JP2013146168A - Power conversion device and method for manufacturing the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve manufacturing efficiency of a power conversion device that includes a semiconductor unit in which semiconductor modules and coolers are laminated.SOLUTION: A power conversion device includes: a semiconductor unit 20; a frame 12 on which the semiconductor unit 20 is to be mounted; and a spring unit 6. The spring unit 6 is inserted between an end surface 20a of the semiconductor unit 20 in a lamination direction and supports 13 that are provided on the frame 12, and supports the semiconductor unit 20 while applying pressure to the semiconductor unit 20. The spring unit 6 has a first plate 4, a second plate 5, and a spring 3 sandwiched therebetween. At least one of the first plate 4 and the second plate 5 is provided with a flange 4a or 5a that extends toward the other plate. The height of the flange is determined in such a manner that a gap formed between the two plates 4 and 5 with a leading end of the flange in contact with the other plate is larger than the length of the spring 3 at the limit of compression thereof.

Description

  The technology disclosed in this specification relates to a power conversion device such as an inverter or a converter and a method for manufacturing the same.

  In an inverter that supplies electric power to a motor or a voltage converter that converts a voltage, a structure in which semiconductor elements having a large amount of heat, such as an IGBT, are integrated separately from other circuits and intensively cooled may be employed. As one of such integrated structures, there is known a semiconductor unit in which flat type semiconductor modules containing semiconductor elements having a large calorific value and flat type coolers are alternately stacked.

  In the stacked type semiconductor unit, the semiconductor module is attached to the frame in a state where pressure is applied in the stacking direction so that heat is well transmitted from the semiconductor module to the cooler. Here, the frame is a structure that supports the semiconductor module, and may typically be a case of a power converter. In order to apply pressure to the semiconductor unit, the semiconductor unit is disposed between a pair of stopper portions provided on the frame (a wall or a support provided on the frame, which is in contact with and supports the semiconductor unit), A spring is inserted between one end of the semiconductor unit stacking direction and one stopper portion of the frame. Such a structure is disclosed in Patent Documents 1 to 3, for example.

  Patent Documents 1 and 2 disclose a technique in which a fixing pin and a leaf spring are arranged between a wall surface of a frame and a semiconductor unit. The fixing pin corresponds to the stopper portion. In the technique of Patent Document 1, the middle of a stacked body of semiconductor units is fixed to a frame, and pressure is applied to the semiconductor units by leaf springs from both sides in the stacking direction. The technology of Patent Document 2 has hysteresis so that the elastic force when the semiconductor unit is deformed in the pressurizing direction is greater than the elastic force when the semiconductor unit is restored in the depressurizing direction. The technique of Patent Document 3 employs a spring unit in which a coil spring is sandwiched between two plates. The technology disclosed in Patent Document 3 is as follows. The spring unit is inserted from the slit provided in the wall of the frame, and the spring unit is compressed while pressing the plate at the tip against the semiconductor unit. When the trailing plate passes through the slit, rotate the plate 90 degrees. When the force pushing the rear end plate is loosened, the rear end plate is caught on both sides of the slit, the spring unit applies pressure to the semiconductor unit, and the semiconductor unit is supported. Both sides of the slit correspond to the stopper portion.

JP 2011-181687 A JP 2007-166819 A JP 2011-200057 A

  In order to increase the adhesion between the semiconductor module and the cooler, the stacked semiconductor units are temporarily applied with a pressure higher than the pressure when they are held on the frame, and the adhesion surface between the semiconductor module and the cooler is adjusted. It is preferable to fix to the frame. For example, paragraph 0024 of Patent Document 1 discloses that a spring is inserted after a high pressure is temporarily applied to stacked semiconductor units. However, in the case of the technique of Patent Document 1, a step of applying a high pressure to the stacked cooling unit, a step of assembling the spring, and a step of biasing the spring are required. On the other hand, when a high pressure is applied to the stacked cooling unit via the spring, the spring may reach the compression limit and deteriorate. In view of such problems, the present specification provides a power conversion device having a structure capable of increasing the manufacturing efficiency of a power conversion device having a stacked semiconductor unit while applying pressure to the stacked cooling unit, and a manufacturing method thereof. To do. In particular, in a power conversion device having a structure in which a semiconductor unit is supported by a frame while applying pressure, a pressure higher than the pressure at which the semiconductor unit is supported by the frame is temporarily applied to the semiconductor module and the cooler. Providing technology to increase the efficiency of the familiarization process.

  One embodiment of the technology disclosed in this specification provides a power conversion device. The power converter includes a semiconductor unit, a frame for attaching the semiconductor unit, and a spring unit. The semiconductor unit is a structure in which flat semiconductor modules containing semiconductor elements and flat coolers are alternately stacked. The frame is provided with a pair of stopper portions (walls or columns), and the semiconductor unit and the spring unit are disposed between the pair of stopper portions. As will be described later, one of the stopper portions may be detachable, and the stopper portion may be attached when the spring unit is attached. The spring unit is a device for fixing the semiconductor unit while applying pressure, and is inserted between one end surface of the semiconductor unit in the stacking direction and a stopper portion facing the end surface. The spring unit is mainly composed of first and second plates and a spring sandwiched between them. Then, at least one of the first plate and the second plate is provided with a protrusion extending toward the other. The height of the projection (the length from the plate surface to the tip of the projection) is such that the gap formed between the plates when the tip of the projection contacts the other plate is larger than the compression limit length of the spring. It is stipulated in.

  Said power converter device can be assembled according to the following process. First, the semiconductor unit is placed on the frame (first step). More specifically, the semiconductor unit is disposed on the frame such that the other end surface faces the first stopper portion provided on the frame. Next, the first plate of the spring unit is applied to one end surface in the stacking direction of the semiconductor units, and a load is applied to the spring unit until the tip of the protrusion comes into contact with the other plate (second step). At this time, since the first and second plates are in contact with each other through the protrusion, a high pressure that cannot be obtained through the spring can be applied to the semiconductor unit. Next, the load applied to the spring unit is removed, and the second plate is brought into contact with a stopper portion (second stopper portion) facing one end surface of the semiconductor unit (third step). When the load is removed, the second plate is pushed back by the reaction force of the spring and comes into contact with the stopper portion so that the column receives the reaction force of the spring unit. Then, the spring unit applies pressure to the semiconductor unit, and the semiconductor unit is supported. In the above spring unit, before the spring reaches the compression limit length, the two plates abut against each other through the protrusion. When a load is applied from the second plate side, a high pressure can be applied to the semiconductor unit without depending on the elastic force of the spring after the two plates abut. In addition, since the first and second plates are in contact with each other, the pressure can be applied stably as compared with the case where the pressure is applied via the spring. After the semiconductor module and the cooler are brought into close contact with each other, the semiconductor unit can be supported by the reaction force of the spring unit by removing the load. That is, the temporary high pressurization and attachment work to the semiconductor unit is completed with one stroke of pushing the spring unit and then releasing the pressed load. The power conversion device improves the efficiency of the process of assembling the semiconductor unit to the frame.

  The first and second stopper portions sandwiching the semiconductor unit may be fixed to the frame in advance, or may be detachable. When the stopper unit (second stopper unit) on the spring unit side is a detachable column, the column is preferably attached to the frame after the second step and before the third step. When the second stopper is previously attached to the frame, the spring unit is compressed and then inserted between the second stopper portion and the semiconductor unit. If the second stopper portion is removed during the second step, the spring unit can be pushed in from the natural length state without being disturbed by the support column. The assembly process can be further simplified.

  The spring may be a leaf spring, but is preferably a coil spring. The coil spring can have a small spring constant as compared with the leaf spring, and the change in load is small with respect to the change in length. Therefore, the distance between the plates can be set wide, and the degree of freedom in designing the height of the protrusion is increased.

  Details of the technology disclosed in this specification and further improvements will be described in the embodiments of the present invention.

It is a perspective view of the power converter device of 1st Example. It is a top view of a power converter device. It is an exploded view of a spring unit. It is a figure which shows the state which has arrange | positioned the spring unit and the semiconductor unit on the flame | frame (before a spring unit compression). It is a figure which shows the state which has arrange | positioned the spring unit and the semiconductor unit to the flame | frame (after compression of a spring unit). It is a figure which shows the state which has arrange | positioned the spring unit and the semiconductor unit to the flame | frame (after insertion of support | pillar). It is a figure which shows the state which has arrange | positioned the spring unit and the semiconductor unit to the flame | frame (after jig | tool retreat). It is a perspective view which shows the modification of the power converter device of 1st Example. It is a perspective view of the power converter device of 2nd Example. It is a perspective view of the power converter device of 3rd Example.

  (First Embodiment) A power converter according to a first embodiment will be described with reference to the drawings. The power converter of an embodiment is an inverter that supplies AC power to a motor. FIG. 1 is a perspective view of the inverter 2. In addition to the semiconductor unit 20 in which power semiconductor elements such as IGBTs and diodes used in the switching circuit are integrated, the inverter 2 is a circuit (substrate) that generates PWM for driving the power semiconductor elements, and smoothes the current. However, the illustration and description thereof are omitted.

  Since power semiconductor elements generate a large amount of heat, they are integrated separately from other circuit components and cooled in a concentrated manner. The inverter 2 uses a large number of power semiconductor elements. A large number of power semiconductor elements are packaged in flat resin molds. Each package is referred to as a semiconductor module 22. The semiconductor unit 20 has a structure in which semiconductor modules 22 and flat plate coolers 21 are alternately stacked. In the semiconductor unit 20 of FIG. 1, six semiconductor modules 22 and seven coolers 21 are alternately stacked. The coolers 21 are connected to each other by pipes. The refrigerant supplied from the supply pipe 24 passes through each cooler 21 and is discharged from the discharge pipe 23. Each semiconductor module 22 sandwiched between both sides by the cooler 21 is efficiently cooled from both sides. Moreover, since the semiconductor unit 20 is a laminated structure, it is excellent in compactness. In FIG. 1, the electrodes of elements extending from the semiconductor module 22 are not shown.

  The semiconductor unit 20 is supported by the frame 12. The frame 12 is further housed in a case (not shown) of the inverter 2. The semiconductor unit 20 is supported between the opposing wall 14 provided on the frame 12 and the column 13. In addition, the support | pillar 13 can be attached or detached. The support | pillar 13 shown with the virtual line of FIG. 1 has shown the removed support | pillar. The spring unit 6 is used to support the semiconductor unit 20. The spring unit 6 is disposed between one end surface 20 a of the semiconductor unit 20 and the support column 13 of the frame 12, and applies pressure in the stacking direction to the semiconductor unit 20. The other end surface 20 b of the semiconductor unit 20 is in contact with the wall 14 fixed to the frame 12.

  The semiconductor module 22 and the cooler 21 before being assembled into a semiconductor unit have slight undulations on their surfaces. Therefore, the degree of adhesion between the semiconductor module 22 and the cooler 21 is not so high simply by stacking the semiconductor module 22 and the cooler 21. Therefore, by applying pressure in the stacking direction to the semiconductor unit 20, the adhesion between the semiconductor module 22 and the cooler 21 is increased, and the thermal conductivity from the semiconductor module 22 to the cooler 21 is improved. Before the semiconductor unit 20 is fixed to the frame 12, it is desirable to apply a high pressure temporarily so that the contact surfaces of the plurality of stacked semiconductor modules 22 and the cooler 21 are well adapted. By temporarily applying a high pressure, the undulations on the surfaces of the semiconductor module 22 and the cooler 21 can be reduced, and the degree of adhesion can be increased. Furthermore, grease is applied to the contact surface between the semiconductor module 22 and the cooler 21, and the familiarity between the grease and the contact surface is improved by temporarily applying a high pressure to the semiconductor unit 20. The spring unit 6 has a function of temporarily applying a high pressure to the semiconductor unit 20 and a function of holding the semiconductor unit 20 at a pressure lower than that pressure. Here, the temporary “high pressure” means that the pressure between the wall 14 and the column 13 is higher than the pressure when the spring unit 6 supports the semiconductor unit 20.

  The spring unit 6 will be described. FIG. 2 shows a plan view of the inverter 2, and FIG. 3 shows an exploded perspective view of the spring unit 6. The spring unit 6 includes two plates (the first plate 5 and the second plate 4) and a coil spring 3 sandwiched between the two plates. The first plate 5 comes into contact with one end face 20 a of the semiconductor unit 20, and the second plate 4 comes into contact with the column 13 attached to the frame 12. In addition, the code | symbol 15 of FIG. 2 has shown the insertion hole which attaches the support | pillar 13. A flange 5a is provided on the back surface of the first plate 5 (the surface facing the second plate 4). A flange 4a is provided on the back surface of the second plate 4 (the surface facing the first plate 5). The flanges 4a and 5a extend toward the other plate. When pressure is applied to the spring unit 6 to contract it, the tip of the flange 5a and the tip of the flange 4a come into contact with each other, and the spring unit 6 cannot be further contracted. The size of the gap between the first plate 5 and the second plate 4 at that time is larger than the compression limit length of the spring 3. In other words, the height of the flanges 5a and 4a (the length from the plate surface to the flange tip) is determined between the two plates when the spring unit 6 is compressed and the tips of both flanges abut against the other plate. Is defined to be larger than the compression limit length of the spring 3. As shown in FIG. 3, when the heights of both flanges are 1 / (2H), the distance between the plates when the flange tips contact each other is H, and the length H is the compression of the spring 3. Greater than the limit length. The compression limit length of the spring 3 means the minimum length of the spring that can be realized when the spring is compressed alone.

  Since the first plate 5 and the second plate 4 have the flanges 5a and 4a described above, the inverter 2 has the following advantages. With the semiconductor unit 20 and the spring unit 6 set on the frame 12, pressure is applied to the spring unit 6 from the second plate 4 side. The spring unit 6 shrinks, but when the flanges come into contact with each other, the spring unit 6 does not shrink any further. When a higher pressure is applied from the second plate 4 side, the first plate 5 and the second plate 4 come into contact with each other via the flange, thereby applying pressure to the semiconductor unit 20 regardless of the rigidity of the spring 3. Can do. Since the flanges are in contact with each other and the first plate 5 and the second plate 4 behave as an integral rigid body, a high pressure can be applied to the semiconductor unit 20, and an unreasonable force is not applied to the spring 3. There is no horizontal displacement. By applying a pressure via the flanges 4a and 5a that are in contact with each other, a uniform pressure can be stably applied to the semiconductor unit 20. Next, a manufacturing process of the inverter 2 (an assembly process of the semiconductor unit 20) will be described.

  (First Step) This step is a step in which the semiconductor unit 20 is arranged on the frame 12 so that one end face 20b of the semiconductor unit 20 faces the wall 14 provided on the frame 12. FIG. 4 shows a plan view of a state where the semiconductor unit 20 is placed. In FIG. 4, the spring unit 6 is also shown. The spring unit 6 is disposed so that the first plate 5 contacts the other end surface 20 a of the semiconductor unit 20. Reference numeral 30 in FIG. 4 indicates a jig for applying a load to the spring unit 6. Reference numeral 15 denotes an insertion hole for attaching the column 13 (see FIG. 1). That is, the column 13 is not yet attached to the frame 12 at this stage.

  (Second Step) In this step, the first plate 5 of the spring unit 6 is applied to the other end surface 20a of the semiconductor unit 20, and the load (pressure) is applied to the spring unit 6 from the second plate 4 side until the flanges come into contact with each other. It is a process of adding. This state is shown in FIG. A load is applied to the spring unit 6 from the second plate 4 side using the jig 30. More specifically, the jig 30 is used to apply a load that is larger than the load at the time when the flanges contact each other. Note that the gap H between the first plate 5 and the second plate 4 when the flange comes into contact is larger than the compression limit length of the spring 3. When the flanges come into contact with each other, the load applied by the jig 30 is directly transmitted to the semiconductor unit 20 through the flanges 5a and 4a regardless of the rigidity of the spring 3. Therefore, the pressure can be stably applied to the semiconductor unit 20, and the close contact surfaces of the stacked semiconductor modules 22 and the cooler 21 can be adjusted. At this time, as shown in FIG. 5, a load (pressure) is applied to the spring unit 6 until the second plate 4 is closer to the semiconductor unit 20 than to the column insertion hole 15. After sufficient pressure is applied to the semiconductor unit 20, the column 13 is fixed to the column insertion hole 15. FIG. 6 shows a state where the support is inserted. At this stage, the jig 30 continues to apply a load, and the second plate 4 of the spring unit 6 is not in contact with the support 13.

  (Third step) This step is a step of removing the load applied to the spring unit 6 by the jig 30. As shown in FIG. 7, the jig 30 is moved backward to remove the load. By removing the load, the second plate 4 is pushed back by the spring 3 and comes into contact with the support 13. That is, the column 13 receives the reaction force of the spring unit 6. On the first plate 5 side, the spring unit 6 applies a predetermined pressure to the semiconductor unit 20. That is, the semiconductor unit 20 is supported between the wall 14 fixed to the frame 12 and the support column 13 while being pressurized by the spring unit 6.

  The advantages of the above manufacturing method will be described. As described in relation to the second step, the spring unit 6 having the flanges 5a and 4a can stably apply a higher pressure to the semiconductor unit 20 than when the spring 3 is interposed. Moreover, the movement of the jig 30 is only a simple movement of first moving toward the spring unit 6 and then moving backward. By simply moving the jig 30 forward and backward, a high pressure can be temporarily applied to the semiconductor unit 20, and then the semiconductor unit 20 can be supported by the spring unit 6 between the column 13 and the wall 14. By introducing the above manufacturing process, the efficiency of the manufacturing process of the inverter (power converter) having the stacked semiconductor units 20 can be improved. That is, the jig 30 for applying a load to the spring unit 6 and then removing the load later only needs to be able to advance and retreat, so that the manufacturing equipment including the equipment for moving the jig 30 back and forth can be simplified. In particular, the step of attaching the support column 13 that supports the semiconductor unit 20 after the second plate 4 of the spring unit 6 is pushed in contributes to the improvement of the manufacturing process of the spring unit 6.

  Next, a modification of the inverter 2 will be described. FIG. 8 is a perspective view of an inverter 2a according to a modification. The structure of the inverter 2a is the same as that of the inverter 2. This modification is characterized in that struts having different diameters are prepared and appropriately selected according to the situation.

  Since the semiconductor unit 20 is formed by stacking the semiconductor module 22 and the cooler 21 in multiple layers, the length in the stacking direction varies. Therefore, in this modified example, a plurality of support columns 13 a, 13 b, and 13 c having different diameters are prepared as the support columns 13. The diameter D2 of the column 13b is larger than the diameter D1 of the column 13a, and the diameter D3 of the column 13c is larger than the diameter D2 of the column 13b. A plurality of struts having different diameters are prepared, and a strut having an appropriate diameter is selected according to the distance between the end surface 20a of the semiconductor unit 20 and the strut insertion hole 15 in the manufacturing process. Then, when manufacturing a large number of inverters, the variation in the distance between the first plate 5 and the second plate 4 when attached to the frame 12 can be made smaller than the variation in the length of the semiconductor unit 20. it can. That is, variation in the length of the spring 3 when the semiconductor unit 20 and the spring unit 6 are set on the frame 12 can be suppressed. When manufacturing a large number of inverters, the length of the spring 3, that is, the variation in pressure applied to the semiconductor unit 20 can be reduced.

  (Second Embodiment) Next, an inverter according to a second embodiment will be described. FIG. 9 shows a perspective view of an inverter 2b (power converter) of the second embodiment. In the inverter 2b, the case 112 corresponds to a frame. Three slits 112 a and 112 b are provided on one side surface of the case 112. The two slits 112b are slits for guiding the refrigerant supply pipe 24 and the discharge pipe 23 to the outside of the case. The slit 112a is provided for passing a jig 30 for applying a load to the spring unit 6 when the semiconductor unit 20 is attached (see FIGS. 4 to 7). Other configurations are the same as those of the inverter 2 of the first embodiment. As shown in the present embodiment, the technology disclosed in this specification can also be applied to a power conversion device in which a frame is a case.

  (Third Embodiment) Next, an inverter 2c according to a third embodiment will be described. FIG. 10 is a perspective view of the inverter 2c. In the inverter 2c of the third embodiment, the first plate 105 and the second plate 104 of the spring unit 106 are simple flat plates and do not have a flange. The inverter 2c of the third embodiment is not suitable for stably applying a high pressure to the spring unit 6 with the flanges in contact with each other, but can simplify the manufacturing process. The inverter 2 c includes a semiconductor unit 20 in which semiconductor modules 22 and a cooler 21 are alternately stacked, a spring unit 106, and a frame 12. The frame 12 is provided with a wall 14 on one side in the stacking direction of the position where the semiconductor unit 20 is to be attached, and a pillar insertion hole 15 for attaching the pillar 13 on the other side. One end surface 20 b of the semiconductor unit 20 is in contact with the wall 14, and a spring unit 106 is inserted between the other end surface 20 a and the column 13 and supported by the frame 12. The manufacturing method of the inverter 2c is as follows.

  (First Step) This step is a step of placing the semiconductor unit 20 on the frame so that one end face 20b of the semiconductor unit 20 faces the wall 14 provided on the frame 12. This step is the same as the first step in the first embodiment.

  (Second Step) In this step, the first plate 105 of the spring unit 106 is applied to the other end surface 20a of the semiconductor unit 20, and the spring unit 106 is moved until the second plate 104 is closer to the semiconductor unit 20 than the column insertion hole 15 is. This is a step of applying a load. In the second step in the first embodiment, a load is applied until the flange comes into contact. However, in this embodiment, since the flange is not provided, the first implementation is performed in that “the load is not applied until the flange comes into contact”. Different from the example.

  (Third step) In this step, the column 13 is attached to the column insertion hole 15, the load applied to the spring unit 106 is removed, and the second plate is placed on the column 13 facing the other surface 20a of the semiconductor unit 20. This is a step of bringing 104 into contact. Here, a plurality of columns 13a, 13b, and 13c having different diameters are prepared, and columns having appropriate diameters are selected and attached according to the distance between the column insertion hole 15 and the end surface 20a of the semiconductor unit 20.

  In the manufacturing process of the third embodiment, since the load is applied to the spring unit 106 before the support 13 is attached, the support 13 is not obstructed. In other words, a jig that applies a load to the spring unit 106 and later removes it (the jig 30 of the first embodiment) may be a facility that simply advances and retreats. Therefore, the production facility can be simplified.

  Points to be noted regarding the technology of the embodiment will be described. The springs constituting the spring units 6 and 106 may be leaf springs, but are preferably coil springs. The coil spring is easy to realize low rigidity. By selecting a low-rigidity spring, the variation in the spring reaction force can be reduced with respect to the variation in the gap width between the support column and the semiconductor unit (that is, the variation in the length of the spring when supporting the semiconductor unit). it can.

  The column 13 and the wall 14 of the embodiment correspond to an example of a stopper portion provided on the frame. The wall 14 corresponds to an example of a first stopper portion that is fixed to the frame 12, and the column 13 corresponds to an example of a second stopper portion that is detachably attached to the frame 12.

  The flanges 4a and 5a correspond to an example of protrusions provided on the plate. In the embodiment, the flange is provided on both the first and second plates, but the flange may be provided only on one of the plates. Whether the flange is provided on one plate or both plates, the distance between the first plate and the second plate when the flanges are in contact with each other only needs to be longer than the compression limit length of the spring.

  The same effect can be obtained by providing a columnar protrusion on the back surface of the first plate or the second plate instead of the flange. Note that the length of the columnar protrusion may be determined by the same technical idea as that of the flange. The protrusions may be provided on both sides of the spring as in the flange of the embodiment. It is because it stabilizes when the tip of the projection comes into contact with the other plate.

  The inverters 2, 2a, 2b, and 2c in the embodiment correspond to an example of a power conversion device. The technology disclosed in the present specification can also be applied to another type of power conversion device such as a voltage converter.

  Specific examples of the present invention have been described in detail above, but these are merely examples and do not limit the scope of the claims. The technology described in the claims includes various modifications and changes of the specific examples illustrated above. The technical elements described in this specification or the drawings exhibit technical usefulness alone or in various combinations, and are not limited to the combinations described in the claims at the time of filing. In addition, the technology exemplified in this specification or the drawings can achieve a plurality of objects at the same time, and has technical usefulness by achieving one of the objects.

2, 2a, 2b, 2c: inverter 3: coil spring 4, 104: second plate 4a, 5a: flange 5, 105: first plate 6, 106: spring unit 12: frames 13, 13a, 13b, 13c: support 14: Wall 15: Supporting hole 20: Semiconductor unit 21: Cooler 22: Semiconductor module 23: Discharge pipe 24: Supply pipe 30: Jig 112: Case

Claims (4)

A semiconductor unit in which flat-plate semiconductor modules containing semiconductor elements and flat-plate coolers are alternately stacked;
A frame for mounting the semiconductor unit;
A spring unit that is inserted between one end surface in the stacking direction of the semiconductor unit and a stopper portion provided in the frame and applies pressure to the semiconductor unit;
The spring unit is
A first and second plate, and a spring sandwiched between them;
A protrusion extending toward the other is provided on at least one of the first plate and the second plate, and a gap formed between the two plates when the tip of the protrusion contacts the other plate is a spring. A power converter characterized in that the height of the protrusion is determined so as to be larger than the compression limit length.   The power converter according to claim 1, wherein the spring is a coil spring. It is a manufacturing method of the power converter according to claim 1 or 2,
A first step of placing the semiconductor unit on the frame;
A second step in which a first plate of the spring unit is applied to one end face of the semiconductor unit, and a load is applied to the spring unit until the tip of the protrusion comes into contact with the other plate;
And a third step of removing the load applied to the spring unit and bringing the second plate into contact with the stopper portion.   The manufacturing method according to claim 3, wherein the stopper portion is a column that is detachably attached to the frame, and the column is attached to the frame after the second step and before the third step.

Images