Semiconductor switches with fail-safe protection

Abstract

Semiconductor switches - a plurality of circuit modules connected in series, each of which is formed with - a first transistor (M3) with a source terminal (source3), a control terminal (Gate3) and a sink terminal, wherein - the source terminal (source3) is a source terminal or an emitter terminal, the control terminal (Gate3) is a gate terminal or a base terminal, and the sink terminal is a drain terminal or a collector terminal, and - a second transistor (Q2) with a source terminal, a control terminal and a sink terminal, wherein - the source terminal is connected to the control terminal (Gate3) of the first transistor (M3), - the sink terminal is connected to the source terminal (source3) of the first transistor (M3), and - with a driver (V2) which is connected via connections to the control terminals of the second transistors (Q1, Q2, Q3, Q4, Q5, Q6) of the individual circuit modules, wherein - the semiconductor switch is designed to interrupt the connection between the control terminal of the second transistor (Q2) of the malfunctioning circuit module and the driver (V2) in the event of a malfunction of one of the circuit modules.

Description

[0001] The invention relates to a semiconductor switch.

[0002] Advances in semiconductor technology have led to new switching concepts that can replace conventional, typically electromechanical, low-voltage switches. These new concepts apply, for example, to circuit breakers or motor starters, but can also be used, in principle, for switching higher currents, such as in power switches. The fast response times of the semiconductor components are of central importance, as they prevent damage from overload. This means the switch interrupts the current before the semiconductor components could be destroyed. Nevertheless, given the greater sensitivity of semiconductor components compared to conventional switching elements, it is still possible for the semiconductor-based switching mechanism to be damaged, and the switch to malfunction.It is desirable that in such situations there is a safeguard in place that protects the user against the consequences of such damage or keeps them within limits.

[0003] DE 691 22 541 T2 shows a static medium-voltage switch with several voltage-dependent power semiconductors, each comprising a gate terminal, a source terminal and a drain terminal, connected in series with a load impedance to the terminals of a supply voltage, and a control circuit which, together with the aforementioned semiconductors, forms several switching stages, each with one of the semiconductors.

[0004] The datasheet for the IRS2092 PROTECTED DIGITAL AUDIO AMPLIFIER, International Rectifier dated October 24, 2013 (https: / / www.infineon.com / cms / en / product / power / class-d-audioamplifier-ic / discrete-class-d-audio-amplifierics / irs2092spbf / ) describes a high-voltage, high-power Class D audio amplifier driver with PWM modulator and overcurrent protection.

[0005] The application note for the IRS2092 (S) Functional Description - Application Note AN-1138, International Rectifier dated September 28, 2007 (https: / / www.infineon.com / dgdl / an-1138.pdf?fileId=5546d462533600a40153559a077610d1) describes application examples for a Class D audio amplifier driver. The audio amplifier driver initiates a sequence of events when it detects an overcurrent condition during the turn-on of a pulse on either the high-side or low-side. This sequence involves turning off outputs. The audio amplifier driver can communicate with an external system controller via a simple interface circuit.

[0006] The invention aims to provide a semiconductor switch with fail-safe protection.

[0007] This problem is solved by a semiconductor switch according to claim 1. Advantageous embodiments are specified in the dependent claims.

[0008] A "semiconductor switch" is a device for switching current in which a current path is routed through at least one semiconductor component (typically a transistor) and can be interrupted by means of this semiconductor component. Such semiconductor switches often incorporate galvanic isolation, which is usually prescribed by a standard; however, the interruption of the current path is always also achieved by at least one semiconductor component. Examples of such semiconductor switches include circuit breakers, power switches, and motor starters in low-voltage applications.

[0009] The semiconductor switch according to the invention is formed with a plurality of circuit modules connected in series.

[0010] Each of these circuit modules comprises a first transistor with a source terminal, a control terminal, and a sink terminal, where the source terminal is either a source or an emitter terminal, the control terminal is either a gate or a base terminal, and the sink terminal is either a drain or a collector terminal. The circuit modules also include a second transistor, which has a source terminal, a control terminal, and a sink terminal, with the source terminal connected to the control terminal of the first transistor and the sink terminal connected to the source terminal of the first transistor. Furthermore, the semiconductor switch includes a driver or driver circuit (preferably only a single driver or driver circuit). This driver is connected via connections to the control terminals of the second transistors of the individual circuit modules.

[0011] The semiconductor switch according to the invention is designed to interrupt the connection between the control terminal of the second transistor of the malfunctioning circuit module and the driver in the event of a failure of one of the circuit modules.

[0012] Additionally, the semiconductor switch can be designed to switch itself off autonomously if one of the circuit modules fails. Specifically, this means that in one configuration, the semiconductor switch could both interrupt the connection between the control terminal of the second transistor of the malfunctioning circuit module and the driver, and also switch itself off to guarantee a higher level of safety.

[0013] In the case of a malfunction in a power electronic semiconductor, two fault conditions can typically be assumed. The semiconductor is either high-resistance and no longer controllable, or—and this is the more common case—it is low-resistance and no longer controllable: the switch is shorted out. In the second case, there is usually a short circuit between all three terminals of the semiconductor or transistor, meaning the gate has a low-resistance connection to the drain and source (in the case of FETs) or collector and emitter (in the case of IGBTs). In this case, the entire parallel circuit is short-circuited, and the semiconductor switch is no longer controllable in the on state (low resistance between drain and source). This condition is a serious fault, especially if the semiconductor switch is intended to be used as a protective element. The term "malfunction" refers specifically to this condition.The semiconductor switch can also be further developed to react to other malfunctions, such as a high-resistance state. While this is a rare occurrence, it also presents a problem in parallel circuits that is difficult to detect. The total resistance – denoted as RDSon – then increases, which can lead to thermal overload of the semiconductor switch.

[0014] According to one embodiment of the semiconductor switch according to the invention, it comprises a number n of circuit modules connected in series and is designed by the individual circuit modules for a maximum permissible value of a load current to be carried through the semiconductor switch such that the cumulative reverse voltage of the transistors of each subset of n-1 circuit modules of the n circuit modules of the semiconductor switch, through which the load current is carried, is greater than or equal to the maximum possible value of the load voltage (i.e., the peak voltage occurring for this maximum permissible load current). That is, the semiconductor switch is dimensioned such that the malfunction of a circuit module can be intercepted, or the reverse voltage is still sufficient. According to a further development, this is also fulfilled in the case of a malfunction of a plurality m > 1 (where m is a natural number) of circuit modules, e.g.In the event of a malfunction of two circuit modules, each subset of nm circuit modules, for example, still possesses a sufficient blocking voltage.

[0015] According to one embodiment of the semiconductor switch according to the invention, (preferably all) circuit modules are formed with a fuse located in the connection between the control terminal of the second transistor of the circuit module and the driver. In the event of a malfunction in a circuit module with a fuse, the semiconductor switch is designed to interrupt the connection between the control terminal of the second transistor of that circuit module and the driver by means of the fuse located there. If the semiconductor switch is appropriately oversized to allow for proper operation with fewer than the total number of circuit modules, it can continue to be operated. However, it is then advisable to send a fault message so that the operator is informed of the damage to the semiconductor switch and can plan for a replacement.

[0016] The interruption of the connection to a malfunctioning circuit module can be achieved, for example, by designing the semiconductor switch so that, if a current threshold is exceeded on the connection between the control terminal of the second transistor of the circuit module and the driver, the fuse located there trips. The term "current threshold" is intended to also include checking whether a value derived from the current has been exceeded, e.g., the voltage drop across a known resistance (e.g., a shunt resistor). The fuse can be designed or dimensioned to trip when the current threshold is exceeded. For example, it could be a thermal fuse or a PTC resistor (i.e., a thermistor with a positive temperature coefficient). However, an electronic fuse is preferred, which, for example,with a switching transistor such as a MOSFET. In the case of an electronic fuse, the corresponding connection between the control terminal of the second transistor (of the failed circuit module) and the driver is formed with a circuit for detecting when a current threshold is exceeded. The semiconductor switch can then include a control unit (e.g., MCU), and the circuit for detecting when the current threshold is exceeded is then designed to output a signal (e.g., via an optocoupler) to the control unit, indicating that the current threshold has been exceeded. This signal can be used to notify the user regarding...The circuit can be used to detect a malfunction (and can also output identification information for the malfunctioning circuit module), as well as to trigger the control unit to generate and transmit a control signal for the (electronic) fuse to trip when the current threshold is exceeded (e.g., via an optocoupler). However, it is also possible for the circuit for detecting the current threshold to itself be designed to generate and transmit a control signal for the fuse to trip when the current threshold is exceeded, meaning that tripping is not triggered by a control unit (with a view to potentially even more reliable tripping or when a dedicated control unit for the semiconductor switch is not required).

[0017] According to a further development of the semiconductor switch according to the invention, it is provided with an evaluation circuit (e.g., part of a control unit or MCU of the switch) which is designed to compare a voltage difference with a threshold value, wherein the threshold value is determined according to the blocking voltage of a first transistor of a circuit module, and voltage measurement circuits are arranged at the input and output of the semiconductor switch, which are designed to transmit measured voltage values ​​to the evaluation circuit. In addition, a circuit for switching off the semiconductor switch when the threshold value is exceeded is provided. This then switches off, for example, the driver voltage. This further development also allows errors or malfunctions to be detected and / or corrected.These are intercepted when a circuit module can no longer be switched on; the on-resistance when the semiconductor switch is turned on is then always greater than the blocking voltage of the circuit module.

[0018] In the embodiments of the semiconductor switch according to the invention, it is not necessary for all circuit modules to be of the same dimensions or even of the same construction (although at least the latter is desirable for the first transistors). Any differences can be taken into account, for example, by software of a control unit of the semiconductor switch, so that the failure protection intended by the invention is also provided in such cases.

[0019] There are various design options for the circuit modules. For example, the first transistor can be a unipolar transistor, such as a MOSFET. However, it is also possible to implement it using a bipolar transistor, such as an IGBT (Insulated-Gate Bipolar Transistor) with an antiparallel protection diode (integrated or as an external freewheeling diode).

[0020] According to one embodiment of the circuit module, the connection to the driver is formed with a conductor section into which a diode is inserted, which blocks current in the direction of the driver.

[0021] According to one embodiment of the circuit module, the control terminal and the source terminal of the second transistor are connected to each other, and a diode is arranged between the control terminal and the source terminal of the second transistor, which blocks current in the direction of the control terminal.

[0022] According to one embodiment of the circuit module, the control terminal and the sink terminal of the second transistor are connected to each other, and a resistor (preferably an ohmic resistor) is arranged between the control terminal and the sink terminal of the second transistor.

[0023] According to one embodiment of the circuit module, a voltage limiter is provided in parallel with the first transistor. A capacitive resistor can also be provided in parallel with the first transistor.

[0024] According to one initial embodiment of the circuit module, the control terminal and the source terminal of the first transistor are connected, and a diode is placed between the control terminal and the source terminal of the first transistor, which blocks current in the direction of the control terminal. This diode is an optional feature, implemented, for example, as a Zener diode, which limits the maximum gate voltage or control terminal voltage.

[0025] Protection diodes and voltage limiters can also be selectively implemented only for those circuit modules where the specific design of the semiconductor switch requires it. Furthermore, the circuit module closest to the driver, or, in the case of a symmetrically designed semiconductor switch, the two circuit modules closest to the driver without a second transistor, can be configured.

[0026] According to a second embodiment of the circuit module, the circuit module is formed with a third transistor, wherein the third transistor is of the same type as the first transistor and the first and third transistors are each directly connected to each other via their source terminals (i.e., they are arranged practically in opposite directions). In this embodiment, the source terminal of the second transistor is connected to the control terminals of the first and third transistors, and the sink terminal of the second transistor is connected to the junction of the source terminals of the first and third transistors.

[0027] The invention is described in more detail below using an exemplary embodiment and the figures. These show... Fig. 1a and Fig. 1b: a circuit module for forming a semiconductor switch according to the invention, Fig. 2: the construction of a semiconductor switch according to the invention with fuses for the circuit modules, Fig. 3: a section of the semiconductor switch according to the invention Fig. 2 with implementation details of the fuse, here as an electronic fuse and a threshold adjustable via a reference, Fig. 4: a section of the semiconductor switch according to the invention Fig. 2 with implementation details of the backup and a higher-level control system that is notified in case of malfunction, and Fig. 5: a section of the semiconductor switch according to the invention Fig. 2 with implementation details of the backup and a higher-level control system that is notified in case of malfunction and triggers the backup.

[0028] The invention is based on a semiconductor switch formed with circuit modules connected in series. Fig. 1a and Fig. Figure 1b shows an embodiment of such a circuit module. The module comprises a MOSFET M3 with a source terminal (source3), a gate terminal (Gate3), and a drain terminal, and a PNP bipolar transistor Q2 with an emitter terminal, a base terminal, and a collector terminal. The emitter terminal is connected to the gate terminal (Gate3) of the MOSFET M3, and the collector terminal is connected to the source terminal (source3) of the MOSFET M3. The gate terminal (Gate3) and the source terminal (source3) of the MOSFET M3 are connected together, and a diode D5 (preferably a Zener diode) is arranged between these terminals, which blocks current in the direction of the gate terminal (Gate3). The base terminal and the emitter terminal of the PNP bipolar transistor Q2 are connected together, and a diode D4 is arranged in the connection, which blocks current in the direction of the base terminal.The base terminal of the PNP bipolar transistor Q2 is also connected to its collector terminal, with a resistor R5 inserted in the connection.

[0029] Furthermore, a connection is provided from the base terminal of the PNP bipolar transistor Q2 to a driver V2. This is due to Fig. Figure 2 shows a semiconductor switch formed from modules connected in series with a driver V2. The driver V2 is connected to the PNP bipolar transistors Q1-Q6 of six circuit modules connected in series. A diode D1, D6, D7, D12, D14, and D16 is placed between the driver V2 and the base terminals of each of the PNP bipolar transistors Q1-Q6, blocking current in the direction of the driver V2. A fuse S1-S6 is also provided between the driver V2 and each of the diodes D1, D6, D7, D12, D14, and D16. In this switch geometry, the semiconductor switch is formed from an even number of modules, with the first half of the circuit modules connected in series with the same switching direction and the second half of the circuit modules connected in series with the opposite switching direction. Fig. 1a and Fig. 1b shows the semiconductor circuit module twice, with different forward bias directions. The semiconductor switch according to... Fig. 2 is therefore composed of modules correspondingly. Fig. 1a and half from modules accordingly Fig. 1b together. The Zener diodes D5 and D18 in Fig. 1a and Fig. 1b are optional. In the middle of the circuit of Fig. In Figure 2, two modules with opposite forward bias are connected in series. The driver V2 is connected to the junction of the two circuit modules with opposite bias (source) and is connected to ground via this junction and resistor R11. Resistors R2 and R6 are placed between this junction and the base terminals of the PNP bipolar transistors of the two circuit modules, respectively. Additionally, resistor R1 is placed between the positive terminal of driver V2 and the junctions of the circuit modules with one forward bias. Capacitors C1-C6 and a voltage limiter U1-U6 (e.g., a suppressor diode or varistor) are connected in parallel to each circuit module. This voltage limiter is an optional component of the circuit modules. A supply voltage V1 and a load R3 powered by this voltage are also shown.

[0030] For semiconductor switch construction according to Fig. The following two points are also relevant: Firstly, this is a circuit for alternating current flow (AC or DC with both current flow directions). For unipolar current flow, only half of the circuit shown would be needed (i.e., only the modules with transistors in one forward direction).

[0031] On the other hand, the driver modules can, in principle, be implemented without transistors Q5 and Q6 for turn-on. Components D13, D14, D17, and R17, or D15, D16, D18, and R10, can also be omitted, so that the modules then consist only of MOSFETs M1 and M4, respectively. However, to ensure that all series-connected components turn off as simultaneously as possible, or to symmetry the turn-off behavior, it is advantageous to construct these two modules in the same way as the others.

[0032] According to one embodiment of the semiconductor switch, it is designed to assume a safe "off" state (high resistance between drain and source) even in the event of a fault. The semiconductor switch can also be designed to maintain functionality even if a circuit module or the MOSFET of that circuit module (or a specified number of circuit modules) fails. In this case, the transition to the safe "off" state is not strictly mandatory; however, continued operation despite a failed circuit module would typically only serve as a temporary measure until the fault is repaired.

[0033] The simpler case is the transition to the safe "off" state. The safety of this state is ensured by the number of circuit modules and the design of the MOSFETs. Both are configured so that if one power semiconductor fails, the remaining sum of the blocking voltages of all still-functional power semiconductors is greater than the peak value of the load voltage.

[0034] As soon as a MOSFET in a series circuit module conducts short, its gate signal is short-circuited, or the MOSFET's drive circuit, formed with a PNP bipolar transistor, is overloaded and the gate signal (driver output) collapses. Therefore, one can monitor either the current of the driver output or the gate voltage or supply voltage of the driver stage to detect an overload.

[0035] Upon detection of this overload, the driver sets the gate signal of the MOSFETs to 0V, thus preventing current flow into the shorted MOSFET (no driving voltage is present). This also initiates the shutdown process for all other functioning MOSFETs or circuit modules. The semiconductor switch enters the "off" state. The occurrence of this fault or fault state can then either be passed to a higher-level control system to block the semiconductor switch from being turned on again, or it can be turned off again after the overload of the drive circuit of the defective MOSFET is detected during each subsequent turn-on process. Modern drive circuits already incorporate this functionality. For example, the TI UCC5390EC (gate driver from Texas Instruments) has a UVLO (undervoltage low-level output), meaning that the output is held "low" in the event of undervoltage at the supply voltage.If the driver's supply voltage is insufficient to provide the required driver power, the driver shuts down, and the semiconductor switch enters the "off" state. This fault state is then locked to prevent unauthorized reactivation, for example, by sending a message to a higher-level control unit (e.g., a microcontroller) of the semiconductor switch, which is programmed to prevent reactivation in the event of a fault. For example, when the semiconductor switch is reactivated, the higher-level control unit checks the driver circuit and, if necessary (e.g., if the driver output remains "low"), switches the semiconductor switch off again.

[0036] According to one embodiment of the invention, the faulty path is selectively isolated, or the faulty circuit module is disconnected. This is achieved with the semiconductor switch according to [reference to relevant document]. Fig. 2 is possible because the galvanic isolation to the driver in the switched-off state is handled by diodes D1 and D6. As in Fig. As shown in Figure 2, a fuse S1-S6 is connected in series with each of the diodes D1-D6. In the event of a fault, the fuse can trip in the path of the fault current, thereby disconnecting the faulty part or circuit module from the assembly of circuit modules. This fuse does not necessarily have to be a physical fuse; electronic solutions are also conceivable, e.g., a so-called "electronic fuse" (see https: / / www.elektronik-kompendium.de / news / thema / elektronische-sicherung / ).

[0037] In Fig. 3 is the use of an electronic fuse in the semiconductor switch according to. Fig. Figure 2 shows. For simplicity, only one circuit module is depicted, which can be disconnected by such a fuse. On the left is the driver V3, on the right a circuit module formed with the MOSFET M7, the PNP bipolar transistor Q7, the Zener diode D21, the diodes D19 and D20 and the resistor R14.

[0038] The electronic safety feature is implemented with a shunt resistor R12, which measures the gate current and checks for exceedance of a limit value. Operational amplifier U8 detects the voltage drop across resistor R12 caused by the gate current and amplifies the voltage signal. This signal is then compared to a predefined reference voltage by the comparator circuit formed by operational amplifier U9 and resistors R13 and R15. If a reference value for this voltage signal (corresponding to a limit value for the maximum gate voltage) is exceeded, a signal is output at the comparator output, triggering transistor U7 to switch off the gate path. To filter out the current pulse when MOSFET M7 is switched on during current measurement, it may be advisable to add a capacitor C7 in parallel with resistor R12.

[0039] According to a further training course (see Fig. 4) When the gate current is measured and a predefined setpoint is detected being exceeded, a message is sent to a higher-level microcontroller or MCU as a warning or parameter for control measures to be carried out by the MCU. Preferably, there is galvanic isolation for the MCU. The message can then be sent, for example, via an optocoupler U10.

[0040] According to one embodiment, the shutdown of the gate path is triggered by a superimposed microcontroller or MCU ( Fig. 5) As in the design of Fig. 4. A message is transmitted to the MCU via the optocoupler U10. The MCU then outputs control information to another optocoupler U11, which is used to split the path by means of a MOSFET M8.

[0041] The one based on the Fig. 2, Fig. 3, Fig. 4 to Fig.The semiconductor switch shown in Figure 5 allows for continued operation of the switch in the majority of failure scenarios based on stochastic failure mechanisms, provided the series connection includes redundant circuit modules. The reduced fit rate (failure in time) resulting from the connection of a larger number of power semiconductors can thus be significantly increased again, and it can also be ensured that the safe "off" state is reached. The fit rate is described, for example, at https: / / de.wikipedia.org / wiki / Failure_In_Time and is, in short, the value FIT = 10⁹h / MTBF. With an MTBF (mean time between failures) of one billion hours, the value for the FIT parameter is then equal to 1.

[0042] The rare occurrence of high resistance can be detected by measuring the voltage at the input and output of the semiconductor switch. In this fault condition, when the series circuit is switched on, a voltage difference on the order of the maximum blocking voltage of a power semiconductor (regardless of whether caused by a power semiconductor avalanche or activation of part of the overvoltage protection) occurs between the input and output, whereupon the switch can and must be switched off again.

Claims

[1] Semiconductor switches with - a plurality of circuit modules connected in series, each of which is formed with - a first transistor (M3) with a source terminal (source3), a control terminal (Gate3) and a sink terminal, wherein - the source terminal (source3) is a source terminal or an emitter terminal, the control terminal (Gate3) is a gate terminal or a base terminal, and the sink terminal is a drain terminal or a collector terminal, and - a second transistor (Q2) with a source terminal, a control terminal and a sink terminal, wherein - the source terminal is connected to the control terminal (Gate3) of the first transistor (M3), - the sink terminal is connected to the source terminal (source3) of the first transistor (M3), and - with a driver (V2) which is connected via connections to the control terminals of the second transistors (Q1, Q2, Q3, Q4, Q5, Q6) of the individual circuit modules, wherein - the semiconductor switch is designed to interrupt the connection between the control terminal of the second transistor (Q2) of the malfunctioning circuit module and the driver (V2) in the event of a malfunction of one of the circuit modules. [2] Semiconductor switch according to claim 1, characterized by that the semiconductor switch is designed to switch itself off in the event of a malfunction of one of the circuit modules. [3] Semiconductor switch according to claim 1 or 2, characterized by, that the semiconductor switch is designed by the number n of circuit modules connected in series and the design of the individual circuit modules for a maximum permissible value of a load current to be carried through the semiconductor switch such that the cumulative blocking voltage of each subset of n-1 circuit modules of the n circuit modules of the semiconductor switch is greater than or equal to the maximum value of the load voltage. [4] Semiconductor switches according to any one of the preceding claims, characterized by , that - Circuit modules of the semiconductor switch are formed with a fuse that is located in the connection between the control terminal of the second transistor (Q2) of the circuit module and the driver (V2), and - the semiconductor switch is designed to interrupt the connection between the control terminal of the second transistor (Q2) of this circuit module to the driver (V2) by means of the fuse located there in the event of a malfunction of a circuit module. [5] Semiconductor switch according to claim 4, characterized by , that the semiconductor switch is designed so that if a current threshold value is exceeded on the connection between the control terminal of the second transistor (Q2) of the circuit module and the driver (V2) due to a malfunction of a circuit module, the fuse located there will trip. [6] Semiconductor switch according to claim 5, characterized by , that - it is an electronic fuse, and - the connection between the control terminal of the second transistor (Q2) to the driver (V2) is formed with a circuit for detecting the exceeding of a current threshold value. [7] Semiconductor switch according to claim 6, characterized by , that - the semiconductor switch includes a control unit, and - the circuit is designed to detect when a current threshold is exceeded and to output a signal to the control unit, which signals that the current threshold has been exceeded. [8] Semiconductor switch according to claim 7, characterized by , that the control unit is designed for generating and transmitting a control signal for the fuse to trigger it when the current threshold is exceeded. [9] Semiconductor switch according to claim 6 or 7, characterized by, that the circuit is designed to detect the exceeding of a current threshold for the generation and transmission of a control signal for the fuse to trigger it when the current threshold is exceeded. [10] Semiconductor switches according to any one of the preceding claims, characterized by - an evaluation circuit designed for comparing a voltage difference with a threshold value, wherein the threshold value is determined according to the blocking voltage of a first transistor (M3) of a circuit module, - Voltage measurement circuits at the input and output of the semiconductor switch, which are designed to transmit measured voltage values ​​to the evaluation circuit, and - a circuit to switch off the semiconductor switch when the threshold is exceeded.

Images