JPS643061B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a semiconductor device capable of controlling a large current at high speed, and particularly to improvements in the structure thereof.

In recent years, high-power semiconductor devices have been used to perform high-precision and high-efficiency control of various power equipment, and in this field, high-power semiconductor devices that are capable of high-speed switching and have low loss are required. ing. If a semiconductor device with a slow switching speed were used to perform high-power switching control, for example, the control frequency would not be raised and would remain in the human audible frequency range, causing the sound generated by the equipment to cause discomfort to the worker. , the switching loss of the semiconductor device itself is large, resulting in financial damage such as difficulty in thermal design of the equipment. On the other hand, due to recent advances in semiconductor technology, the switching speed of high-power semiconductor devices themselves has improved dramatically, and the switching speed of several hundred A current can be reduced to 0.1 μsec.
Devices are being realized that can be turned on and turned off in a fraction of a second or less. If such a device is applied to the above-mentioned purpose, it will become possible to realize a power equipment in which switching loss hardly becomes a problem even if the operating frequency is sufficiently higher than the audio frequency range.

By the way, when a semiconductor device is actually used, it is usually sealed in a package. FIG. 1 shows an exploded sectional view of an apparatus in which a semiconductor device is mounted on a conventional ceramic seal pressure-contact type package. As shown in the figure, this device includes a semiconductor device 1 such as a transistor or a thyristor, molybdenum plates 2a and 2b, metal conductor blocks 3a and 3b such as copper, flanges 4a and 4b, welding plates 5a and 5b, a ceramic case 6, and a seal. It consists of a stop pipe/control electrode extraction hole 7 and a control electrode lead 8. This package is usually cylindrical. This example is designed to encapsulate one semiconductor device substrate. Therefore, it is unsuitable for use in encapsulating a plurality of semiconductor device substrates. For example, when trying to realize a high power semiconductor device of several 100A,
By configuring a plurality of power semiconductor devices connected in parallel, it is possible to achieve this without significantly increasing the area of each semiconductor device substrate. Normally, yields can be expected to improve by manufacturing semiconductor device substrates divided into appropriate sizes rather than manufacturing one large-area, large-current-capacity semiconductor device substrate. Encapsulating device substrates is difficult.

FIG. 2 shows an example of a conventional ceramic case flat type package in which a plurality of semiconductor device substrates are mounted. In FIG. 2, a semiconductor device is constructed of semiconductor device substrates 11a, 11b, metal base 12, electrode leads 13a, 13b, ceramic case 14, and bonding wires 15a, 15b. This example is designed to be able to mount multiple semiconductor device substrates, but since it is intended for ultra-high frequency semiconductor devices, the semiconductor substrate to be mounted is considerably smaller than a high-power semiconductor device substrate, and the electrodes The leads 13a and 13b are narrowed and taken out to the outside of the package to facilitate impedance matching with the ultra-high frequency circuit. The term "electrode lead" as used herein refers to the connection part that is metallized on ceramic by vapor deposition or printing and is to be connected to the electrode part on the semiconductor device substrate with a wire or ribbon lead, and the lead part that is taken out to the outside of the ceramic case. and electrode conductors that are brazed or welded to the external metallized part.

Normally, high-power circuits are constructed with an impedance that is one order of magnitude or less smaller than the impedance of ultra-high frequency circuits, so the semiconductor device shown in the example in Figure 2 is designed for high-power applications that rapidly switch currents of several hundred amperes. It cannot be expected to be used as a semiconductor device. This situation is the same even when a thin wire is used like the control electrode lead 8 in the example shown in FIG.

The next semiconductor device capable of high-speed switching of large currents was the electrostatic induction thyristor (SITh).
The above situation will be explained using an example. A typical n-channel SITh has electrodes of an anode A, a cathode K, and a gate G, as shown in FIG. 3, and the main current flows from the anode A to the cathode K. The gate G is used as a control electrode for flowing or cutting off the main current. The ideal operation of switching is that the externally applied gate signal is transmitted to the gate of the device without delay, and as a result, the main current is turned on and off instantly, and when on, the anode and cathode The voltage drop between them is close to OV, and very high voltages can be blocked when they are off. SITh
In the device itself, for example, when several hundred A is applied, the voltage drop between the anode and cathode is about 1V.
The blocking voltage when off is several KV, and switching on and off can be performed in approximately 0.1 μsec or less. FIG. 4 shows a lumped constant equivalent circuit when such SITh is enclosed in a package. Here, R _A and L _A are the resistance and inductance of the anode electrode lead, respectively, and R _K and L _K
are the resistance and inductance of the cathode electrode lead, and R _G and L _G are the resistance and inductance of the gate electrode lead, respectively.
It is inductance. In other words, both are the impedance of the package itself containing SITh.
The voltage drop when a time-varying current I flows through a conductor having a resistance R and an inductance L is given by RI+LdI/dt. Now, if we were to change a current of 100A in a conductor with a resistance of 1mΩ and an inductance of 10nH in a time of 0.1μsec, the voltage drop would be 10V, mostly due to the effect of the inductance component. Let's consider the switching operation of SITh based on these considerations. Generally, when switching a semiconductor device (thyristor or transistor) at high speed, a current is rapidly supplied to the gate electrode or base electrode of the semiconductor device to reach a predetermined voltage when it is turned on, and conversely when it is turned off. Sometimes it is necessary to draw current at high speed to reach a predetermined voltage, but in some cases, the control electrode lead 8 is linear as in the conventional example shown in Fig. 1, or in the conventional example shown in Fig. 2. If the electrode leads 13a and 13b are narrow as in the case of narrow electrode leads 13a and 13b, their impedance causes a delay time for a signal applied to the control electrode lead end from the outside of the semiconductor device to reach the control electrode of the semiconductor device. This component may cause overshoot, which causes waveform disturbance. In particular, when turning off thyristors and transistors that switch large currents, a larger current than the control electrode must be drawn, and the above-mentioned tendency becomes noticeable and hinders high-speed operation. In SITh, a large amount of electrons and holes are injected into the channel part when conduction occurs, so it is necessary to suck them out of the channel part when turn-off. The faster the turn-off speed is attempted, the greater the current instantaneously flowing into the control electrode. This situation is explained by the following equation (1). The turn-off time t _pff of SITh is approximately given by t _pff = τ _eff · ln (1 + I _A /I _GP ) (1). However, τ _eff is the effective carrier life,
I _A is the anode current and I _GP is the peak gate current at cut-off. Equation (1) shows that the larger I _GP becomes, the shorter t _pff becomes. If the gate impedance of SITh including the outside is Z _G , then I _GP Z _G cannot be greater than the reverse gate bias V _GK applied to the gate. The smaller the Z _G , the smaller the V _GK and the larger the
It can flow I _GP and perform high-speed shutoff. That is, it is necessarily required that the sum of the control electrode impedance of the device itself and the impedance of the control electrode lead of the package be small. In particular, high-speed, large-current switching requires a small inductance component. However, as mentioned above, there are only a few
What happens if there is a voltage drop of 10V with an inductance of 10nH? In other words, the drive circuit for driving the semiconductor device must supply more voltage than the voltage required to control the semiconductor device itself by this voltage drop. On the other hand, the voltage applied to the control electrode of a semiconductor device is usually at most several
It is about 10V. Therefore, in high-current, high-speed switching, even if the inductance is small, most of the voltage that can be applied to the control electrode is accounted for by the voltage drop at the control electrode lead of the package.

Now, if we consider the self-inductance of the control electrode lead, the self-inductance L of a wire with diameter 2r and length l is L = μol/2π (ln2l/r-1) (H) (2) Width W, The self-inductance L of a plate-shaped conductor of length l is L=μol/2π(ln2l/W+1/2+W/3l)(H)(3
) is given by 2r=1mm, l=50mm, and W=5
When mm and l=50mm, L is 43nH and 35.3nH, respectively.
Therefore, it is impossible to perform high-speed, large-current switching just by considering the control electrode.

Next, let's consider the cathode electrode and the anode electrode. Since the main current to be switched will flow through these electrodes, it is natural that they should have an area that can withstand a current of several hundred amperes, but when considering the effect of self-inductance, the following disadvantages arise. will occur. Suppose that SITh is now in the process of being turned on by the control signal.
When trying to turn on SITh, apply the gate bias voltage when turning off
The gate voltage must be increased from V _GKpff to a sufficient gate bias voltage V _GKpo to induce turn-on. The gate bias voltage at this time is the voltage of the gate electrode based on the cathode electrode of the SITh device itself. gate voltage
When the main current begins to flow in the process of increasing from V _GKpff to V _GKpo , the cathode voltage of the device increases by L _K dl/dt due to the inductance of the cathode electrode lead. Due to this increase in the voltage of the cathode electrode, the gate voltage V _GK with respect to the cathode decreases relatively, thereby delaying the turn-on time. That is,
This is why the inductance causes a negative feedback effect. The same thing can be said for the turn-off process if the sign is reversed. This means that the gate drive circuit must have additional voltage supply capability. W = 50mm, L = 20mm, which seems to be quite wide as an electrode lead.
In this case, when calculating the self-inductance L using equation (3), it becomes 16.1 nH, which is quite small.
If a current of 100A were to flow in a time of 0.1μsec, the voltage drop would be 16.1V, which means that high-speed switching of extremely large currents would not be possible with this kind of thinking. Also, even if we consider the anode electrode lead, it is explained in exactly the same way as in the above case that it is impossible to quickly reach an ON voltage as low as 1V from a high voltage of hundreds or thousands of V just by considering its self-inductance. be done. In this way, high-speed, large-current switching cannot be achieved with the concept of a multi-layer constant model as shown in FIG. 4, which considers only self-inductance, and the package must be constructed in the form of a transmission line as a distributed constant circuit.

In view of the above points, the present invention aims to provide a semiconductor device in which all the electrode leads of the package are made wide in the form of a plate, making it easier to obtain a desired low impedance, and capable of high-speed operation with a large current.

The present invention will be described below with reference to the drawings.
FIG. 5 shows an exploded top view (a) and an exploded cross-sectional view (b) of an embodiment of a semiconductor device in which all electrode leads of the present invention are made into plate shapes to reduce the inductance component. This device includes semiconductor devices 21a, 21b...21j such as thyristors and transistors, and a molybdenum plate 22.
a, 22b...22j, a conductor base 23 which itself becomes a first electrode and is connected to the second main surface (lower surface in the figure) of the semiconductor device with the molybdenum plate; A conductor plate lead 24 serving as a second electrode connected to the main surface (the upper surface in the figure), a conductor plate lead 25 serving as a third electrode, bonding wire leads (or ribbon leads) 26a, 26b, and ceramic It is composed of a case 27 made of an insulator such as the like. Here, the conductor plate leads 24 and 25, which serve as electrodes, are metalized by vapor deposition or printing on an insulator such as ceramic, and are connected to the electrode parts on the semiconductor device substrate by wires or ribbon leads, and the case. It collectively refers to the lead part taken out to the outside and the electrode conductor brazed or welded to the external metallized part. Of course, these may be a single conductor plate.
The conductor base 23 is usually made of metal having high electrical conductivity and high thermal conductivity. Further, although the case where a plurality of molybdenum plates 22 are provided is shown, the molybdenum plate may be a single elongated plate. In this way, the electrode conductor base 23 and the conductor plate leads 24, 25 are wider than the sum of the dimensions of one side of the plurality of semiconductor device substrates accommodated, and the conductor plate The leads 24 and 25 have a width greater than the distance between them. The electrode conductor base 23 can mount all the devices accommodated thereon, and the conductor plate leads 24, 25 are electrically coupled to the electrode portions of all the devices by a number of parallel bonding wire leads 26a, 26b. In this example, semiconductor devices 21a, 2
1b, . . . 21j are all separated, but it is not necessary to do so. Semiconductor devices 21a, 21 are placed on the silicon wafer.
b, . . . 21j may be formed into a long and thin piece and configured integrally, or, of course, a predetermined number of pieces may be formed into a single piece and arranged in a long and thin manner. These matters can be determined based on each technology level so that the cost can be kept low considering the yield issues in the semiconductor process and the yield of the assembly process such as the yield when uniformly metalizing the semiconductor device onto the molybdenum plate. In short, it is important to arrange the device so that the rapidly changing current spreads out and does not flow locally and increase the inductance. In FIG. 5, the conductor plate leads 24, 25 and the semiconductor device are connected by wire leads or ribbon leads. In the configuration shown in FIG. 5, this is the part where the current flows in a concentrated manner. Therefore, it is necessary that the distance between the conductor plate leads 24 and 25 and the semiconductor device 21 be as narrow as the physical arrangement allows. Further, if the semiconductor device itself is too long in the direction connecting the conductor plate leads 24 and 25, the inductance of the metal electrode wiring on the semiconductor device will become a problem, so it cannot be made that long. Moreover, when a large current is to flow, the total area of the semiconductor device must be increased. In the end, the arrangement must be elongated in the vertical direction as shown in FIG. 5a. The package must be constructed such that the width of the conductor plate leads 24 and 25 in the direction perpendicular to the direction of current flow is wider than the distance between the conductor plate leads 24 and 25. For example, cathode,
Anode spacing 450μm, P ⁺ gate diffusion depth 4μm,
In SITh, in which a basic channel is formed in a stripe structure with a P ⁺ gate and a gate spacing of 1.5 μm, operation with an on-voltage of 1.0 to 1.2 V is achieved at a current density of 800 A/cm ² in the channel portion. If an SITh with an average current of 100 A is constructed, the chip size of the semiconductor device will be approximately 7 x 40 mm ² . That is, the conductor plate lead 24
The length of the semiconductor device in the direction connecting 25 and 25 is 7
mm, and the width of the device in the direction perpendicular to the so-called current flow direction is 40 mm. At this time, the distance between the conductor plate leads 24 and 25 is approximately 10 mm, and the width of the leads in the direction perpendicular to the current flow direction is approximately 40 mm.

Packages with such a configuration are usually mounted on wide conductor plates. Therefore, both the control electrode lead portion and the anode and cathode electrode portions have a transmission line-like configuration. In such a transmission line configuration, self-inductance has almost no effect, and mutual inductance has an effect.

A comparison of transmission lines between conventional linear electrode leads or narrowly designed plate electrode leads and the wide plate electrode lead of the present invention will be described below. The characteristic impedance Z ₀ when a wire with a diameter of 2r is placed in parallel as a wide conductor plate in the air and separated by a distance D is: It is. The characteristic impedance Z ₀ when a plate-shaped conductor with a width W is placed parallel to a wide conductor plate in the air and separated by a distance D is approximately Z ₀ ≒377/W/D {1+1.735 (W/D ) ^-0.836 ｝[Ω]
... is given by (5).

In the case of a diameter of r = 0.5 mm, when D = 5 mm and substituted into equation (4), Z ₀ ≒ 180 [Ω] ... (6) In the case of a narrow plate shape when W = 5 mm and D = 5 mm,
From formula (5), Z ₀ ≒138 [Ω] ...(7) In the case of a wide plate when W = 50 mm and D = 5 mm,
From formula (5), Z ₀ ≒30 [Ω] ...(8).

As is clear from equations (6), (7), and (8), the impedance of a wide plate electrode structure is considerably smaller than that of a linear electrode structure or a narrow plate electrode structure. When using high power devices, especially high current devices, the input impedance of the device,
Since both the output impedance becomes small, it is desirable that the impedance of the input/output circuit is also small. Particularly suitable for high current switching
When performing high-speed operation with a semiconductor device such as, it is necessary to attract a large gate current at turn-off, so the effect of reducing the impedance of all electrode leads according to the present invention is significant.

The embodiment shown in FIG. 5 is an example in which a semiconductor device substrate is mounted on a conductive base that itself serves as an electrode, but depending on the purpose of use, a metal base that fixes the semiconductor device and releases heat, and the semiconductor device In some cases, it may be more convenient to mount the substrate and electrically insulate it from the conductor plate that serves as the electrode. An example is shown in FIG. Furthermore, it may be easier to use if the electrodes are taken out at a 90° angle with other electrodes, for example, rather than taken out in the same direction. An example is shown in FIG. In the embodiment shown in FIGS. 6 and 7, the same symbols are used for the same parts as in the embodiment shown in FIG. In FIGS. 6 and 7, 29 is an insulating plate made of silicon carbide, beryllia, alumina, etc., and 30 is a metal base such as copper.

SITh, bipolar mode SIT (BSIT)
or any other bipolar transistor (BJT) or gate turn-off thyristor (GTO)
However, it is common to have a control electrode, such as a gate or base, in close proximity to a main carrier-injecting electrode, such as a source, emitter, or cathode. Therefore, in the exploded top view of the semiconductor device shown in FIG. 5a, the conductor plate leads 24 and 25 become leads connected to the control electrode and the carrier injection main electrode, respectively. The conductor base 23 is
This serves as the lead for the carrier extraction main electrode, such as the drain, collector, or anode of the device.

By the way, in normal high-power high-frequency control circuits, the circuit structure is often such that the carrier injection electrode is used as a common potential terminal. For example, if the conductor plate lead 25 is used as a carrier injection main electrode, it is possible to actually form a circuit even if the conductor plate lead 25 has a shape as shown in FIGS. 5, 6, and 7. However, it is also true that a considerable degree of ingenuity is required. If you think about the circuit configuration, the 8th
If the package is formed as shown in the figure or FIG. 9, the construction of the circuit becomes easy.

In FIG. 8, the basic structure is the same as in FIGS. 5, 6, and 7, but the circuit layout is simplified when the carrier injection main electrode is used as a common potential terminal. In FIG. 8, 31 is a metal such as copper, and 32 is also a metal plate such as copper. 33 is an insulating material such as ceramic. In a semiconductor device having such a structure, the carrier injection main electrode, which serves as a common potential terminal, can be simply bonded to a metal plate 32, such as a large-area copper plate, which serves as a common potential in the circuit, making the arrangement extremely simple. The control electrode drive circuit and the main inter-electrode circuit can be neatly arranged, for example, separated into left and right sides. Since everything can be constructed on a wide common potential copper plate, distributed constant circuit layout is facilitated. When a semiconductor device becomes a high-voltage device with several thousand volts,
If the conductor plate lead 24 serving as the control electrode and the conductor plate lead 23 serving as the carrier lead-out main electrode are bent vertically downward as shown in FIG. 8, the circuit arrangement will be easier.

In the example of Figure 8, the heat dissipation of the device is 22,2
This is mostly done only through 3, 29, and 30. Of course, it is also possible to dissipate heat from the other side. An example is shown in FIG. Figure 9 has a structure that is easy to use when the semiconductor device is not composed of multiple chips but is integrated. That is, the carrier injection main electrode of the semiconductor device is also connected to the conductor plate lead 35 via the molybdenum plate 34.
connected directly to. 36 and 37 are insulators such as ceramics. These insulators are also preferably made of beryllia or silicon carbide, which have high thermal conductivity. In the structure shown in Fig. 9, the carrier injection main electrode is directly connected to the conductor plate lead 35 which becomes the main electrode via a molybdenum plate or, if necessary, a copper plate, and there is no wire lead or ribbon lead, so the inductance is even smaller. As a result, it is extremely desirable. Since the inductance of the electrode serving as the common potential terminal has a negative feedback effect as described above, it is desirable that it be as small as possible.

The semiconductor device having the structure shown in FIG. 9 has an extremely small inductance and is suitable for high-speed, large-current switching, and is also excellent in terms of ease of circuit construction and heat dissipation.

In Figures 5, 6, 7, 8, and 9, the electrode lead structure, which is the main focus of the present invention, is shown for easy understanding.
For example, only the main parts are depicted, omitting details such as seals. An example of a specific shape of the embodiment shown in FIG. 5 is shown in FIG. 10. Note that since the current flowing through the conductor plate lead 24 serving as the control electrode is relatively smaller than the current through the conductor plate lead 25, the width of the conductor plate lead 24 may be made slightly narrower than the conductor plate lead 25.

In the semiconductor device of the present invention, the semiconductor device substrate is arranged in a substantially rectangular shape, whether it has a plurality of chips or an integrated structure, and the conductor plate leads extend along the long sides of the semiconductor device substrate. The end faces in the width direction are arranged, and the width of the conductor plate lead is made larger than the length of the long side of the rectangle, and the width of the conductor plate lead is wider than the interval between the opposing conductor plate leads. It is.

As an example showing the effect of the wide electrode according to the present invention, a 2000V, 10μF capacitor (60 x 100 x 150 mm ³
An example of an experiment carried out using a volume (with two terminals provided at the top) will be described below. A high frequency, large current switching circuit shown in FIG. 11 was constructed using a semiconductor device T having a structure based on the present invention. This semiconductor device can turn on and off a current of 200A. Since the experiment aims to observe current switching, R _L =0,
R _M =0.01Ω.

First, in the circuit of Figure 11, capacitor C _D
The above-mentioned high frequency capacitors were used for C and C _g , and these terminals were wired with 2 cm wide mesh wires, approximately 10 cm on each side and 20 cm in total. The current switching waveform obtained with this circuit configuration was as shown in FIG. 12a. That is, both the rise and fall were slow, and the result was that almost no current flowed. This result is due to the voltage drop caused by inductance in the wide lead wire connected to the capacitor. Next, I connected 10 small capacitors with leads on both sides in parallel to form a 300μF capacitor.
At the same time as the lead inductance decreases, the effective inductance further decreases because the current no longer flows concentrated in a narrow portion.
The current waveform at that time is as shown in FIG. 12b.

Current switching of 200A is achieved with a switching time of less than 0.15μsec. Furthermore, the capacitor section of C _D in Fig. 11 was devised to form the circuit shown in Fig. 11 in a circuit configuration that has almost no lead inductance, and a 7 x 40 mm ² capacitor was added in the package of the embodiment shown in Fig. 10. Chip-sized bipolar mode SIT
When high-frequency, large-current switching was performed using the semiconductor device of the present invention into which (BSIT) was inserted, a record-setting on-voltage of 0.7V, current of 100A, turn-on time of 0.1μsec, and turn-off time of 10ns was achieved. High-current, high-speed switching was realized. The waveform at this time is shown in FIG. Switching at 100A and 10nsec is a high-speed switching that has never been recorded before. The voltage amplitude at this time was 300V, and the load resistance R _L was 3Ω. This is a value that cannot be achieved with a semiconductor device in which the above-mentioned BSIT is enclosed in a conventional package.

As described above, the effect of dispersing the current and lowering the inductance according to the present invention is clearly seen in the results of this experiment. The results of this experiment also show that even if a semiconductor device capable of high-speed, high-speed switching is created, the method of configuring the circuit, including the package that encloses it, and the structure of the elements that make up the circuit, such as the capacitors and resistors used, are This also shows that unless the device is suitable for switching, it will not be possible to take advantage of the high-speed performance of semiconductor devices.

The semiconductor device of the present invention can switch large currents ranging from several tens of amperes to several thousand amperes at extremely high speeds if the circuit configuration is considered, and can efficiently control large amounts of power.Moreover, the operating frequency can easily be adjusted to an audible frequency. Because it can be brought into an area beyond the specified area, it does not cause discomfort to the operator, reduces switching loss, and does not require heavy loads unlike pressure-welding type packages, making it easy to handle and mounting design. The industrial value is extremely large.

[Brief explanation of the drawing]

Figure 1 is an exploded cross-sectional view of a device in which a semiconductor device is mounted on a conventional ceramic seal press-contact type package, and Figure 2 shows an example of a conventional ceramic case flat type package in which multiple semiconductor device substrates are mounted. , a is an exploded top view,
b is an exploded sectional view. Figure 3 shows n channel.
Figure 4, which symbolically shows SITh, is a lumped constant equivalent circuit when SITh is enclosed in a package. FIG. 5 is an exploded top view (a) and an exploded cross-sectional view (b) of an embodiment of a semiconductor device according to the present invention in which all electrode leads are made into plate shapes to reduce an inductance component. FIG. 6 is an exploded cross-sectional view of another embodiment of the present invention in which a metal base for fixing a semiconductor device and dissipating heat is electrically insulated from a conductor on which a semiconductor device substrate is mounted and serving as an electrode. FIG. 7 is an exploded sectional view of another embodiment of the present invention in which the directions of extraction are not the same as in FIG. 6, but are taken out at an angle of 90°. FIG. 8 shows still another embodiment configured to facilitate circuit arrangement when the carrier injection main electrode is used as a common potential terminal. FIG. 9 shows another embodiment of the present invention in which the carrier injection main electrode of the semiconductor device is directly connected to the conductor plate lead via a molybdenum plate.
FIG. 10 is a perspective view showing a specific example of the shape of the embodiment shown in FIG. FIG. 11 shows a circuit for examining the characteristics of the semiconductor device of the present invention, and FIG. 12a shows switching waveforms when the lead wires of capacitors C _D and C _g are made to have large inductance in the circuit shown in FIG. 11. Figure b shows the switching waveform when the lead wires of the capacitors C _D and C _g are configured to effectively reduce the inductance. FIG. 13 shows the switching waveform when a large current is switched by the circuit of FIG. 11 using the semiconductor device of the present invention in which a BSIT is inserted into the package of the embodiment of FIG. 10. be. 1, 11a, 11b, 21a, 21b, 21j
...Semiconductor device substrate, 2a, 2b, 22a,
22b, 22j, 34...Molybdenum plate, 3a,
3b... Metal conductor block, 4a, 4b... Flange, 5a, 5b... Welding plate, 6... Ceramic case, 7... Sealing pipe/control electrode extraction hole, 8... Control lead electrode, 12... ...Metal base, 13a, 13b...Electrode lead, 14...Ceramic case, 15a, 15b, 26a, 26
b... Bonding wire (or ribbon lead), 23... Conductor base, 24, 25... Conductor plate lead, 27... Insulator case, 29... Insulating plate, 30... Metal base, 31... Copper, etc. 32... Metal plate, 33, 36, 37... Insulator, 35... Conductor plate lead, 38... Seal plate.

Claims (1)

[Scope of Claims] 1. A first, a second, and a semiconductor device having first and second main surfaces, encapsulating a substantially rectangular three-terminal semiconductor device substrate, and electrically separated from each other by an insulator.
In a semiconductor device having a third lead electrode conductor, the first lead electrode conductor is connected to the second main surface of the semiconductor device substrate, and has a width sufficient to place at least the entire semiconductor device substrate. and the second and third lead electrode conductors are formed on the long sides of the rectangle of the first main surface of the rectangular semiconductor device substrate. It is a plate-shaped thing connected to the electrode part via a bonding wire, a ribbon lead, or a molybdenum plate, and the width dimension of at least one of the second and third lead electrode conductors is the second and third lead electrode conductor. A semiconductor device characterized in that the spacing between the lead electrode conductors is wider than that of the lead electrode conductor. 2. The semiconductor device according to claim 1, wherein the width of the second and third lead electrode conductors is approximately the same as or larger than the length of the long side of the rectangle of the semiconductor device substrate. . 3. The substantially rectangular three-terminal semiconductor device substrate is formed by arranging a plurality of semiconductor device substrates in a row, and the width dimension of the second and third lead electrode conductors is set in the row shape. The semiconductor device according to claim 1, characterized in that the dimension in the column direction of the plurality of semiconductor device substrates arranged is larger than that of the plurality of semiconductor device substrates arranged.