JP4853928B2 - Control device and control method for silicon carbide static induction transistor - Google Patents

Description

  The present invention relates to a control apparatus and a control method for a high voltage, low on-resistance vertical electrostatic induction transistor made of silicon carbide, and more particularly to a control method when an overcurrent exceeding a rating flows.

  Various power conversion devices using power semiconductor elements are located between a power supply source (power supply) and various electric appliances (loads) that consume power, and convert power from AC to DC or from DC to AC. It plays the role of adjusting the output of the load by controlling AC / DC conversion, cutting off the current, and controlling the amount of power passing through. Such a power conversion device is required to have an overload resistance capable of withstanding a short-time overcurrent operation of the load. Normally, the device must operate normally even if the operation continues for several minutes at an output of about 1.5 to 2 times the rated capacity of the conversion device.

  In the operation for several minutes, the temperature rise of the power semiconductor element caused by the overcurrent during that time cannot be absorbed only by the heat capacity of the semiconductor element. For this reason, the maximum temperature of the semiconductor element used in the converter must be kept from exceeding the upper limit temperature even in an overload operation of 1.5 to 2 times the rated value. As a result, the capacity of the semiconductor elements used is larger than the rated capacity of the converter, and the converter cooling system is also enlarged accordingly. This leads to an increase in the size and weight of the conversion device, which is a major factor in increasing the price of large-capacity power conversion devices that are used especially in power distribution systems, etc. Yes.

  In order to solve this problem, it is considered to use a power semiconductor element made of silicon carbide whose upper limit temperature of operation can be higher by 100 ° C. or more than a power element made of silicon. That is, in the semiconductor element for silicon carbide power, the temperature of the element at the rated operation has a sufficient margin with respect to the upper limit temperature of the operation, so that the rated capacity is 1 without increasing the capacity of the element used or the cooling system of the converter. This is because it is relatively easy to prevent the maximum temperature of the element from exceeding the upper limit temperature of operation in an overcurrent state of 5 to 2 times. Thus, if silicon carbide power semiconductor elements having a significantly lower operating loss at the rated capacity than silicon elements and having a high operating upper limit temperature are used, not only the conversion efficiency during rated operation of the converter is increased. Also, a great effect can be expected in reducing the size and weight of the device.

  However, in the power conversion device connected in the system for the purpose of stabilizing the power transmission / distribution system and maintaining the power quality, the conventional silicon semiconductor element is simply replaced with the silicon carbide semiconductor element for the following reasons. It's not enough. That is, in the electric power system, the converter used in the system has a rated current of 10 in case of a short circuit accident of various electric devices on the load side linked to the system, an accident such as an interphase short circuit or a ground fault in the transmission and distribution lines. Strict specifications are attached so that the function of the converter is not lost even if the overcurrent of double to 20 times is applied for a period of half cycle to several cycles (5 msec to 100 msec). It is because it is required to endure.

  In the case of power conversion devices such as motor rotation speed control, it is often installed near the motor that is the load, so in the event of an abnormality such as a short circuit of the load, the abnormality is quickly transmitted to the control circuit of the conversion device. Controls such as shutting off the converter or reducing the output can be made relatively easy. However, in converters used in power transmission / distribution systems, the distance between the fault point of the grid and the point where the converter is installed is usually long, so it takes a long time to control the system. In order to prevent the power supply from other normal points as much as possible, it is not permitted to immediately shut off only a part of the converters in the system or to reduce the current. In other words, there are a number of mechanical contact type circuit breaker devices that cut off the current using a protective relay and electromagnetic force in the power system, but these devices can quickly cut off current due to the overcurrent that flows in the event of an abnormality. In order to operate these devices safely and to stabilize the entire system, when a part of the system malfunctions, the linked converter is instantly shut off. This is because it is not allowed to do or to narrow down the current. For this reason, the converter for the power system is required to have an overload resistance that can withstand energization of extremely severe surge current as described in the previous section [0005].

  Conventional power converters for power transmission and distribution systems that withstand overloading of surge current include thyristors and gate tan-off thyristors (GTO thyristors) that are particularly resistant to surge current among silicon power semiconductors. ) Or a bipolar operation type element such as an insulated gate bipolar transistor (IGBT) is used. In the thyristor element, the conductivity of the internal high resistance layer is remarkably increased by injecting electrons and holes from the two emitters of the cathode and the anode in the energized state, so that even if a current of 10 to 20 times the rated current flows, the element The voltage drop increment is relatively small and has the advantageous feature that it does not lead to breakdown for short periods of half or several cycles. In addition, since the IGBT element modulates the conductivity of the high resistance layer by injecting holes from the anode emitter, if the current capacity of the element is increased to about 2 to 3 times the rated capacity, an excessive surge current can be generated. Even when energized, it can operate safely without a significant increase in on-voltage.

  However, when using a power semiconductor device made of silicon carbide for a power converter for a power transmission / distribution system for the reason described in [0004], the thyristor, GTO, or IGBT shown in [0008] is used. When the bipolar operation type power semiconductor element is applied, there is a problem that the conversion efficiency during normal operation at or below the rating of the apparatus is significantly impaired as described below. That is, when these bipolar devices are in the ON state, conductivity modulation occurs due to minority carrier injection from the pn junction, and a large current flows. In order for this carrier injection to occur, a forward voltage higher than the built-in voltage (threshold voltage) must be applied to the pn junction. By the way, the built-in voltage of the pn junction depends on the size of the band gap (forbidden band width) of the crystal material of the semiconductor element. Since the built-in voltage of a silicon carbide pn junction with a large band gap is about 2.5 V, which is more than twice as large as about 1.0 V of silicon, when the silicon carbide bipolar device is in the ON state, the forward voltage drop In this case, the voltage of 2.5 V is always applied, and this loss increases the generation loss of the element and the power loss of the converter. As a result, the conversion efficiency of the apparatus in a normal operation state at a rated current or lower is remarkably impaired. Furthermore, in high-voltage power converters such as power distribution systems, semiconductor elements are often used in the form of series connection, but current loss due to built-in voltage is accumulated in proportion to the number of elements in series. Such a bipolar operation type element is disadvantageous in terms of increasing the efficiency of the converter.

  On the other hand, a unipolar operation type electrostatic induction transistor (junction field effect transistor; sometimes referred to as JFET) is applied as a power semiconductor element made of silicon carbide used in a power converter. It is advantageous. FIG. 1 shows what was described in Patent Document 1 as a typical example of a silicon carbide static induction transistor. A low-concentration n-type layer 2 is laminated on a silicon carbide high-concentration n-type single crystal substrate 1, and a plurality of high-concentration p-type layers 3 are embedded in the low-concentration n-type layer 2. A high-concentration n-type layer 4 is laminated. A drain electrode 5 is formed on the surface of the high-concentration n-type single crystal substrate 1, a source electrode 6 is formed on the surface of the high-concentration n-type layer 4, and a gate electrode 7 is formed on a part of the surface of the high-concentration p-type layer 3. Each has a low resistance ohmic contact. The n-type region 21 sandwiched between the two high-concentration p-type layers 3 is called a channel and spreads in the channel controlled by the polarity and magnitude of the gate voltage Vg applied between the gate electrode 7 and the source electrode 6. The depletion layer turns on and off. That is, the channel is pinched off by a sufficiently large reverse gate voltage Vg to be turned off, and turned on when the gate voltage Vg is removed or a low voltage is applied in the forward direction. In addition to this example, various structures of transistors have been proposed as the silicon carbide static induction transistor, but the operation is the same. Therefore, the operation and features will be described in detail below based on this example.

  FIG. 2 shows a representative example of output characteristics of a silicon carbide static induction transistor. This is a voltage / current characteristic between the drain voltage Vd and the drain current Id when a gate voltage Vg of 0 V, 1.0 V, 2.5 V, and 3.0 to 5.0 V is applied between the gate and the source. Id hardly flows when Vg = 0V, but when Vg ≧ 0V, Id increases linearly as Vd increases and eventually saturates. The on-resistance of this linear region is about 0.1 ohm when Vg = 1.0V, 0.05 ohm when Vg = 2.5V, and about 0.01 ohm when Vg = 3.0 to 5.0V. These are extremely low on-resistances and become smaller as Vg increases. Further, the saturation current also increases as Vg increases. This is because when Vg ≦ 2.5V, the application of the forward gate voltage suppresses the spread of the depletion layer extending to the channel region, and the substantial width of the channel region serving as a current path is increased. When Vg = 3.0 to 5.0 V, holes are injected from the high-concentration p-type layer 3 into the channel region 21 in addition to the substantial expansion of the channel region, and the conductivity there is remarkably high. Is due to the result. Vg = 2.5V is approximately equal to the built-in voltage of the gate-source pn junction of silicon carbide. In the bias state of 2.5 V or less, since minority carrier injection from the gate electrode does not occur, Vg ≦ 2.5 V needs to be satisfied in order to retain the characteristics of the unipolar operation type semiconductor device that switches at high speed. On the other hand, when a high voltage of 3.0V to 5.0V exceeding the built-in voltage 2.5V of the silicon carbide pn junction is applied as the gate voltage, the on-resistance in the linear region is further reduced to 0.01 ohm. The saturation current also increases drastically to about 200 A, the on-resistance is reduced to about 1/5, and the saturation current is increased about 10 times compared to the steady operation at Vg = 2.5V. However, when Vg> 2.5 V, the switching speed is slow as in the bipolar operation type due to the effect of minority carrier accumulation, and there is a problem in characteristics that circuit components cannot be miniaturized due to high frequency operation.

  In these output characteristics, since the drain current rises from the drain voltage Vd = 0V, the loss due to the built-in voltage at the pn junction is not added to the conduction loss as in the bipolar operation type element described in the previous section [0008]. The conduction loss can be reduced. Also, even in series-connected high voltage operation, voltage loss is not accumulated by built-in in proportion to the number of series. In addition, since the on-resistance of the element increases as the temperature rises, current concentration is unlikely to occur, and parallel connection of elements in the converter and parallel mounting of semiconductor chips in one element can be realized relatively easily. This is superior to conventional bipolar operation type elements such as thyristors and IGBTs.

  Compared to MOSFETs that are widely used as field effect transistors, both are elements that are turned on and off by a voltage signal from a gate circuit applied between the gate and the source. Insulated by an oxide film or the like, a channel serving as a current conduction path is formed on the semiconductor surface immediately below the gate insulating film by application of the gate voltage signal, and is turned on. Although the on-resistance strongly depends on the electron mobility in this channel, the channel mobility of the MOSFET is as low as 1/10 or less of the bulk mobility due to the presence of interface defects between the silicon carbide semiconductor and the oxide insulating film. Therefore, the on-resistance of the MOSFET is not so small. On the other hand, the electrostatic induction transistor has a feature that the channel mobility is as high as the bulk mobility, and a low on-resistance is easily realized. In addition, since an insulating film such as SiO2 is interposed between the gate and silicon carbide in the MOSFET, the on-resistance is reduced or the saturation current is increased by injecting carriers from the gate electrode as in the electrostatic induction transistor. Control is not possible.

  In the silicon carbide static induction transistor having the output characteristics shown in FIG. 2, when an operation with Vg ≦ 2.5 V is performed with the intention of a unipolar operation in a steady operation, an overcurrent exceeding twice the rated value is energized. There is a problem that the energization loss at the time increases significantly. In other words, in the forward ON state at Vg = 2.5 V, if the rating of this element is set to about 10 A, the energization loss of the element in the operation below the rated current can be made sufficiently small, even twice the rating. Even if an overcurrent of a certain level flows, the generated loss can be kept small without a significant increase in the drain voltage Vd. However, when a current greater than the current value at which the output current saturates flows, the drain voltage Vd increases rapidly, resulting in a significant increase in current loss in the device. When the current flow time is long, the device temperature exceeds the upper limit temperature. In some cases, the device may be destroyed. Therefore, a device having a considerably large area is required to be applied to a converter in which a surge current 10 to 20 times the rated value flows.

On the other hand, when a high voltage exceeding the built-in voltage such as Vg = 3.0V to 5.0V is supplied to the gate in order to reduce the on-resistance during forward energization in the normal operating state below the rated current, the Although the resistance can be reduced, there arises a problem of a so-called bipolar operation type element in which the switching speed is impaired by the effect of accumulation of minority carriers. Furthermore, since the gate current is always applied during the ON operation due to the forward bias of the gate junction, there is a problem that not only the gate circuit is enlarged but also the loss of control power cannot be ignored.
JP 2006-253292 A

  Distribution using a silicon carbide static induction transistor of unipolar operation because high conversion efficiency in normal operation below the rated capacity of the power converter, high-speed switching operation, and high resistance to overload current twice the rating are obtained. In a power converter for high power such as for a system, when a surge current of 10 to 20 times the rated current flows in the converter, excessive energization due to a significant increase in on-resistance in the surge current region of the transistor used There is a problem that the transistor is thermally destroyed due to loss, and the power converter is broken down.

  In view of these problems, an object of the present invention is to provide a control device and a control method for a silicon carbide static induction transistor element in which a significant increase in on-resistance loss is suppressed in an overcurrent of a power converter.

  Another object of the present invention is to provide a transistor control device and a control method for realizing a power converter with high conversion efficiency that is small, light, and low in cost using a silicon carbide static induction transistor. That is.

  In order to solve the above problems, the present invention provides a gate voltage as a control signal for turning on a transistor during normal operation within the rated capacity of a power converter using a silicon carbide electrostatic induction transistor. The built-in voltage (threshold voltage) of the pn junction between the transistors is set to a voltage lower than the built-in voltage (threshold voltage) only during overload energization to significantly reduce the on-resistance of the transistor and prevent thermal destruction of the element. Features.

  As the gate voltage, a voltage of 2.5 V or less during steady operation is set to a voltage in the range of 3.0 V to 5.0 V during overload energization. At the time of the overload energization, the current is at least five times the rated capacity of the converter.

  As described above, the present invention has the following effects. Even when a surge overcurrent is applied that significantly exceeds the rated capacity of the power converter element to which a silicon carbide static induction transistor with high speed, low loss and strong overcurrent resistance is applied, the loss generated by the applied element is significantly increased. The transistor's current control method eliminates the need for the current capacity of the transistor used to be significantly larger than the rated capacity of the device, and simultaneously reduces the size, weight, efficiency, and cost of power converters. realizable.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings. FIG. 3 shows one embodiment of the control apparatus for the silicon carbide static induction transistor of the present invention. This device is roughly composed of a gate drive circuit, a gate voltage control circuit, and a current detection circuit. The gate driving circuit includes a capacitor C1 serving as a 2.5V on-gate voltage source, a capacitor C2 serving as a 15V off-gate voltage source, and a voltage of C1 and C2 via a gate resistor _RG according to an on / off command from a photocoupler. It is composed of switches Trs1 and Trs2 for supplying to the gate terminal G of the silicon carbide electrostatic induction transistor, and the like, during normal operation in which a current less than the rated capacity or about twice the rated current flows through the silicon carbide electrostatic induction transistor. It is a drive circuit. The gate voltage control circuit is composed of capacitors C3 and Cs3 that are connected in series to capacitors C3 and C3 serving as a 5V on-gate voltage source, and a gate circuit of Trs4, and the 2.5V on-gate voltage source C1 of the gate driving circuit. Are connected in parallel, and when the current flowing through the silicon carbide static induction transistor greatly exceeds the rated current, the gate terminal G of the transistor has a gate voltage exceeding 2.5V, in this example 5.0V , Is a circuit for supplying. The current detection circuit is composed of a resistor RL connected in series with a current flowing through the silicon carbide electrostatic induction transistor, and a circuit that compares a current value detected by a voltage drop at both ends thereof with a specified value. It is determined that the current value has greatly exceeded the rated current of the transistor, and a command signal for turning on the switch Trs4 of the gate voltage control circuit is output.
This control device uses a gate drive circuit to turn on the gate voltage at the time of normal operation when the current flowing through the transistor is equal to or lower than the rated current or about twice the rated value, and the built-in voltage of the pn junction between the gate and the source of the transistor (approximately 2.5 V) or less. When a current significantly exceeding the rating flows, the switch Trs4 of the gate voltage control circuit is turned on by a signal output when the current detected by the current detection circuit exceeds a predetermined value, and exceeds the built-in voltage. The gate voltage (5.0 V in this example) is controlled. A specific example of the control method of the present invention using such a control device will be described below.

  FIG. 4 illustrates a method for controlling the silicon carbide static induction transistor of the present invention. The upper stage is a command signal for holding the electrostatic induction transistor on or off, the middle stage is a gate voltage Vg applied from the gate circuit to the electrostatic induction transistor, and the lower stage is a drain / source of the electrostatic induction transistor. A voltage (drain voltage) Vd between the drain and the source (drain current) Id is shown. The change of each value is represented by the time elapsed. In the off command state before time t1, the electrostatic induction transistor is turned off by the gate voltage of Vg = -15V, and the state of the drain current Id = 0 is maintained while the drain voltage Vd is applied. Note that the gate voltage Vg in the off state is not necessarily -15V, and may be another value such as Vg = 0 or -50V depending on the transistor. When the command signal is switched from OFF to ON at time t1, a voltage of Vg = 2.5V is applied to the gate, the transistor is turned on, and the drain current Id starts to flow. The drain voltage is held at a low value of about 1 V in proportion to the magnitude of Id. Note that the gate voltage Vg at the time of turning on does not necessarily need to be 2.5V. It may be less than the built-in voltage (threshold voltage) of the gate junction of the silicon carbide static induction transistor in which minority carriers from the gate junction are not injected in order to cause a high-speed switching operation by the unipolar type operation. Other values such as .5V may be used. As shown in FIG. 2, the higher the Vg, the lower the on-resistance, so a voltage in the range of 2.0 to 2.5 V that does not exceed the built-in voltage is good. When the operation is normal, the command signal is turned off at t5, and the gate voltage Vg is applied as it is until the gate voltage of Vg = -15V is applied again and the transistor is turned off (dotted line (a) in the figure). Continue on state.

  In FIG. 4, during the period when the on-command signal from t1 to t5 is applied, if a current that greatly exceeds the rated current starts flowing through the transistor at a certain time t2 that is not predicted, the gate voltage Vg is immediately set to 5. Boost to 0V. The step-up of Vg is continued during a period up to time t4 when a drain current Id decreases by a surge current exceeding 5 times the rated current of the transistor and decreases below this predetermined current value. Although the case where the time t4 is before the time t5 when the command signal is switched from on to off is illustrated here, the time t4 may be after the time t5. In that case, control is performed so as to continue boosting Vg in preference to the command signal. Further, the gate voltage Vg during the overcurrent period from the time t3 to the time t4 is not necessarily 5.0V, and may be a voltage exceeding the built-in voltage.

  In the specific embodiment of FIG. 4, in the conventional control method such as the dotted line (a) of the gate voltage in the figure in which the gate voltage is set to 2.5 V which is the same as in normal operation even when the overcurrent is abnormal, the time in the figure As the drain voltage Vd of the element from t3 to time t4 is applied with a high voltage equal to the same power supply voltage as in the off state as indicated by a dotted line (a), the loss generated in the element becomes extremely large. On the other hand, in the control method of the solid line (b) of the gate voltage in the figure in which the gate voltage at the time of overcurrent abnormality is boosted to 5.0 V as in this specific embodiment, the element from time t3 to time t4 The drain voltage Vd becomes a low value as shown by the solid line (b), and the generated loss in the device is suppressed to a small value. Note that Vd has a high value during the period from when an overcurrent in the vicinity of time t3 is detected until the gate voltage is boosted, but since this period is an extremely short time of 10 microseconds or less, it has an effect on the generation loss of the element There are few.

  As described above, the device and the control method at the time of overcurrent when the electrostatic induction transistor is applied to the power converter using silicon carbide as a material have been described. As the power converter to which the electrostatic induction transistor is applied, an inverter device or a converter is used. Also included are power order, reverse conversion devices, various chopper devices that increase and decrease the voltage, and AC and DC circuit breakers. Further, the silicon carbide electrostatic induction transistor has been described based on the example of the control circuit and the control method shown in the figure, but the present invention is not limited to the above-described example, and within the scope of the claims. It includes other configurations that can be easily modified by a trader.

Cell cross section of a typical silicon carbide static induction transistor Typical example of output characteristics of silicon carbide electrostatic induction transistor One Embodiment of Control Device for Silicon Carbide Static Induction Transistor of the Present Invention One Embodiment of Control Method of Silicon Carbide Static Induction Transistor of the Present Invention

Explanation of symbols

1. 1. High concentration n-type single crystal substrate Low concentration n-type deposited film (drift layer)
3. 3. High concentration p-type layer 4. High concentration n-type source layer 5. drain electrode 6. Source electrode Gate electrode 21. n-type channel region 51. Drain electrode terminal 61. Source electrode terminal 71. Gate electrode terminal

Claims (5)

In an electrostatic induction transistor control device made of silicon carbide,
A circuit for detecting an element current flowing in the electrostatic induction transistor;
A circuit for controlling the gate voltage based on the output of the element current detection circuit,
The gate voltage control circuit controls the gate voltage at the time of on-state during normal operation below the rated current to a voltage below the pn junction built-in voltage between the gate and source of the electrostatic induction transistor and exceeds the rating. A device for controlling an electrostatic induction transistor comprising preventing a device breakdown due to an overcurrent by boosting a gate voltage to a built-in voltage or higher only when a current flows.   2. The electrostatic induction transistor according to claim 1, wherein the gate voltage not exceeding 2.5 V as the gate voltage less than the built-in voltage and the gate voltage exceeding the built-in voltage is 3.0 to 5.0 V. 3. Control device.   3. The electrostatic induction transistor control device according to claim 1, wherein the overcurrent exceeding the rated current is a current that is approximately five times or more the rated current. 4.   3. The electrostatic induction transistor control device according to claim 1, wherein the overcurrent exceeding the rated current is a current that is approximately 10 times or more the rated current. 4. In the control method of the electrostatic induction transistor made of silicon carbide,
A circuit for detecting an element current flowing through the electrostatic induction transistor, and a circuit for controlling a gate voltage based on an output of the element current detection circuit;
The gate voltage control circuit controls the gate voltage at the time of on-state during normal operation below the rated current to a voltage below the pn junction built-in voltage between the gate and source of the electrostatic induction transistor and exceeds the rating. A method for controlling an electrostatic induction transistor, comprising preventing element destruction due to overcurrent by boosting a gate voltage to a built-in voltage or higher only when a current flows.

Images