiesbaden, 2007, ISBN 978-3-8348-0239-2
Web links
- Explanation of driver circuits
- Overview of commercially available driver modules
- Jörg Rehrmann, the new Internet power supply and converter manual (many application examples)