JP5691916B2 - Power converter - Google Patents

Description

  The present invention relates to a power conversion device capable of cooling heat generated by switching of a semiconductor element, and in particular, includes a first power module and a second power module, and each power module is connected to a semiconductor element via an insulating member. It is related with the power converter device provided with the cooling case joined together.

  In a hybrid vehicle or the like, power conversion is performed by a power conversion device (inverter device), and the power conversion device is configured to cool heat generated by switching of a semiconductor element. In recent years, such power conversion devices have been required to be smaller and lighter and to have higher output, and various devices have been proposed.

  For example, as shown in FIG. 41, the power converter described in Patent Document 1 below includes a first power module 210L, a second power module 210H, and a cooling jacket 290. The first power module 210L and the second power module 210H include a semiconductor element 220 that is an IGBT or a diode, a copper base 240 joined to the semiconductor element 220 via an insulating member 230, and a cooling case in which a refrigerant flows in or out. 250.

  The first power module 210L and the second power module 210H are assembled via a cooling jacket 290 in a direction orthogonal to the planar direction of the copper base 240, and the refrigerant is the cooling jacket 290 and the copper base 240. It flows in the refrigerant space RK between the fins 240a. Thus, the heat transferred from the semiconductor element 220 to the fins 240a is cooled by the flow of the refrigerant.

  Here, in the power converter 1V described in Patent Document 1 below, as shown in FIG. 42, an inflow structure RC1 into which the refrigerant flows and an outflow structure RC2 from which the refrigerant flows out are configured. 42 is a cross-sectional view of the power conversion device 1V shown in FIG. 41 in the ZZ direction.

  As shown in FIG. 42, in the inflow structure RC1, one end of the upper copper base 240 is assembled to the upper cooling jacket 290U via the O-ring 271 and one end of the lower copper base 240 is O It is assembled to the lower cooling jacket 290D via a ring 272. Similarly, in the outflow structure RC2, the other end of the upper copper base 240 is assembled to the upper cooling jacket 290U via the O-ring 271 and the other end of the lower copper base 240 is the O-ring. 272 is attached to the lower cooling jacket 290D. Thus, the strength of the copper base 240 is ensured by assembling the copper base 240 to the cooling jacket 290.

  Furthermore, as shown in FIG. 42, the power conversion device 1 </ b> V is configured such that the size of the upper (first power module 210 </ b> L) copper base 240 in the planar direction is the lower (second power module 210 </ b> H) copper base 240. The size of the refrigerant inflow portion RV1 and the size of the refrigerant outflow portion RV2 are ensured by making the size larger than the size in the planar direction. This is based on the following reason.

  If the size of the copper bases 240 in the plane direction is made equal while maintaining the height of the fins 240a in the height direction (vertical direction in FIG. 42), by assembling the cooling jackets 290U and 290D, The flow path of inflow part RV1 and outflow part RV2 becomes narrow, and a pressure loss becomes large. Further, if the fins 240a are increased in length in the height direction so that the copper bases 240 have the same size in the plane direction, the flow rate of the refrigerant at the root portion of the fins 240a decreases, It becomes easy to stay and cooling performance falls. In this way, the copper bases 240 are configured to have different sizes in the plane direction in order to appropriately secure the flow paths of the inflow portion RV1 and the outflow portion RV2 while preventing a decrease in cooling performance and an increase in pressure loss. Was.

JP 2005-259748 A

  However, the power converter 1V described above has the following problems. That is, the size of the power converter 1 </ b> V in the planar direction is increased due to the structure in which the cooling jackets 290 </ b> U and 290 </ b> D are separately assembled to the copper base 240 and the sizes of the copper bases 240 are different from each other. As a result, there is a problem that vehicle mountability deteriorates.

  Moreover, since it is necessary to seal between the copper base 240 and the cooling jackets 290U and 290D using two O-rings 271 and 272, there is also a problem that the sealing structure is not simple. Furthermore, it is necessary to firmly seal the two O-rings 271 and 272 with the copper base 240, respectively, and in order to ensure the strength of the copper base 240, the length of the copper base 240 in the thickness direction (vertical direction in FIG. 39). There is also a problem that the height in the height direction of the power conversion device 1V increases.

  The present invention has been made in order to solve the above-described problems, and provides a power conversion device that has a simple seal structure and can be reduced in size (size in the planar direction and length in the height direction). The purpose is to provide.

(1) A power conversion device according to the present invention is a power conversion device capable of cooling heat generated by switching of a semiconductor element, and includes a cooling case joined to the semiconductor element via an insulating member. The cooling case of each power module includes a heat dissipating part that extends in a planar shape, the insulating member is joined to the front surface, and fins are formed on the back surface. A water channel part extending in a direction perpendicular to the plane direction and flowing in and out of the refrigerant, and an edge part that surrounds the fin and forms a space through which the refrigerant flows, and the heat radiation part, the water channel part, and the edge part. The first power module and the second power module are assembled in a state where the fins face each other and the edges of the first power module and the second power module are sealed. It has been, that the edge of said the edge of the first power module second power module is welded to said edge, easily transferred heat by welding to the inside from its outer surface The heat buffer structure is formed, and the heat buffer structure is a slit .

(2) A power conversion device according to the present invention is a power conversion device capable of cooling heat generated by switching of a semiconductor element, and includes a cooling case joined to the semiconductor element via an insulating member. The cooling case of each power module includes a heat dissipating part that extends in a planar shape, the insulating member is joined to the front surface, and fins are formed on the back surface. A water channel part extending in a direction perpendicular to the plane direction and flowing in and out of the refrigerant, and an edge part that surrounds the fin and forms a space through which the refrigerant flows, and the heat radiation part, the water channel part, and the edge part. The first power module and the second power module are assembled in a state where the fins face each other and the edges of the first power module and the second power module are sealed. The edge of the first power module and the edge of the second power module are welded to each other, and it is difficult for the edge to transmit heat due to welding from the outer surface to the inside. The heat buffer structure is formed, and the heat buffer structure is a step junction between the edge of the first power module and the edge of the second power module.

In has been power conversion device according to (3) (1) or (2), wherein the first power module edge and the edge of the second power module are assembled via an O-ring It is preferable.

( 4 ) In the power conversion device described in any one of (1) to ( 3 ), the cooling case includes a header portion that continuously extends from the heat radiating portion in a planar shape and has no fins. A sealing member made of a thermosetting resin is formed on the surfaces of the heat radiating portion and the header portion so as to cover the semiconductor element, and the header portion protrudes from the surface and is sealed It is preferable that an engaging convex portion that engages with the member is formed.

( 5 ) In the power conversion device described in ( 4 ), it is preferable that the engagement protrusion is formed with a first protrusion that extends in a planar direction of the header portion and holds the sealing member. .

( 6 ) In the power conversion device described in any one of (1) to ( 3 ), the cooling case includes a header portion that continuously extends in a planar shape from the heat radiating portion and does not have the fins formed thereon. A sealing member made of a thermosetting resin is formed on the surfaces of the heat radiating portion and the header portion so as to cover the semiconductor element, and the heat radiating portion is recessed from the surface and is sealed It is preferable that an engaging recess to be engaged with the member is formed.

( 7 ) In the power conversion device described in ( 6 ), the engaging recess is formed with a second projecting piece projecting inward so as to close the engaging recess and holding the sealing member. Is preferred.

( 8 ) In the power conversion device described in any one of (1) to ( 3 ), a thermosetting resin sealing member is formed on the surface of the heat dissipation portion so as to cover the semiconductor element, The sealing member projects outward from the end portion of the cooling case, and a third projecting piece is formed at the end portion of the cooling case, extending in the planar direction of the heat radiating portion and holding the sealing member. Preferably it is.

( 9 ) In the power conversion device described in any one of (1) to ( 3 ), a thermosetting resin sealing member is formed on the surface of the heat dissipation portion so as to cover the semiconductor element, It is preferable that the 2nd sealing member of a thermoplastic resin is formed so that the whole sealing member and the whole edge part of a 1st power module and the said 2nd power module may be covered.

( 10 ) In the power converter described in any one of (1) to ( 9 ), the cooling case is formed by forging using an aluminum material in which aluminum is 99% or more by mass. Preferably there is.

( 11 ) In the power conversion device described in any one of (1) to ( 10 ), each of the first power module and the second power module includes a pair of diodes and transistors, and includes one motor. Preferably, an inverter circuit for driving is formed.

( 12 ) In the power conversion device described in any one of (1) to ( 11 ), the pair of the first power module and the second power module assembled in the direction orthogonal to the planar direction of the heat radiating unit. It is preferable that a plurality of layers are stacked.

( 13 ) In the power conversion device described in ( 12 ), the plurality of stacked first power modules and second power modules are assembled so as to be covered with a housing and a bottom cover. A first elastic member is interposed between the bottom and the first power module and the second power module positioned at the bottom of the stacked first power module and second power module, and the plurality It is preferable that a second elastic member is interposed between the first power module and the second power module positioned at the top of the stacked first power module and second power module and the housing. .

( 14 ) In the power converter described in ( 12 ) or ( 13 ), the pair of first power module and second power module assembled together, and the pair of first power module and first assembled It is preferable that a third elastic member is interposed between the two power modules.

The effect of the power converter device in this invention is demonstrated.
In said structure (1), the power converter device is comprised by two components, the 1st power module provided with the cooling case, and the 2nd power module. And the 1st power module and the 2nd power module are assembled in the state where a mutual fin opposed and a mutual edge was sealed. For this reason, the seal structure is simple. Furthermore, in the cooling case, the heat radiation part, the water channel part, and the edge part are integrally formed, and the length of the fin can be set appropriately while ensuring the size of the opening part of the water channel appropriately. Therefore, it is not necessary to increase the size of the cooling case (the size in the plane direction and the length in the height direction) more than necessary. In this way, the power conversion device can be configured by reducing the size of the cooling case.
In addition, by welding the edges, it is possible to sufficiently secure the bonding strength between the first power module and the second power module.
Further, for example, the heat buffer structure hardly causes defects due to welding, such as heat transmitted to the O-ring causing damage to the O-ring, and melting of the base material of the cooling case entering the assembly groove for assembling the O-ring. can do.
Moreover, since the thermal buffer structure is configured as a slit, the thermal buffer structure can be easily configured.

In said structure (2), the power converter device is comprised by two components, the 1st power module provided with the cooling case, and the 2nd power module. And the 1st power module and the 2nd power module are assembled in the state where a mutual fin opposed and a mutual edge was sealed. For this reason, the seal structure is simple. Furthermore, in the cooling case, the heat radiation part, the water channel part, and the edge part are integrally formed, and the length of the fin can be set appropriately while ensuring the size of the opening part of the water channel appropriately. Therefore, it is not necessary to increase the size of the cooling case (the size in the plane direction and the length in the height direction) more than necessary. In this way, the power conversion device can be configured by reducing the size of the cooling case.
In addition, by welding the edges, it is possible to sufficiently secure the bonding strength between the first power module and the second power module.
Further, for example, the heat buffer structure hardly causes defects due to welding, such as heat transmitted to the O-ring causing damage to the O-ring, and melting of the base material of the cooling case entering the assembly groove for assembling the O-ring. can do.
Moreover, since the thermal buffer structure is configured as a step junction between edges, the thermal buffer structure can be easily configured.

In the configuration (3), the liquid tight structure in which the refrigerant does not leak can be easily formed by assembling the edge of the first power module and the edge of the second power module using the O-ring.

In the configuration ( 4 ), the engagement convex portion protruding from the surface of the header portion is engaged with the sealing member, so that the bonding strength between the sealing member and the cooling case can be increased. Furthermore, since the fins are formed on the back surface of the heat radiating portion, the engaging projections are formed on the surface of the header portion, so that fluctuations in the thickness distribution in the heat radiating portion and the header portion can be reduced. That is, the difference between the thickness of the header portion and the thickness of the heat radiating portion is reduced by forming the engaging convex portion on the surface of the header portion where no fin is formed on the back surface. Therefore, it becomes easy to form the cooling case by forging.

In the above configuration ( 5 ), since the sealing resin inserts the first projecting piece formed on the engaging convex portion, the bonding strength between the sealing member and the cooling case can be further increased. Thus, it is possible to easily ensure the rigidity required for the cooling case.

In the above configuration ( 6 ), the engagement recess recessed from the surface of the heat radiating portion engages with the sealing resin, so that the bonding strength between the sealing member and the cooling case can be increased. Furthermore, since fins are formed on the back surface of the heat radiating portion and an engagement recess is formed on the surface of the heat radiating portion, fluctuations in the thickness distribution in the heat radiating portion and the header portion can be reduced. That is, an increase in the thickness of the heat radiating portion caused by forming the fin can be suppressed by forming the engaging recess, and the difference between the thickness of the header portion and the thickness of the heat radiating portion is reduced. Therefore, it becomes easy to form the cooling case by forging.

In the configuration ( 7 ), since the second projecting piece formed in the engaging recess holds the sealing resin, the bonding strength between the sealing resin and the cooling case can be further increased. Thus, it is possible to easily ensure the rigidity required for the cooling case.

In the configuration ( 8 ), since the sealing resin inserts the third projecting piece formed at the end of the cooling case, the bonding strength between the sealing resin and the cooling case can be further increased. Thus, it is possible to easily ensure the rigidity required for the cooling case.

In the above configuration ( 9 ), the second sealing member made of the thermoplastic resin covers the entire sealing member and the entire edge of the first power module and the second power. Bonding strength with the power module can be increased, and refrigerant leakage can be prevented.

In said structure ( 10 ), a desired shape is obtained by forging an aluminum material with high purity. Moreover, since the purity of the aluminum material is high, the heat transfer coefficient of the cooling case is improved and the cooling performance is improved. Furthermore, since the impurities contained in the aluminum material are extremely small, even when laser welding is performed, the impurities do not bump to form blowholes.

In the configuration ( 11 ), the semiconductor element disposed on the downstream side of the refrigerant in the low-side circuit has the highest temperature. Therefore, by observing the temperature sensor built in the semiconductor element, abnormal overheating of each of the six semiconductor elements can be prevented. In other words, the temperature sensor incorporated in the remaining semiconductor elements (semiconductor elements other than the semiconductor elements arranged on the downstream side of the refrigerant in the low-side circuit) can be omitted.

In the configuration ( 12 ), a pair of the first power module and the second power module, which are assembled in advance, are stacked in a direction perpendicular to the plane direction of the heat radiating portion, and therefore flow toward the fins. The seal of the refrigerant space and the seal of the water channel through which the refrigerant flows in or out can be made independent. For this reason, the applied pressure at the time of sealing a water channel can be made small. Therefore, the strength required for the cooling case is reduced, and the cooling case can be reduced in size.

In the configuration ( 13 ), the entire power conversion device (cooling case) can be reinforced by the housing, the bottom cover, the first elastic member, and the second elastic member, and leakage of the refrigerant can be prevented. Furthermore, the impact acting on the bottom lid and the housing can be absorbed by the first elastic member and the second elastic member, and the power converter can be made to have a structure resistant to the impact.

In the above configuration ( 14 ), the entire power conversion device can be reinforced by filling the gap generated between the assembled power modules with the third elastic member.

It is a fragmentary sectional view when the right half of a power converter is fractured. (A) It is a top view of a 1st power module. (B) It is a fragmentary sectional view when the right half of a 1st power module is fractured | ruptured. (C) It is a rear view of a 1st power module. (A) It is a top view of a 2nd power module. (B) It is a fragmentary sectional view when the right half of the 2nd power module is fractured. (C) It is a rear view of a 2nd power module. It is the figure which showed the inverter circuit of the power converter device. (A) It is the figure which showed the state before a semiconductor element was soldered. (B) It is the figure which showed the state after the semiconductor element was soldered. (A) It is the figure which showed the state by which the insulating member was arrange | positioned on the upper surface of a cooling case. (B) It is the figure which showed the state by which an insulation state is thermocompression-bonded on the upper surface of a cooling case. It is the figure which showed the state by which the semiconductor element unit and the cooling case were arrange | positioned in the transfer mold molding machine. It is the figure which showed the state by which resin was poured between the upper metal mold | die and the lower metal mold | die. It is the figure which showed the state from which the molding was taken out from the transfer mold molding machine. It is the figure which showed the part to which a mutual edge part is welded. (A) It is sectional drawing along the XX line of FIG. (B) It is a figure equivalent to FIG. 11 (A) in case a heat buffer structure is the thick part of an edge. FIG. 11 (C) is a view corresponding to FIG. 11 (A) in the case where the heat buffering structure is a step junction between edges. It is a fragmentary sectional view when the right half of the power converter device constituted as a comparative product is fractured. It is an exploded view of the power converter device shown in FIG. It is the figure equivalent to FIG. 12 which showed the power converter device which removed the fin space | interval wall shown in FIG. It is the figure equivalent to FIG. 12 which showed the power converter device which removed the fin space | interval wall shown in FIG. 12, and enlarged the length of the pin fin. It is the figure equivalent to FIG. 12 which showed the power converter device which made small the length of the height direction of the cooling case shown in FIG. (A) It is a rear view of the power module of this embodiment. (B) It is a side view of the thermal radiation part and sealing resin which were shown to FIG. 17 (A). (C) It is the figure which showed the stress distribution which acts on the adhesion interface of the thermal radiation part shown in FIG.17 (B), and sealing resin. (A) It is a rear view of a power module at the time of forming a straight fin in the back surface of a thermal radiation part. (B) It is a side view of the thermal radiation part and sealing resin which were shown to FIG. 18 (A). (C) It is the figure which showed the stress distribution which acts on the adhesion interface of the thermal radiation part shown in FIG.18 (B), and sealing resin. It is the table | surface which compared the result of the shape, cooling performance, laser welding, and ultrasonic flaw detection, when cooling case was shape | molded by changing the purity of aluminum material and the shaping | molding method. It is the figure which showed the state which has test | inspected the ultrasonic flaw detection with respect to the power module. It is sectional drawing to which the power module in 2nd Embodiment was expanded partially. (A) It is the perspective view which showed the 1st example of the shape of the rib shown in FIG. (B) It is the perspective view which showed the 2nd example of the shape of the rib shown in FIG. (A) It is the figure which showed the state before a rib was pressed with a pressing pin. (B) It is the figure which showed the state by which the rib was pressed with the press pin. It is sectional drawing to which the power module in 3rd Embodiment was expanded partially. It is the perspective view which showed the 1st example of the shape of the ditch | groove shown in FIG. (B) It is the perspective view which showed the 2nd example of the shape of the ditch | groove shown in FIG. (A) It is the figure which showed the state before a ditch | groove was pressed with a pressing pin. (B) It is the figure which showed the state by which the ditch | groove was pressed with the press pin. It is a fragmentary sectional view when the right half of the power converter of a 4th embodiment is fractured. It is an exploded view of the power converter device shown in FIG. It is a fragmentary sectional view when the right half of the power converter device constituted as a comparative product is fractured. It is an exploded view of the power converter device shown in FIG. (A) It is the schematic which showed the pressurization area of the O-ring in the power converter device shown in FIG. (B) It is the schematic which showed the pressurization area of the O-ring in the power converter device shown in FIG. It is the figure which expanded a part of power converter shown in FIG. It is the figure which showed the state with which the control terminal and control board of the power converter device shown in FIG. 27 were connected. It is sectional drawing along the YY line shown in FIG. It is the figure which showed the state which pinched | interposed the edge parts of the power module with the clip in the power converter device of deformation | transformation embodiment. In the power converter device of deformation | transformation embodiment, it is the figure which showed the state cast-in with sealing resin so that the whole edge part and whole sealing resin of a power module might be covered. In the rib of deformation | transformation embodiment, (A) It is the figure which showed the state before a rib is pressed with a pressing pin. (B) It is the figure which showed the state by which the rib was pressed with the press pin. It is the figure which showed the state in which the rectification rib is formed in the back surface of a header part in the power module of deformation | transformation embodiment. The figure which showed the state in which the elastic member was interposed between sealing resin and a housing, and the elastic member was interposed between sealing resin and a housing in the power converter device of deformation | transformation embodiment. It is. (A) In the power converter device of deformation | transformation embodiment, it is the figure which showed the state which sealing resin has protruded outside the edge part of the cooling case. (B) It is sectional drawing along the JJ line of FIG. 40 (A). It is sectional drawing of the conventional power converter device. It is sectional drawing along the ZZ line | wire shown in FIG.

<First Embodiment>
A power converter according to the present invention will be described below with reference to the drawings. FIG. 1 is a partial cross-sectional view when the right half of the power converter 1 is broken. This power conversion device 1 is composed of two parts, a low-side power module 10L (hereinafter simply referred to as “power module 10L”) and a high-side power module 10H (hereinafter simply referred to as “power module 10H”). Has been. The power module 10L corresponds to the first power module of the present invention, and the power module 10H corresponds to the second power module of the present invention.

  The power module 10 </ b> L has a symmetrical structure in the left-right direction in FIG. 1, and includes a semiconductor element 20, an insulating member 30, a sealing resin 40, and a cooling case 50. Here, FIG. 2 (A) is a plan view of the power module 10L, FIG. 2 (B) is a partial sectional view when the right half of the power module 10L is broken, and FIG. It is a rear view of power module 10L.

  The semiconductor element 20 is an electronic component that constitutes an inverter circuit, and is a heating element that generates heat by switching. As shown in FIGS. 1 and 2, the lower surface of the semiconductor element 20 is bonded to the upper surface of the collector electrode 21 by soldering, and the upper surface of the semiconductor element 20 is bonded to the lower surface of the emitter electrode 22 by soldering. . As shown in FIGS. 2A and 2C, the semiconductor element 20 is connected to the control terminal 71, and the emitter electrode 22 is connected to the power terminal 72. The semiconductor element 20 is specifically an IGBT 20A and a diode 20B, and this power module 10L is mounted with three pairs of IGBT 20A and diode 20B.

  The insulating member 30 is for electrically insulating the semiconductor element 20 and the cooling case 50. Specifically, the insulating member 30 is a sheet-like epoxy resin in which an insulating inorganic filler having a good thermal conductivity is blended to improve the thermal conductivity. The upper surface of the insulating member 30 is thermocompression bonded to the lower surface of the collector electrode 21, and the lower surface of the insulating member 30 is thermocompression bonded to the upper surface of the cooling case 50.

  The sealing resin 40 is for reducing external force acting on the semiconductor element 20. The sealing resin 40 is, for example, a thermosetting epoxy resin. This sealing resin 40 corresponds to the sealing member of the present invention. The sealing resin 40 is formed so as to cover the insulating member 30, the collector electrode 21, the semiconductor element 20, and the emitter electrode 22 from the upper surface of the cooling case 50.

  The cooling case 50 is a case for the refrigerant 60 to flow, and has a substantially rectangular parallelepiped shape with the lower side opened as shown in FIG. The cooling case 50 is made of aluminum having good thermal conductivity, and has a heat radiating portion 51, a header portion 52, an inflow water channel portion 53 </ b> A and an outflow water channel portion 53 </ b> B, and an edge 54.

  The heat dissipating part 51 extends in a planar shape, the insulating member 30 is bonded to the front surface (upper surface) 51a, and a plurality of pin fins 51c are formed on the rear surface (lower surface) 51b. Each pin fin 51c is formed in a columnar shape and is arranged in a lattice shape. The header portion 52 continuously extends in a planar shape from the heat radiating portion 51, and no pin fins 51 c are formed on the header portion 52.

  The inflow water channel portion 53 </ b> A is a portion into which the refrigerant 60 flows, is formed in a stepped cylindrical shape, and extends in a direction orthogonal to the planar direction of the heat radiation portion 51. An O-ring 73 is assembled to the step portion 53a of the inflow water channel portion 53A. The outflow water channel portion 53 </ b> B is a portion through which the refrigerant 60 flows out, is formed in a stepped cylindrical shape, and extends in a direction orthogonal to the planar direction of the heat radiation portion 51. An O-ring 74 is assembled to the step portion 53b of the outflow water channel portion 53B.

  The edge portion 54 is an annular frame, surrounds the pin fin 51c, and forms a refrigerant space RK in which the refrigerant 60 flows. The refrigerant 60 flowing through the refrigerant space RK contacts the pin fins 51c, so that the heat transmitted from the semiconductor element 20 to the pin fins 51c is efficiently cooled. Further, as shown in FIG. 2 (C), the edge portion 54 has an annular assembly groove 54a for assembling the O-ring 75, and as shown in FIGS. 2 (A) and 2 (C), The four corners have four insertion holes 54b for inserting the fixing pins 76 (see FIG. 1).

  The heat radiation part 51, the header part 52, the inflow water channel part 53A, the outflow water channel part 53B, and the edge part 54 are integrally formed by forging. Here, FIG. 3 (A) is a plan view of the power module 10H, FIG. 3 (B) is a partial cross-sectional view when the right half of the power module 10H is broken, and FIG. It is a rear view of the power module 10H.

  As shown in FIGS. 3A, 3B, and 3C, the power module 10H includes the semiconductor element 20, the insulating member 30, the sealing resin 40, the cooling case 50 (the heat radiating portion 51) as in the power module 10L described above. , Header portion 52, inflow water channel portion 53A, outflow water channel portion 53B, edge portion 54). For this reason, about the same structure, the same code | symbol is attached | subjected and the description is abbreviate | omitted.

  As shown in FIG. 3B, the shapes of the inflow water channel portion 53A and the outflow water channel portion 53B of the power module 10H are the shapes of the inflow water channel portion 53A and the outflow water channel portion 53B of the power module 10L (see FIG. 2B). ) Is a little different. The inflow water channel 53A and the outflow water channel 53B of the power module 10H have a stepped cylindrical fitting hole KG so that the inflow water channel 53A and the outflow water channel 53B of the power module 10L can be fitted therein. Yes. Note that the inflow water channel portion 53A and the outflow water channel portion 53B of the power module 10L are not assembled with O-rings.

  The power module 10L and the power module 10H are assembled so that the pin fins 51c face each other. Specifically, as shown in FIG. 1, the O-ring 75 is assembled in the assembly groove 54 a of each edge 54, and the fixing pin 76 is inserted from the upper side of the power module 10 </ b> L into the four edges 54. The tip of the fixing pin 76 is crimped through the hole 54b.

  Thus, the power module 10L and the power module 10H are assembled in a state where the pin fins 51c abut against each other and the edge portions 54 are sealed. Further, the edge portions 54 of the power modules 10L and 10H are assembled with each other using the O-ring 75, so that a liquid-tight structure in which the refrigerant 60 does not leak is easily formed. Then, the refrigerant 60 flows in from the inflow water channel portion 53A, flows in the refrigerant space RK in the direction indicated by the arrow in FIG. 1 (left direction in FIG. 1), and flows out to the outflow water channel portion 53B. In addition, although the front-end | tip of each pin fin 51c is facing in the state contact | abutted, you may oppose in the state which is not contact | abutting.

  Here, FIG. 4 is a diagram illustrating an inverter circuit of the power conversion device 1. In FIG. 4, the low side circuit which the power module 10L comprises is shown as the dashed-dotted line LS. In the low-side circuit LS, three pairs of IGBTs 20A and diodes 20B, which are the semiconductor elements 20, are arranged in parallel. In FIG. 4, the IGBT 20A and the diode 20B arranged on the upstream side (right side in FIG. 1), the middle stream side (center in FIG. 1), and the downstream side (left side in FIG. 1) of the refrigerant 60 flowing in the refrigerant space RK are sequentially arranged. It will be shown as IGBT 20A1, diode 20B1, IGBT 20A2, diode 20B2, IGBT 20A3 and diode 20B3.

  Further, the high-side circuit that the power module 10H configures is shown as a one-dot chain line HS. In the high side circuit HS, three pairs of IGBTs 20A and diodes 20B are arranged in parallel. In FIG. 4, the IGBT 20 </ b> A and the diode 20 </ b> B arranged on the upstream side (right side in FIG. 1), the middle stream side (center in FIG. 1), and the downstream side (left side in FIG. 1) of the refrigerant 60 flowing in the refrigerant space RK are sequentially arranged. And diode 20B1, IGBT 20A2, diode 20B2, IGBT 20A3 and diode 20B3.

  Between the IGBT 20A1 and the diode 20B1 (hereinafter referred to as “upstream semiconductor element LS1”) of the low side circuit LS and the IGBT 20A1 and the diode 20B1 (hereinafter referred to as “upstream semiconductor element HS1”) of the high side circuit HS. In addition, a W-phase electrode W1 is connected. Further, between the IGBT 20A2 and the diode 20B2 (hereinafter referred to as “middle stream side semiconductor element LS2”) of the low side circuit LS and the IGBT 20A2 and the diode 20B2 (hereinafter referred to as “middle stream side semiconductor element HS2”) of the high side circuit HS. In addition, a V-phase electrode V1 is connected.

  Further, between the IGBT 20A3 and the diode 20B3 (hereinafter referred to as “downstream semiconductor element LS3”) of the low side circuit LS and the IGBT 20A3 and the diode 20B3 (hereinafter referred to as “downstream semiconductor element HS3”) of the high side circuit HS. In addition, a U-phase electrode U1 is connected. These U-phase electrode U1, V-phase electrode V1, and W-phase electrode W1 are connected to one three-phase AC motor (not shown). Thus, the power conversion device 1 on which the six sets of IGBT 20A and diode 20B are mounted constitutes an inverter circuit that drives one three-phase AC motor.

  The effect of the power converter device 1 which comprises an inverter circuit as mentioned above is demonstrated. In the low side circuit LS, the upstream side semiconductor element LS1, the midstream side semiconductor element LS2, and the downstream side semiconductor element LS3 drive the same three-phase AC motor, and thus the same load acts. For this reason, all the heat generated by the energization current is the same. However, since the refrigerant 60 is gradually warmed by the heat generated by the semiconductor elements LS1, LS2, and LS3 from the upstream side to the downstream side of the refrigerant space RK, the downstream semiconductor element LS3 is more refrigerant than the upstream semiconductor element LS1. It is difficult to be cooled by 60. That is, the downstream semiconductor element LS3 is warmer than the upstream semiconductor element LS1.

  Therefore, in the low-side circuit LS, by observing the temperature sensor incorporated in the downstream semiconductor element LS3, the semiconductor elements LS1, LS2 are observed without observing the temperature sensor incorporated in each semiconductor element LS1, LS2, LS3. , LS3 abnormal overheating can be prevented. In other words, the temperature sensor built in the upstream side semiconductor element LS1 and the midstream side semiconductor element LS2 can be omitted.

  The above-described operation and effect cannot be obtained when the low-side circuit LS includes a semiconductor element that drives a different three-phase AC motor. Further, a three-phase AC motor can be constructed by configuring three stages of power modules (2-in-1 type power modules) including one semiconductor element constituting the low-side circuit and one semiconductor element constituting the high-side circuit. In the case of driving, the effect of cooling each semiconductor element becomes almost equal. On the other hand, the influence of the difference in the amount of heat generated due to the variation in characteristics of each semiconductor element and the like becomes relatively large, and the regularity that the temperature of the semiconductor element at any part of the structure is always high is lost. Can't get.

  That is, the above-described operation and effect can be obtained because the low-side circuit LS is composed of three semiconductor elements LS1, LS2, and LS3 (three sets of IGBT 20A and diode 20B), and one power module is composed of these three semiconductors. Only when the elements LS1, LS2 and LS3 are included (when the power module is a 3 in 1 type).

  The operational effects in the low side circuit LS described above can be similarly applied to the operational effects in the high side circuit HS. Here, when comparing the low-side circuit LS that drives one three-phase AC motor with the high-side circuit HS, the gate voltage that acts on the low-side circuit LS is always slightly higher due to the characteristics of the gate circuit. The current flowing through is slightly larger.

  For this reason, the temperature of the downstream semiconductor element LS3 of the low side circuit LS is higher than the temperature of the downstream semiconductor element HS3 of the high side circuit HS. Therefore, by observing the temperature sensor built in the downstream semiconductor element LS3 of the low side circuit LS, abnormal heating of each of the six semiconductor elements LS1, LS2, LS3, HS1, HS2, HS3 can be prevented. In other words, the temperature sensors built in the remaining semiconductor elements LS1, LS2, HS1, HS2, HS3 can be omitted.

  Next, a procedure for assembling each member constituting the power conversion device 1 will be described with reference to FIGS. 5A is a diagram illustrating a state before the semiconductor element 20 is soldered, and FIG. 5B is a diagram illustrating a state after the semiconductor element 20 is soldered. As shown in FIG. 5A, first, the solder foil 23 is disposed between the semiconductor element 20 and the collector electrode 21, and the solder foil 24 is disposed between the semiconductor element 20 and the emitter electrode 22. Next, by melting the solder foils 23 and 24 in a furnace, the lower surface of the semiconductor element 20 and the upper surface of the collector electrode 21 are soldered as shown in FIG. And the lower surface of the emitter electrode 22 are soldered. Thus, the semiconductor element unit 20U is formed.

  In addition, the cooling case 50 having the heat radiating portion 51, the header portion 52, the inflow water channel portion 53A, the outflow water channel portion 53B, and the edge portion 54 is previously formed by forging. Here, FIG. 6A is a diagram illustrating a state in which the insulating member 30 is disposed on the upper surface of the cooling case 50. FIG. 6B is a diagram illustrating a state in which the insulating member 30 is thermocompression bonded to the upper surface of the cooling case 50. As shown in FIG. 6A, after the insulating member 30 is arranged, the insulating member 30 is pressed toward the cooling case 50 using a heat press HP as shown in FIG. 6B. Thereby, the insulating member 30 is thermocompression bonded, and the lower surface of the insulating member 30 and the upper surface of the cooling case 50 are temporarily joined.

  7, 8, and 9 are diagrams illustrating a process of molding the sealing resin 40 using the transfer mold molding machine TF. First, as shown in FIG. 7, the semiconductor element unit 20 </ b> U and the cooling case 50 are disposed in the transfer mold forming machine TF. At this time, the lower surface of the collector electrode 21 is disposed on the upper surface of the insulating member 30.

  Next, as shown in FIG. 8, the cooling case 50 is sandwiched between the upper mold TF1 and the lower mold TF2 of the transfer mold molding machine TF to form a space ZR into which resin is poured. Then, the liquid sealing resin 40 is poured into the space ZR, and the insulating member 30 is heated at the temperature of the sealing resin 40 and is pressurized by the injection pressure of the sealing resin 40. Furthermore, the insulating member 30 is also heated by the temperatures of the upper mold TF1 and the lower mold TF2. Thereafter, the crosslinking reaction proceeds in the sealing resin 40 made of epoxy resin, and the lower surface of the insulating member 30 and the upper surface of the cooling case 50 are firmly bonded, and the lower surface of the collector electrode 21 and the upper surface of the insulating member 30 are bonded to each other. Adhere firmly.

  Finally, as shown in FIG. 9, the molded product is taken out from the transfer molding machine TF. Thus, the power module 10L in which the semiconductor element unit 20U and the cooling case 50 are joined is molded. The power module 10H is also formed in the same manner as the power module 10L. Then, as described above, the power conversion device 1 is manufactured by assembling the power module 10L and the power module 10H in a state where the pin fins 51c face each other and the edges 54 are sealed.

  By the way, as described above, the power module 10L and the power module 10H are joined by the fixing pins 76 being inserted into the four insertion holes 54b of the respective edge portions 54 and the tips of the fixing pins 76 being caulked. Yes. Usually, the joining strength can be sufficiently obtained by joining with the fixing pin 76. However, due to the physique of the power modules 10L and 10H, the pressure of the refrigerant 60, and the reaction force of the O-ring 75, the joining strength is not sufficient only by joining with the fixing pin 76, and the coolant 60 may leak from the edge portion 54. . For this reason, in this embodiment, the edge part 54 of the power module 10L and the edge part 54 of the power module 10H are welded.

  FIG. 10 is a diagram showing a portion where the edge portions 54 are welded. In the present embodiment, the edge portions 54 are joined by laser welding at a portion IS shown by a one-dot chain line in FIG. Thereby, the joining of the power modules 10L and 10H can be reinforced, and a sufficiently large joining strength can be obtained. In this embodiment, laser welding is exemplified, but the welding method may be other than laser welding (for example, MIG, MAG, TIG, plasma welding, etc.). Here, FIG. 11A is a cross-sectional view taken along line XX in FIG.

  In the present embodiment, as shown in FIG. 11A, a slit SR is formed between the outer surface 54c of the edge 54 of the power modules 10L and 10H and the assembly groove 54a of the power modules 10L and 10H. Has been. The slit SR makes it difficult to transfer heat from welding from the outer surface 54c of the edge portion 54 toward the inside. The slit SR corresponds to the heat buffer structure of the present invention.

  The slit SR is formed to be annular around the entire circumference of the edge portion 54. However, a plurality of slits SR may be partially formed in the edge portion 54 corresponding to the portion to be laser-welded (the portion IS indicated by the one-dot chain line in FIG. 10). An air heat insulating layer is formed in the slit SR, and the heat transfer coefficient of the heat insulating layer is sufficiently smaller than the heat transfer coefficient of the cooling case 50 made of aluminum. For this reason, it is difficult for the heat rays of the laser to be transmitted to the O-ring 75. Further, it is possible to prevent the molten base material of the cooling case 50 from entering the assembly groove 54a. In this way, heat due to welding is hardly transmitted from the outer surface 54c of the edge portion 54 to the inside, and damage to the O-ring 75 can be prevented.

  Here, in this embodiment, as shown in FIG. 11A, the thermal buffer structure is the slit SR, but, as in the modified embodiment shown in FIG. 54 thick part NA may be sufficient. The thick portion NA is configured such that the distance N1 between the outer surface 54c of the edge 54 and the assembly groove 54a is 1.5 mm or more. The thick portion NA increases the heat capacity of the edge portion 54 and makes it difficult for heat to be transferred, and a thermal diffusion effect is produced. In this way, heat due to welding is hardly transmitted from the outer surface 54c of the edge portion 54 to the inside, and damage to the O-ring 75 can be prevented.

  Further, as in the modified embodiment shown in FIG. 11C, the thermal buffer structure may be a step junction (maze structure) DS of the edge portion 54. This step joint DS is configured by joining the edges 54 in a crank shape between the outer surface 54c and the assembly groove 54a. That is, the edges 54 are not joined in a planar shape. This makes it difficult for the heat rays of the laser to be transmitted to the O-ring 75 and prevents the molten base material of the cooling case 50 from entering the assembly groove 54a. In this way, heat due to welding is hardly transmitted from the outer surface 54c of the edge portion 54 to the inside, and damage to the O-ring 75 can be prevented.

  Next, before describing the advantages of the power conversion device 1 of the present embodiment, a power conversion device 1W configured as a comparative product will be described. FIG. 12 is a partial cross-sectional view when the right half of the power converter 1W is broken. FIG. 13 is an exploded view of power converter 1W shown in FIG. As shown in FIG. 12, the power conversion device 1W includes a low-side power card 110L (hereinafter simply referred to as “power card 110L”) and a high-side power card 110H (hereinafter simply referred to as “power card 110H”). )) And the cooling case 190.

  The power card 110L includes a semiconductor element 120, an insulating member 130, a sealing resin 140, and a heat sink 150. Since the semiconductor element 120, the insulating member 130, and the sealing resin 140 are the same as the semiconductor element 20, the insulating member 30, and the sealing resin 40 described above, description thereof is omitted. The heat sink 150 is formed in a flat plate shape and has a plurality of pin fins 150a on the back surface. Since the power card 110H has the same configuration as the power card 110L, the description thereof is omitted.

  As shown in FIGS. 12 and 13, the cooling case 190 is a case for forming a flow path through which the refrigerant 160 flows. The cooling case 190 includes a substantially rectangular parallelepiped main body portion 191, a substantially cylindrical inflow water passage portion 192 </ b> A and an outflow water passage portion 192 </ b> B extending in a direction orthogonal to the planar direction of the main body portion 191, and the main body portion 191 and the inflow water passage portion. 192A and the outflow water channel part 192B, and the header part 193 arrange | positioned. The main body 191 has an opening 191a for assembling the pin fin 150a of the power card 110L on the upper side, and an opening 191b for assembling the pin fin 150a of the power card 110H on the lower side.

  The main body portion 191 and the header portion 193 are formed with a rectangular parallelepiped flow passage hole RA through which the refrigerant 160 flows. The header portion 193 has a concave groove 193a for assembling the O-ring 194 on the upper side, and a concave groove 193b for assembling the O-ring 195 on the lower side. The O-rings 194 and 195 are for sealing the ends of the heat sinks 150 of the power cards 110L and 110H. The inflow water channel portion 192A and the outflow water channel portion 192B have step portions 192a and 192b for assembling the O-ring 196, respectively.

  By the way, in the power converter device 1 </ b> W illustrated in FIG. 12, the fin interval wall 197 is provided in the flow path hole RA of the cooling case 190. This is based on the following reason. First, as shown in FIG. 14, consider the power conversion device 1 </ b> X in which the fin interval wall 197 is removed from the power conversion device 1 </ b> W shown in FIG. 12. In this power conversion device 1X, the refrigerant 160 easily flows in the central portion CO indicated by the one-dot chain line in FIG. 14, and does not easily flow in the vicinity FS of the pin fin 150a indicated by the one-dot chain line in FIG. For this reason, the refrigerant | coolant 160 stagnates in the vicinity FS of the pin fin 150a, heat exchange is not appropriately performed between the pin fin 150a and the refrigerant | coolant 160, and cooling performance falls.

  Next, as shown in FIG. 15, a power converter 1Y in which the fin interval wall 197 is removed from the power converter 1W shown in FIG. 12 and the length of the pin fin 150a is increased will be considered. In this power conversion device 1Y, the root portion of the pin fin 150a is closer to the semiconductor element 120, which is a heating element, than the tip portion of the pin fin 150a, and thus becomes high temperature. In other words, the root portion of the pin fin 150a is a portion having better thermal efficiency than the tip portion of the pin fin 150a. However, since the length of the pin fin 150a is large, the refrigerant 160 easily flows in the vicinity SB near the tip end portion of the pin fin 150a shown by the one-dot chain line in FIG. 15, and the vicinity of the root portion of the pin fin 150a shown by the one-dot chain line in FIG. It is hard to flow with NB. For this reason, heat exchange is not appropriately performed between the root portion of the pin fin 150a with good thermal efficiency and the refrigerant 160, and cooling performance is deteriorated.

  Next, as shown in FIG. 16, consider the power converter 1Z in which the dimension in the thickness direction of the cooling case 190 of the power converter 1W shown in FIG. 12 is reduced. In this power conversion device 1Z, the opening KK of the water channel indicated by the one-dot chain line in FIG. 16 is narrowed. For this reason, the flow resistance of the refrigerant 160 is very large at the opening portion of the water channel, and the pressure loss becomes extremely large.

  The reason why the opening KK of the power conversion device 1W shown in FIG. 16 is narrow is that the heat sink 150 and the cooling case 190 are separated. That is, in order to combine the heat sink 150 and the cooling case 190 so that the refrigerant does not leak, it is necessary to provide the O-rings 194 and 195 between the heat sink 150 and the cooling case 190 (concave grooves 193a and 193b). Because there is.

  From the above, in the power conversion device 1W shown in FIG. 12, by providing the fin interval wall 197, an appropriate size of the opening portion KK of the water channel can be secured, and the length of the pin fin 150a is appropriately set. Thus, heat exchange can be appropriately performed between the pin fin 150a and the refrigerant 160. However, even this power converter 1W has the following problems.

  First, as shown in FIG. 12, since the fin interval wall 197 is provided, the dimension in the height direction of the cooling case 190 (power converter 1W) is increased, and the weight of the cooling case 190 (power converter 1W) is increased. growing. For this reason, the cost of the power converter device 1W increases. Moreover, in this power converter device 1W, in order to assemble the power card 110L and the power card 110H on the upper side and the lower side of the cooling case 190, the plane portion indicated by the one-dot chain line in FIG. It is necessary to form HB. When the cooling case 190 having such a pair of upper and lower planar portions HB is formed, it cannot be formed by a single movement of the mold. For this reason, it is necessary to join by welding etc. after shape | molding the workpiece previously divided | segmented up and down. That is, the cooling case 190 has poor moldability.

  On the other hand, the power conversion device 1 of the present embodiment has the following advantages. As shown in FIG. 1, in the power conversion device 1, the size of the opening KK of the water channel can be appropriately ensured without providing the fin interval wall 197 (see FIG. 12) as described above, and the pin fin 150 a Can be set appropriately. For this reason, since the dimension in the height direction of the cooling case 50 (power converter 1) can be prevented by not providing the fin interval wall 197, the opening portion KK of the water channel becomes wide, and the pressure loss increases. Can be prevented.

  Further, the cooling case 50 is formed by integrally forming a heat radiating portion 51, a header portion 52, water channel portions 53A, 53B, and an edge portion 54, and an O-ring 194 like the power conversion device 1W shown in FIG. , 195 need not be provided between the heat sink 150 and the cooling case 190. For this reason, in the power converter device 1 shown in FIG. 1, compared with the power converter device 1W shown in FIG. 12, the opening part KK of a wide water channel can be formed, and the increase in the flow resistance of the refrigerant | coolant 60 can be suppressed. And increase in pressure loss can be suppressed.

  Further, the length of the pin fin 51c is not increased more than necessary, and heat exchange can be appropriately performed between the root portion of the pin fin 51c with good thermal efficiency and the refrigerant 60. In addition, in the cooling case 50 of each power module 10L, 10H, since there is no pair of upper and lower plane portions HB as shown by the one-dot chain line in FIG. 13, the cooling case 50 can be formed by moving the mold once. it can. That is, the cooling case 50 has good moldability.

  Next, the effect that the pin fins 51c are formed in the heat radiation part 51 of the cooling case 50 of the present embodiment will be described with reference to FIGS. FIG. 17A is a rear view of the power module 10 of the present embodiment, and FIG. 17B is a side view of the heat radiating portion 51 and the sealing resin 40 shown in FIG. FIG. 17C shows a stress distribution acting on the adhesion interface KM1 between the heat radiation part 51 and the sealing resin 40 shown in FIG.

  On the other hand, FIG. 18A is a rear view of the power module 10Z in the case where the straight fin SF is formed on the back surface 51b of the heat radiating portion 51, and FIG. 18B is shown in FIG. FIG. 4 is a side view of a heat dissipation part 51 and a sealing resin 40. FIG. 18C shows a stress distribution acting on the adhesion interface KM2 between the heat radiation part 51 and the sealing resin 40 shown in FIG.

By the way, the linear expansion coefficient of aluminum, which is the material of the heat radiation portion 51, is 23 × 10 ⁻⁶ / K, and the linear expansion coefficient of the epoxy resin, which is the material of the sealing resin 40, is 14 × 10 ⁻⁶ / K. There is a difference in heat shrinkage between the heat radiation part 51 and the sealing resin 40. For this reason, stress may generate | occur | produce in the adhesion interface KM1 and KM2 of the thermal radiation part 51 and the sealing resin 40 by the temperature change at the time of shaping | molding or a market, and there exists a possibility that the sealing resin 40 may peel from the thermal radiation part 51.

  Accordingly, when comparing FIG. 17C and FIG. 18C, the stress acting on the adhesion interface KM1 when the pin fin 51c is formed acts on the adhesion interface KM2 when the straight fin SF is formed. It is about 10% smaller than the stress to be applied. That is, when the pin fin 51c is formed, it can be said that peeling does not easily occur at the adhesive interface KM1. This is based on the following reason.

  As shown in FIG. 18A, since the straight fin SF extends in the direction in which the refrigerant 60 flows (hereinafter, referred to as “flow direction”), the heat dissipation portion 51 has high rigidity in the flow direction. For this reason, when the heat radiation part 51 and the sealing resin 40 are thermally contracted, the heat radiation part 51 is not easily deformed, and the stress acting on the adhesive interface KM2 is not relaxed. On the other hand, as shown in FIG. 17A, since the pin fins 51c are scattered in a lattice shape, the rigidity of the heat radiating part 51 is smaller than when the straight fins SF are formed.

  For this reason, if it is the pin fin 51c, when the thermal radiation part 51 and the sealing resin 40 heat-shrink, the thermal radiation part 51 will be easy to deform | transform and the stress which acts on the adhesion interface KM1 can be relieved. Therefore, when the pin fins 51 c are formed, the sealing resin 40 can be made difficult to peel from the heat radiating portion 51. Note that, as shown in FIGS. 17C and 18C, the peeling of the adhesion interfaces KM1 and KM2 occurs from an outer portion far from the central portion when viewed from the planar direction.

  Next, the material and molding method of the cooling case 50 will be described with reference to FIG. FIG. 19 is a table comparing the shape, cooling performance, laser welding, and ultrasonic flaw detection results when the cooling case 50 is formed by changing the purity of the aluminum material and the forming method. Here, the purity of the aluminum material refers to a ratio of aluminum contained in mass% of the aluminum material constituting the cooling case 50. The chemical composition of the aluminum material is accurately measured by chemical analysis using, for example, ICP-MS (inductively coupled plasma mass spectrometer).

  In the shape shown in FIG. 19, it is determined whether or not the cooling case 50, which is a molded product, has been formed into a desired shape. Further, in the cooling performance, it is determined whether or not the cooling case 50 that is a molded product has a desired heat transfer coefficient. In laser welding, it is determined whether or not many blow holes are generated when laser welding is performed between edges. Further, in ultrasonic flaw detection, whether or not peeling occurs at the adhesive interface KM1 (see FIG. 17B) between the heat radiation part 51 and the sealing resin 40, and whether or not many voids are generated in the heat radiation part 51. Is judged.

  Here, the inspection for ultrasonic flaw detection will be described with reference to FIG. FIG. 20 is a diagram illustrating a state where the power module 10L is inspected for ultrasonic flaw detection. In the ultrasonic flaw detection, as shown in FIG. 20, with the power module 10L arranged in water, the ultrasonic oscillator CO is directed to the power module 10L by ultrasonic waves of several MHz from the pin fin 51c side. Is incident. As a result, when voids are generated in the heat radiation part 51 or when peeling occurs at the adhesive interface KM1, the ultrasonic waves are reflected or scattered. Thus, the presence or absence of peeling and voids is determined.

  As is clear from the results of FIG. 19, in die casting, low pressure casting, squeeze-casting, and vacuum die casting, the desired shape could not be obtained when the purity of the aluminum material was 99% or more. This is because it becomes difficult for the molten metal to wrap around the mold and warping of the molded product occurs. However, regardless of the forming method, the higher the purity of the aluminum material, the higher the heat transfer coefficient of the cooling case 50 and the better the cooling performance.

  Regardless of the forming method, when the purity of the aluminum material is less than 99%, the impurities contained in the aluminum material are bumped to form blowholes when laser welding is performed. In addition, in forging and low-pressure casting, when the aluminum purity was 90% or more, good results were obtained in the ultrasonic inspection. From the above results, the shape, the cooling performance, and the cooling case 50 are formed by forging using an aluminum material whose mass% is aluminum and 99% or more as shown by the hatched portion TZ in FIG. Good results were obtained in all items of laser welding and ultrasonic flaw detection.

  The effect of the power converter device 1 of 1st Embodiment is demonstrated. As shown in FIG. 1, the power conversion device 1 includes two components, a power module 10 </ b> L including a cooling case 50 and a power module 10 </ b> H. The power module 10L and the power module 10H are assembled in a state where the pin fins 51c face each other and the edge portions 54 are sealed. For this reason, the seal structure is simple. Further, in the cooling case 50, the heat radiating portion 51, the header portion 52, the inflow water channel portion 53A, the outflow water channel portion 53B, and the edge portion 54 are integrally formed to ensure the size of the opening of the water channel appropriately. The length of the pin fin 51c can be set appropriately, and the size (the size in the plane direction and the length in the height direction) of the cooling case 50 does not need to be increased more than necessary. Thus, the power conversion device 1 has a small physique of the cooling case 50, and the vehicle mountability can be improved.

  Note that, in the size of the power conversion device 1 of the present embodiment, the dimension in the horizontal direction (left-right direction in FIG. 1) is about 140 mm in the plane direction, and the dimension in the vertical direction (direction perpendicular to the paper surface in FIG. 1). Is about 50 mm. The dimension in the height direction (vertical direction in FIG. 1) is about 35 mm.

  Moreover, in this power converter device 1, as shown in FIG. 10, the bonding strength between the power module 10L and the power module 10H can be sufficiently ensured by laser welding the edge portions 54 to each other. In addition, as shown in FIG. 11, a slit SR as a heat buffer structure is formed between the outer surface 54c of the edge 54 of the power modules 10L and 10H and the assembly groove 54a of the power modules 10L and 10H. Therefore, it is possible to reduce the heat transmitted by the laser welding to the O-ring 75. In particular, the heat buffer structure can be easily configured by configuring the heat buffer structure as the slit SR.

  Moreover, in this power converter 1, the cooling case 50 is shape | molded by forging using the aluminum material whose aluminum is 99% or more by mass%. Thus, a desired shape is obtained by forging an aluminum material with high purity. Moreover, since the purity of the aluminum material is high, the heat transfer coefficient of the cooling case 50 is large and the cooling performance is good. Furthermore, since the impurities contained in the aluminum material are extremely small, the impurities are not bumped by laser welding, and no blowholes are generated.

Second Embodiment
Next, a second embodiment will be described with reference to FIGS. FIG. 21 is a partially enlarged cross-sectional view of the power module 10L in the second embodiment. As shown in FIG. 21, ribs 52b are formed on the surface 52a of the header portion 52 of the cooling case 50 of the second embodiment. The ribs 52 b are for improving the adhesion between the sealing resin 40 and the cooling case 50 by engaging with the sealing resin 40. This rib 52b is equivalent to the engaging convex part of this invention.

  The rib 52b protrudes from the surface 52a of the header part 52, and is formed in the reverse taper shape which becomes so thick that it leaves | separates from the surface 52a of the header part 52 (it goes to the upper side of FIG. 21). As shown in FIG. 21, the rib 52b has only to have a reverse tapered shape at least in cross section. The rib 52b has an inverted truncated cone shape (axisymmetric shape) as shown in FIG. 22A, or FIG. The cross section as shown may be an inversely tapered prismatic shape (axisymmetric shape). Thus, the first projecting piece 52b1 that extends in the plane direction of the header portion 52 is formed on the tip side of the rib 52b. The first projecting piece 52b1 is one in which the sealing resin 40 is added. Since the other configuration of the second embodiment is the same as the configuration of the first embodiment described above, description thereof is omitted.

  Here, a method of forming the rib 52b will be described with reference to FIG. FIG. 23 is a diagram illustrating a process for forming the rib 52b. As shown in FIG. 23A, ribs 52B are formed on the surface 52a of the header portion 52 when the cooling case 50 is forged and formed. The rib 52B has a cylindrical shape or a rectangular parallelepiped shape. As described above, when the insulating member 30 is thermocompression bonded to the upper surface of the cooling case 50 (see FIG. 6B) or when the sealing resin 40 is injected (see FIG. 8), FIG. As shown in B), the tip of the rib 52B is pressed toward the surface 52a of the header 52 using the pressing pin OP1. Here, the cross-sectional area of the pressing pin OP1 is larger than the cross-sectional area of the rib 52B. In this way, the above-described rib 52b is formed in a process in the middle of molding the power module 10L, that is, without causing time loss due to the addition of a new separate process.

  The effect of 2nd Embodiment is demonstrated. In the second embodiment, since the rib 52b and the sealing resin 40 are engaged, the bonding strength between the sealing resin 40 and the cooling case 50 can be increased. In particular, since the sealing resin 40 holds the first projecting piece 52b1 formed on the rib 52b, the bonding strength between the sealing resin 40 and the cooling case 50 can be further increased. Thus, it becomes easy to ensure the rigidity required for the cooling case 50.

  By the way, as shown in FIG. 17B, stress may act on the adhesion interface KM1 between the heat radiation part 51 (cooling case 50) and the sealing resin 40 to cause peeling. In particular, when the area of the semiconductor element 20 is increased so that a large current can be applied due to an increase in the size of a drive motor of an automobile, the power conversion device 1 is increased in size. As a result, the stress that causes the adhesive interface KM1 to peel off also increases and the risk of peeling increases. On the other hand, in 2nd Embodiment, peeling of the adhesion interface KM1 can be prevented by increasing the joint strength between the sealing resin 40 and the cooling case 50.

  In the second embodiment, as shown in FIG. 21, the pin fins 51 c are formed on the back surface 51 b of the heat radiating part 51, whereas the ribs 52 b are formed on the front surface 52 a of the header part 52. And the fluctuation | variation of the thickness distribution in the header part 52 is reduced. That is, by forming the rib 52b on the front surface 52a of the header portion 52 where the pin fins 51c are not formed on the back surface, the difference between the thickness of the header portion 52 and the thickness of the heat radiating portion 51 is reduced. Therefore, it becomes easy to form the cooling case 50 by forging. Other functions and effects of the second embodiment are the same as the functions and effects of the first embodiment described above, and a description thereof will be omitted.

<Third Embodiment>
Next, a third embodiment will be described with reference to FIGS. FIG. 24 is a partially enlarged cross-sectional view of the power module 10L according to the third embodiment. As shown in FIG. 24, a recess 51d is formed on the surface 51a of the heat radiation part 51 of the cooling case 50 of the third embodiment. The recess 51 d is for improving the adhesion between the sealing resin 40 and the cooling case 50 by engaging with the sealing resin 40. The recess 51d corresponds to the engagement recess of the present invention.

  The concave portion 51d is recessed from the surface 51a of the heat radiating portion 51, and the bottom side (lower side in FIG. 24) of the concave portion 51d becomes thicker as it is away from the surface 51a of the heat radiating portion 51 (toward the lower side in FIG. 24). It is formed in a reverse taper shape. As shown in FIG. 24, the bottom side of the recess 51d only needs to have at least a cross section that is reversely tapered, and may have an inverted truncated cone shape (axisymmetric shape) as shown in FIG. The cross section as shown in B) may be an inversely tapered prismatic shape (axisymmetric shape). And the 2nd protrusion 51d1 which protrudes inside is formed in the recessed part 51d so that the recessed part 51d may be plugged up. The second projecting piece 52d1 holds the sealing resin 40 therein. Since the other configuration of the third embodiment is the same as the configuration of the first embodiment described above, the description thereof is omitted.

  Here, a method for forming the recess 51d will be described with reference to FIG. FIG. 26 is a diagram illustrating a process for forming the recess 51d. As shown in FIG. 26A, when the cooling case 50 is forged and formed on the surface 51a of the heat radiating portion 51, a recess 51D is formed. The recess 51D has a cylindrical shape or a rectangular parallelepiped shape. As described above, when the insulating member 30 is thermocompression bonded to the upper surface of the cooling case 50 (see FIG. 6B) or when the sealing resin 40 is injected (see FIG. 8), FIG. As shown to B), the surface 51a of the thermal radiation part 51 located around the recessed part 51D is pressed using pressing pin OP2. Here, the tip of the pressing pin OP2 is notched in a tapered shape. In this way, the above-described recess 51d is formed in a process in the middle of molding the power module 10L, that is, without causing time loss due to the addition of a new separate process.

  The effect of 3rd Embodiment is demonstrated. In 3rd Embodiment, since the recessed part 51d and the sealing resin 40 engage, the joining strength of the sealing resin 40 and the cooling case 50 can be enlarged. In particular, since the second protrusion 51d1 formed in the recess 51d holds the sealing resin 40, the bonding strength between the sealing resin 40 and the cooling case 50 can be further increased. Thus, it becomes easy to ensure the rigidity required for the cooling case 50.

  By the way, as shown in FIG. 17B, stress may act on the adhesion interface KM1 between the heat radiation part 51 (cooling case 50) and the sealing resin 40 to cause peeling. In particular, when the area of the semiconductor element 20 is increased so that a large current can be applied due to an increase in the size of a drive motor of an automobile, the power conversion device 1 is increased in size. As a result, the stress that causes the adhesive interface KM1 to peel off also increases and the risk of peeling increases. On the other hand, in 3rd Embodiment, peeling of the adhesion interface KM1 can be prevented by increasing the joint strength between the sealing resin 40 and the cooling case 50.

  In the third embodiment, since the pin fins 51 c are formed on the back surface 51 b of the heat radiating portion 51 and the concave portions 51 d are formed on the front surface 51 a of the heat radiating portion 51, fluctuations in the thickness distribution in the heat radiating portion 51 and the header portion 52. Can be reduced. That is, the increase in the thickness of the heat radiation part 51 caused by forming the pin fins 51c can be suppressed by forming the recess 51d, and the difference between the thickness of the heat radiation part 51 and the thickness of the header part 52 is small. Become. Therefore, it becomes easy to form the cooling case 50 by forging. Other functions and effects of the third embodiment are the same as the functions and effects of the first embodiment described above, and a description thereof will be omitted.

<Fourth embodiment>
Next, a fourth embodiment will be described with reference to FIGS. FIG. 27 is a partial cross-sectional view when the right half of the power conversion device 1S of the fourth embodiment is broken. FIG. 28 is an exploded view of power converter 1S shown in FIG. This power conversion device 1S is configured to form an inverter circuit that drives three three-phase AC motors. As shown in FIGS. 27 and 28, the power conversion device 1S includes a first power conversion device 1S1, a second power conversion device 1S2, a third power conversion device 1S3, a housing 80, a bottom cover 81, A water supply pipe 82 and a water distribution pipe 83 are provided. Since the configuration of the first, second, and third power converters 1S1, 1S2, and 1S3 and the configuration of the power converter 1 of the first embodiment are the same, the corresponding parts are denoted by the same reference numerals, and the description thereof is given. Omitted.

  The housing 80 is for covering the first, second, and third power converters 1S1, 1S2, and 1S3, and the lower side is open. Further, the housing 80 has an inflow hole 80a and an outflow hole 80b for allowing the refrigerant 60 to flow in or out on the upper wall. The bottom lid 81 covers the housing 80 from below, and has sealing portions 81a and 81b for sealing the water channel. The water supply pipe 82 is a pipe for feeding the refrigerant 60 into the housing 80, and is fitted to the upper part of the inflow hole 80 a of the housing 80. The water distribution pipe 83 is a pipe for sending the refrigerant 60 out of the housing 80, and is fitted in the upper part of the outflow hole 80 b of the housing 80.

  The assembly of this power conversion device 1S will be described. The assembly of the power conversion device 1S is performed after the first, second, and third power conversion devices 1S1, 1S2, and 1S3 are assembled. That is, the power modules 10L and 10H are assembled with the fixing pins 76 (see FIG. 1) to constitute the first, second and third power converters 1S1, 1S2 and 1S3, and then the first, second and third The power conversion devices 1S1, 1S2, and 1S3 are stacked in a direction orthogonal to the planar direction of the heat dissipation portion 51.

  When stacking the power converters, the inflow water channel 53A and the outflow water channel 53B of the power module 10L are connected to the fitting holes KG of the inflow water channel 53A and the outflow water channel 53B of the power module 10H via the O-rings 73 and 74. To fit. Thus, the sealing resin 40 of the power module 10L and the sealing resin of the power module 10H are adjacent to each other. A slight gap is formed between the sealing resin 40 of the adjacent power module 10L and the sealing resin of the power module 10H.

  The housing 80 covers the first, second, and third power conversion devices 1S1, 1S2, and 1S3, and the inflow water channel portion 53A and the outflow water channel portion of the power module 10L are provided between the housing 80 and the third power conversion device 1S3. 53B is fitted into the inflow hole 80a and the outflow hole 80b of the housing 80 via the O-rings 73 and 74. Further, between the first power conversion device 1S1 and the bottom lid 81, the sealing portions 81a and 81b of the bottom lid 81 are connected to the inflow water channel portion 53A and the outflow water channel portion 53B of the power module 10H via O-rings 77 and 78. Fit into the fitting hole KG. Finally, the lower end of the side wall of the housing 80 and the end of the bottom lid 81 are connected via a bolt 84. In this way, the refrigerant 60 is sealed at the fitted portion so as not to leak, and the power conversion device 1S is assembled.

  Next, before explaining the advantages of the power conversion device 1S of the fourth embodiment, a power conversion device 1T configured as a comparative product will be described. FIG. 29 is a partial cross-sectional view when the right half of the power converter 1T is broken. FIG. 30 is an exploded view of the power converter 1T shown in FIG. The power conversion device 1T is configured to form an inverter circuit that drives three three-phase AC motors.

  As shown in FIGS. 29 and 30, the power conversion device 1T has a first power conversion device 1T1 and a second power conversion device 1T2 in the center, and a power card 110L and a cooling case 190M on the lower side. In addition, a power card 110H and a cooling case 190N are provided on the upper side. The power conversion device 1T includes a housing 80, a bottom cover 81, a water supply pipe 82, and a water distribution pipe 83, as in the power conversion apparatus 1S described above. Since the configuration of the first and second power conversion devices 1T1 and 1T2 and the configuration of the power conversion device 1W as a comparative product described in the first embodiment are the same, the corresponding parts are denoted by the same reference numerals, The description is omitted.

  The power card 110L disposed on the lower side has the same configuration as that of the power card 110L of the power conversion device 1W as a comparative product described in the first embodiment. The cooling case 190M is open on the lower side, and is the same as the configuration on the upper side of the cooling case 190 of the power conversion device 1W. The power card 110H has the same configuration as that of the power card 110H of the power conversion device 1W. The cooling case 190N is open on the upper side, and has the same configuration as the lower side of the cooling case 190 of the power conversion device 1W.

  The assembly of this power conversion device 1T will be described. The power conversion device 1T is assembled so that the cooling case 190M, the power card 110L, the first and second power conversion devices 1T1 and 1T2, the cooling case 190N, and the power card 110H are assembled together. That is, the above-described components are not fixed in advance with a fixing pin or the like, and are stacked in a direction orthogonal to the planar direction of the heat radiating portion 51 using the housing 80 and the bottom cover 81.

  When stacking the power conversion devices, the inflow water channel portions 192A and the outflow water channel portions 192B of each cooling case are fitted through the O-ring 196, and the end portions of the heat sinks 150 of the power cards 110L and 110H are The cooling case 190 is assembled to the concave groove 193a through the rings 194 and 195. Thus, the power conversion device is laminated so that the sealing resin 140 of the power card 110L and the sealing resin 140 of the power card 110H are adjacent to each other.

  Then, the cooling case 190M, which covers each part with the housing 80 and is arranged on the lower side, is assembled to the bottom lid 81 in a state where the O-ring 198 is assembled to the annular groove KO. The cooling case 190N arranged on the upper side is assembled to the upper wall of the housing 80 in a state where the O-ring 199 is assembled to the annular groove KO. In this manner, the lower end of the side wall of the housing 80 and the end of the bottom cover 81 are connected via the bolt 84 in a state where the respective components are assembled. Thus, the power converter 1T is assembled by pressurizing and fixing the entire module.

  This power converter 1T has the following problems. When assembling the power converter 1T, the O-ring 196 that seals the inflow water channel portions 192A and the outflow water channel portions 192B of the cooling case and the O-rings 194 and 195 that seal between the power cards 110L and 110H and the cooling case. The O-ring 198 that seals between the cooling case 190M and the bottom lid 81 and the O-ring 199 that seals between the cooling case 190N and the upper wall of the housing 80 must all be pressurized together.

  For this reason, if the pressure area of the O-ring to be pressurized is large and sufficient pressure cannot be applied to all the O-rings, refrigerant leakage occurs. In order to cope with this, when a very large pressure is applied to all the O-rings, the housing 80 and the cooling cases 190, 190M, and 190N require a very large strength. Thus, the power converter 1T has a problem that the housing 80 and the cooling case are increased in size and weight.

  In addition, due to the structure in which pressure is applied to all O-rings at once, the relative height of the pressure surface of each O-ring must be ensured with high accuracy. For this reason, for example, when the pressure surface of one O-ring is high while the pressure surface of the other O-ring is low, pressure cannot be applied evenly to the O-ring. There is a risk of refrigerant leakage. Thus, the power conversion device 1T has a problem that refrigerant leakage is likely to occur unless the relative height of the pressure surface of each O-ring is ensured with high accuracy.

  Further, due to the structure in which the components are assembled together, a leak inspection for inspecting refrigerant leakage must be performed after the components are covered and fixed by the housing 80. For this reason, in the leak inspection, if a leak product in which the refrigerant leaks due to an improper assembly of the O-ring is found, all components must be disassembled to find the cause of the refrigerant leak. Thus, in this power conversion device 1T, there is a problem that in the production process, it takes a lot of time and effort to find a seal failure, which is inefficient.

  On the other hand, the power converter 1S of the fourth embodiment has the following advantages. In assembling the power converter 1S, the first, second, and third power converters 1S1, 1S2, and 1S3 are assembled, and then the housing 80 covers the first, second, and third power converters 1S1, 1S2, and 1S3. And stack. For this reason, the seal of the refrigerant space RK in which the refrigerant 60 flows toward the pin fins 51c (the seal by the O-ring 75) and the seal of the water channel into which the refrigerant 60 flows in or out (the seal by the O-rings 73 and 77) are independent. Yes. Therefore, only the O-rings 73 and 77 need to be pressurized in the seal when the housing 80 covers the first, second, and third power converters 1S1, 1S2, and 1S3, and the pressure area of the O-ring is sufficient. Is small. Therefore, in this power conversion device 1S, compared with the above-described power conversion device 1T, the pressurizing force of the housing 80 is small, the housing 80 and the cooling case do not need great strength, and the housing 80 and the cooling case can be downsized.

Here, the pressing force of the housing 80 of the power conversion device 1T and the pressing force of the housing 80 of the power conversion device 1S of the fourth embodiment will be described in comparison. First, FIG. 31 (A) is a schematic diagram showing pressurization areas of the O-ring 195 and the O-ring 196 in the power conversion device 1T. Here, when the O-ring 195 has a wire diameter of 4 mm, a horizontal dimension D1 of 80 mm, and a vertical dimension D2 of 40 mm, the pressure area of the O-ring 195 is 4 × (80 + 40) × 2 = 960 mm ² . If the O-ring 196 has a wire diameter of 2 mm and a diameter of 20 mm, the pressure area of the two O-rings 196 is 2 × (10 ² −8 ² ) × π = 226 mm ² . And the pressurization force required per unit area shall be 1 MPa. Thereby, the applied pressure of the housing 80 of the power converter 1T is (186 + 226) mm ² × 1 MPa, which is 1.186 kN.

Next, FIG. 31 (B) is a schematic diagram showing the pressure areas of the O-rings 73 and 74 in the power conversion device 1S. Here, when the wire diameter of the O-rings 73 and 74 is 2 mm and the diameter is set to 20 mm, the pressure area of the O-rings 73 and 74 is 2 × (10 ² −8 ² ) × π = 226 mm ² . And the pressurization force required per unit area shall be 1 MPa. Thereby, the applied pressure of the housing 80 of the power conversion device 1S is 226 mm ² × 1 MPa, which is 0.226 kN. From the above, the pressurizing force of the housing 80 of the power conversion device 1S is approximately 1 / 5.2 (0.226 ÷ 1.186) of the pressurizing force of the housing 80 of the power conversion device 1T. Therefore, in the power conversion device 1S, the strength required for the housing 80 and the cooling case is reduced, and the housing 80 and the cooling case can be downsized.

  Further, in the power conversion device 1S of the fourth embodiment, as shown in FIG. 32, the pressure surface of the O-ring 75 is structured so that the seal by the O-ring 75 and the seal by the O-rings 73 and 77 are independent. It is easy to ensure the relative height H1 of the O-ring 73 with high accuracy, and it is easy to ensure the relative height H2 of the pressure surface of the O-ring 73 with high accuracy. That is, since the dimensions C1 and D1 shown in FIG. 32 are determined by the size of the forging or casting mold, the dimensional error is about several tens of μm and can be set with high accuracy. Moreover, when it is necessary to set the C1 dimension and the D1 dimension with higher accuracy, they may be corrected by cutting after being formed by forging or casting. FIG. 32 is an enlarged view of a part of the power conversion apparatus 1S shown in FIG.

  Thus, since the C1 dimension and the D1 dimension can be set with high accuracy, the relative height H1 of the pressure surface of the O-ring 75 can be secured with high accuracy, and the height H1 can be set with high accuracy. Accordingly, the height H2 can be set with high accuracy. Therefore, in this power conversion device 1S, the relative height shift of the pressure surfaces of the O-rings 75, 73, 77 is difficult to occur, and it is easy to apply pressure equally to the O-rings 75, 73, 77.

  Here, the advantage that the dimension H1 shown in FIG. 32, that is, the dimension between the power converters 1S1, 1S2, and 1S3 can be set with high accuracy will be described with reference to FIG. 33 and FIG. FIG. 33 is a diagram illustrating a state in which the control terminal 71 of the power conversion device 1S and the control board 85 are connected. As shown in FIG. 33, a flat substrate fixing table 86 is installed, and three boss portions 87 are provided on the substrate fixing table 86. The control board 85 is assembled to these boss portions 87 via fixing screws 88. Each control terminal 71 extends from each power conversion device 1S1, 1S2, 1S3 in a direction orthogonal to the planar direction of the control board 85, and as shown in FIG. 34, a through hole formed in the control board 85. It penetrates 85a. FIG. 34 is a cross-sectional view taken along line YY of FIG.

  By the way, as shown in FIG. 34, the cross section of the control terminal 71 is substantially square, and the length S1 of one side of the control terminal 71 is 0.8 mm. The diameter R1 of the through hole 85a is 1.2 mm. For this reason, there is a margin of the gap f1 by a maximum of ± 0.2 mm between the through hole 85a and the control terminal 71. In other words, in order to allow each control terminal 71 to pass through all the through holes 85a, the relative positional deviation between the control terminals 71 must be suppressed within ± 0.2 mm.

  In the power conversion device 1S of the fourth embodiment, the dimension H1 shown in FIG. 32 can be set with high accuracy. Therefore, in the one-stage power conversion device 1S1, the control terminal 71 at one end is controlled from the control terminal 71 at the other end. The deviation (error) of the distance E1 is up to ± 0.05 mm. Further, in the two-stage power converters 1S1 and 1S2, the deviation of the distance E2 from the control terminal 71 at one end to the control terminal 71 at the other end is ± 0.10 mm or less. Thus, in the three-stage power converters 1S1, 1S2, and 1S3, the deviation of the distance E3 from the control terminal 71 at one end to the control terminal 71 at the other end can be ± 0.15 mm or less. Therefore, in the power conversion device 1S, even if the power conversion devices 1S1, 1S2, and 1S3 are stacked, the control terminals 71 can be passed through all the through holes 85a without correcting the control terminals 71. it can.

  In the case of the power conversion device 1T shown in FIG. 29, since all the O-rings 196, 195, 197, and 198 are pressurized together, the control terminal 71 on one end controls the other end. The deviation of the distance to the terminal 71 (the distance corresponding to E3 in FIG. 31) is about ± 1.2 mm. For this reason, unless the control terminal 71 is corrected, each control terminal 71 cannot be penetrated through all the through holes 85a.

  Further, in the power conversion device 1S of the fourth embodiment, before the first, second, and third power conversion devices 1S1, 1S2, and 1S3 are covered and stacked by the housing 80, the power conversion devices 1S1, 1S2, and 1S3 are stacked. On the other hand, a leak test for inspecting refrigerant leakage can be performed. For this reason, it is possible to easily find the cause of leakage of the refrigerant (leak product) by individually inspecting each of the power conversion devices 1S1, 1S2, and 1S3. Thus, in this power conversion device 1S, the effort for finding a seal failure is small and efficient in the production process.

  Moreover, the power conversion device 1S of the fourth embodiment also has the following advantages. In the power conversion device 1S, as shown in FIG. 28, a total of 11 O-rings (O-rings 73, 74, 75, 77, 78) are sealed, whereas in the power conversion device 1T, 30, a total of 14 pieces (O-rings 194, 195, 196, 198, 199) are used for sealing. Therefore, the power conversion device 1S can reduce the number of O-rings and can be configured at a lower cost than the power conversion device 1S. Further, in the power conversion device 1S, since it is not necessary to use the fin interval wall 197 unlike the power conversion device 1T, the length in the height direction (the vertical dimension in FIG. 29) can be reduced, and the vehicle Mountability can be improved.

  Moreover, in the power converter device 1S of 4th Embodiment, between 1st power converter device 1S1 and 2nd power converter device 1S2 (refer FIG. 32), between 2nd power converter device 1S2 and 3rd power converter device 1S3. In the meantime, the semiconductor element 20 of the power module 10H constituting the high side circuit and the semiconductor element 20 of the power module 10L constituting the low side circuit are arranged to face each other. Thereby, since the high-side circuit and the low-side circuit are close to each other, the mutual inductance can be kept low, and the surge generated by the switching of the semiconductor element 20 can be reduced.

  The effect of 4th Embodiment is demonstrated. In the fourth embodiment, a pair of power modules 10 </ b> L and power modules 10 </ b> H (each power conversion device 1 </ b> S <b> 1, 1 </ b> S <b> 2, 1 </ b> S <b> 3) in a pre-assembled state are stacked in a direction orthogonal to the planar direction of the heat dissipation part 51. Therefore, the seal of the refrigerant space that flows toward the pin fins 51c (the seal by the O-ring 75) and the seal of the water channel through which the refrigerant flows in or out (the seal by the O-rings 73 and 77) are made independent. be able to. For this reason, the applied pressure at the time of sealing a water channel can be made small. Accordingly, the strength required for the cooling case is reduced, and the cooling case can be reduced in size. Other functions and effects of the fourth embodiment are the same as the functions and effects of the first embodiment described above, and a description thereof will be omitted.

<Modified Embodiment>
As mentioned above, although demonstrated in the power converter device which concerns on this invention, this invention is not limited to this, A various change is possible in the range which does not deviate from the meaning. For example, in the first embodiment, as illustrated in FIG. 10, the edge portions 54 of the power modules 10 </ b> L and 10 </ b> H are laser-welded to increase the bonding strength between the power module 10 </ b> L and the power module 10 </ b> H. However, as shown in FIG. 35, the edge portions 54 of the power modules 10L and 10H may be sandwiched by the clip CR. The clip CR is made of a phosphor bronze spring material. The clip CR may sandwich the sealing resin 40 of the power module 10L and the sealing resin 40 of the power module 10H.

  Moreover, as shown in FIG. 36, it may be cast with a second sealing resin HZ of a thermoplastic resin. In this case, since the second sealing member HZ covers the entire edge 54 of the power modules 10L and 10H and the entire sealing resin 40, the bonding strength between the power module 10L and the power module 10H can be increased. The leakage of the refrigerant can be prevented.

  In the first embodiment, the power module 10L and the power module 10H are caulked by inserting the fixing pin 76 (see FIG. 1) through the insertion hole 54b (see FIGS. 2 and 3) of the cooling case 50. Assembled. However, the assembly of the power module 10L and the power module 10H is not limited to caulking using the fixing pin 76, and can be changed as appropriate. For example, without using the fixing pin 76, the power module 10L and the power module 10H may be assembled by edge welding 54 only by laser welding.

  Moreover, in 1st Embodiment, the cooling case 50 was shape | molded by forging using the aluminum material whose aluminum is 99% or more by mass%. However, the purity and forming method of the aluminum material are not limited to those described above, and can be changed as appropriate. For example, the cooling case 50 may be formed by low pressure casting using an aluminum material whose mass% is aluminum of 90% or more and less than 99%.

  In the first embodiment, the material of the cooling case 50 is aluminum (aluminum alloy). However, the material of the cooling case 50 is not limited to aluminum. For example, copper (copper alloy), resin, or the like is appropriately used. It can be changed. Moreover, although the fin formed in the back surface of the thermal radiation part 51 of the cooling case 50 was the pin fin 51c, a fin is not limited to the pin fin 51c, A corrugated fin, a straight fin, etc. can be changed suitably.

  Moreover, in 2nd Embodiment, as shown in FIG. 21, the reverse taper-shaped rib 52b which becomes thick, so that it leaves | separates from the surface 52a of the header part 52 was formed. However, the shape of the rib 52b is not limited to the reverse tapered shape, and may be a cylindrical shape or a tapered shape, and can be changed as appropriate. For example, as shown in FIG. 37B, the rib may be a rib 52c whose tip is formed in a V-shaped wedge shape. As shown in FIG. 37A, the rib 52c is formed by using a pressing pin OP2 whose tip is notched in a tapered shape with respect to the columnar rib 52C formed on the surface 52a of the header portion 52. Press. Thus, the first projecting piece 52c1 extending in the plane direction of the header portion 52 is formed on the tip side of the rib 52c.

  Further, as shown in FIG. 38, the back surface 52d of the header portion 52 may be formed with a rectifying rib 52e extending in the direction in which the refrigerant 60 flows. The rectifying ribs 52e adjust the flow of the refrigerant 60 and are formed on the back surface 52d of the header portion 52 by six. The three rectifying ribs 52e are formed to extend radially from the inflow hole RN of the inflow water channel portion 53A toward the pin fins 51c. The remaining three rectifying ribs 52e are formed to extend radially from the outflow hole RS of the outflow water channel portion 53B toward the pin fins 51c.

  By the way, since the refrigerant 60 generally proceeds straightly, the flow rate of the refrigerant 60 is smaller in the end portion HB indicated by the one-dot chain line in FIG. 38 than in the central portion CB indicated by the one-dot chain line in FIG. . For this reason, in the end portion HB, the heat generated by the semiconductor element 20 is difficult to cool compared to the central portion CB. Thus, there is a problem that the semiconductor element 20 is locally heated.

  Therefore, as described above, when the flow regulating rib 52e is formed on the back surface 52d of the header portion 52, the refrigerant 60 flows so as to spread from the outflow hole RN and flows so as to gather in the outflow hole RS. Thereby, between the edge part HB and the center part CB, the flow rate difference of the refrigerant | coolant 60 becomes small and the cooling condition of the semiconductor element 20 becomes uniform. Thus, local overheating of the semiconductor element 20 can be prevented by the rectifying rib 52e. Moreover, since the rectifying rib 52e is formed on the back surface 52d of the header portion 52 and the pin fins 51c are formed on the back surface 51b of the heat radiating portion 51, the difference between the thickness of the header portion 52 and the thickness of the heat radiating portion 51 becomes small. . Therefore, it becomes easy to form the cooling case 50 by forging.

  Moreover, in 3rd Embodiment, as shown in FIG. 24, the recessed part 51d which has the 2nd protrusion 51d1 which was depressed from the surface 51a of the thermal radiation part 51 and protruded inside was formed. However, the shape of the recess 51d is not limited to the above, and can be changed as appropriate. For example, the recess 51d may have a cylindrical shape that does not include the second projecting piece 51d1.

  Moreover, in 4th Embodiment, as shown in FIG. 27, the power converter device 1S was comprised by laminating | stacking three power converter device 1S1, 1S2, 1S3. However, the number of stacked power conversion devices is not limited to three and can be changed as appropriate, and may be two or four, for example.

  Further, instead of reinforcing the edge portions 54 by welding or the like, as shown in FIG. 39, between the sealing resin 40 of the power module 10H of the first power conversion device 1S1 and the bottom cover 81, a rubber or plate A first elastic member DB1 such as a spring is interposed, and a second elastic member DB2 such as a rubber or a leaf spring is interposed between the sealing resin 40 of the power module 10L of the third power converter 1S3 and the housing 80. And you may comprise the power converter device 1U. In this power conversion device 1U, the sealing resin 40 of the power module 10L of the first power conversion device 1S1 and the sealing resin 40 of the power module 10H of the second power conversion device 1S2 (a pair assembled) A third elastic member DB3 is interposed between the pair of power modules 10L and 10H assembled with the power modules 10L and 10H, and the power module 10L of the second power converter 1S2 is sealed. A third elastic member DB3 is interposed between the stop resin 40 and the sealing resin 40 of the power module 10H of the third power converter 1S3.

  Thus, the entire cooling case 50 can be backed up and leakage of the refrigerant 60 can be prevented even if reinforcement such as welding is omitted with respect to the pressure of the refrigerant 60 and the reaction force of the O-ring 75. That is, the entire power conversion device 1U (cooling case 50) can be reinforced by the housing 80, the bottom cover 81, the first elastic member DB1, and the second elastic member DB2, and leakage of the refrigerant 60 can be prevented. Furthermore, in this power conversion device 1U, the impact acting on the bottom lid 81 and the housing 80 can be absorbed by the elastic members DB1 and DB2, and the structure is strong against the impact. Further, the entire power conversion device 1U is further reinforced by filling the gap formed between the sealing resins 40 of the assembled power modules 10L and 10H with the third elastic member DB3.

  40A, the thermosetting resin sealing resin 40K that covers the semiconductor element 20 projects outward from the end 50t of the cooling case 50, and the power conversion device 1K shown in FIG. As shown in B), the end 50t of the cooling case 50 extends in the planar direction of the heat dissipating part 51 (particularly, the direction in which the control terminal 71 and the power terminal 72 extend), and the sealing resin 40K is inserted into the third part. A protruding piece 50t1 may be formed. In this case, since the sealing resin 40K includes the third projecting piece 50t1, the bonding strength between the sealing resin 40K and the cooling case 50 can be further increased, and the rigidity required for the cooling case 50 can be increased. It becomes easy to secure.

  In each embodiment, the number of semiconductor elements 20 (a pair of IGBTs 20A and diodes 20B) mounted on the power modules 10L and 10H of the power conversion device is not limited to three and can be changed as appropriate. .

1, 1S, 1K Power conversion device 10L, 10H Power module 20 Semiconductor element 30 Insulating member 40 Sealing resin 50 Cooling case 50t1 Third protrusion 51c Heat radiation part 51c Pin fin 51d Recessed part 51d1 Second protrusion 52 Header part 52b, 52c Rib 52b1, 52c1 First protrusion 52e Flow regulating rib 53A Inflow water channel portion 53B Outflow water channel portion 54 Edge portion 54a Assembly groove 60 Refrigerant 71 Control terminal 72 Power terminals 73, 74, 75 O-ring 80 Housing 81 Bottom lid SR Slit DB1 First Elastic member DB2 Second elastic member DB3 Third elastic member

Claims (14)

In the power conversion device that can cool the heat generated by the switching of the semiconductor element,
It is composed of a first power module and a second power module having a cooling case joined to the semiconductor element via an insulating member,
The cooling case of each power module is
A heat dissipating part extending in a planar shape and having the insulating member bonded to the front surface and fins formed on the back surface;
A water channel portion extending in a direction orthogonal to the planar direction of the heat radiating portion and into which refrigerant flows in and out,
An edge that surrounds the fin and forms a space through which the refrigerant flows,
These heat radiation part, water channel part and edge part are molded integrally,
The first power module and the second power module are assembled with the fins facing each other and the edges of each other sealed .
The edge of the first power module and the edge of the second power module are welded;
The edge is formed with a heat buffering structure that makes it difficult to transfer heat from the outer surface toward the inside from the outer surface,
The power conversion device , wherein the heat buffer structure is a slit . In the power conversion device that can cool the heat generated by the switching of the semiconductor element,
It is composed of a first power module and a second power module having a cooling case joined to the semiconductor element via an insulating member,
The cooling case of each power module is
A heat dissipating part extending in a planar shape and having the insulating member bonded to the front surface and fins formed on the back surface;
A water channel portion extending in a direction orthogonal to the planar direction of the heat radiating portion and into which refrigerant flows in and out,
An edge that surrounds the fin and forms a space through which the refrigerant flows,
These heat radiation part, water channel part and edge part are molded integrally,
The first power module and the second power module are assembled with the fins facing each other and the edges of each other sealed .
The edge of the first power module and the edge of the second power module are welded;
The edge is formed with a heat buffering structure that makes it difficult to transfer heat from the outer surface toward the inside from the outer surface,
The power conversion device according to claim 1, wherein the thermal buffer structure is a step junction between an edge of the first power module and an edge of the second power module . In the power converter device according to claim 1 or 2 ,
The edge part of the said 1st power module and the edge part of the said 2nd power module are assembled | attached through the O-ring, The power converter device characterized by the above-mentioned. In the power converter device according to any one of claims 1 to 3 ,
The cooling case has a header portion that extends continuously in a planar shape from the heat radiating portion and has no fins formed thereon,
A thermosetting resin sealing member is formed on the surfaces of the heat radiating part and the header part so as to cover the semiconductor element,
The power conversion device according to claim 1, wherein an engagement convex portion that protrudes from a surface and engages with the sealing member is formed on the header portion. In the power converter device described in Claim 4 ,
The power conversion device according to claim 1, wherein a first projecting piece that extends in a planar direction of the header portion and that holds the sealing member is formed on the engagement convex portion. In the power converter device according to any one of claims 1 to 3 ,
The cooling case has a header portion that extends continuously in a planar shape from the heat radiating portion and has no fins formed thereon,
A thermosetting resin sealing member is formed on the surfaces of the heat radiating part and the header part so as to cover the semiconductor element,
The power conversion device according to claim 1, wherein an engagement recess that is recessed from the surface and engages with the sealing member is formed in the heat dissipation portion. In the power converter device described in Claim 6 ,
The power conversion device, wherein the engagement recess is formed with a second protrusion that protrudes inward to close the engagement recess and holds the sealing member. In the power converter device according to any one of claims 1 to 3 ,
On the surface of the heat dissipation part, a thermosetting resin sealing member is formed so as to cover the semiconductor element,
The sealing member projects outward from the end of the cooling case,
A power conversion device, wherein a third projecting piece is formed at an end of the cooling case, extending in a planar direction of the heat radiating portion and holding the sealing member. In the power converter device according to any one of claims 1 to 3 ,
On the surface of the heat dissipation part, a thermosetting resin sealing member is formed so as to cover the semiconductor element,
A power conversion device, wherein a second sealing member made of a thermoplastic resin is formed so as to cover the entire sealing member and the entire edge of the first power module and the second power module. In the power converter according to any one of claims 1 to 9 ,
The power conversion device according to claim 1, wherein the cooling case is formed by forging using an aluminum material whose mass% is aluminum of 99% or more. In the power converter device according to any one of claims 1 to 10 ,
The first power module and the second power module each have three pairs of diodes and transistors to form an inverter circuit that drives one motor. The power conversion device according to any one of claims 1 to 11 ,
A plurality of the assembled first power module and second power module are stacked in a direction orthogonal to the planar direction of the heat radiating portion. In the power converter device described in Claim 12 ,
The plurality of stacked first power modules and second power modules are assembled so as to be covered with a housing and a bottom cover,
A first elastic member is interposed between the bottom and the first power module and the second power module positioned at the bottom of the plurality of stacked first and second power modules,
A second elastic member is interposed between the housing between the first power module and the second power module positioned at the top of the plurality of stacked first power modules and second power modules. The power converter characterized by the above-mentioned. In the power converter device described in Claim 12 or Claim 13 ,
A third elastic member is interposed between the pair of first power module and second power module assembled and the pair of first power module and second power module assembled. The power converter characterized by the above-mentioned.