Convolutional neural network-based semiconductor device heat loss prediction system

Abstract

The invention relates to the technical field of semiconductor testing, in particular to a semiconductor device heat loss prediction system based on a convolutional neural network, which comprises a test case and a test platform arranged in the test case, a probe is preassembled on the test platform, and a cold plate is arranged on the test platform. The semiconductor device body is positioned and clamped by the clamping head on the mounting side plate, and the device endpoint of the semiconductor device body is accurately heated through the heating channel, so that the test temperature of an important structural layer in the semiconductor device body is obtained. The first driving motor and the second driving motor are used for driving the second sliding blocks on the two sides to move in the opposite or opposite direction, so that multi-contact heat loss prediction is carried out on the semiconductor device body, and errors existing in acquisition of temperature sensitive parameters of an important structural layer in the semiconductor device are avoided; therefore, real thermal simulation model experiment data can be smoothly generated.

Description

Technical Field

[0001] The present invention relates to the field of semiconductor testing technology, and in particular to a semiconductor device heat loss prediction system based on a convolutional neural network. Background Art

[0002] Semiconductor devices are mainly used in motor drives or AC / DC converters. Continuous motor drive operation will eventually affect the reliability of semiconductor devices, such as the breakage of the power transmission or steering system on electric vehicles. Existing technologies usually use power cycle testers to specifically test or evaluate the thermal stability and reliability of power semiconductor devices. These tests are crucial to ensure that electronic equipment can withstand thermal stress during long-term operation and prevent thermal failure.

[0003] For example, the Chinese patent with application number CN202410060016.5 discloses a semiconductor tester, including: a tester body, and also including: a needle loading mechanism, the needle loading mechanism is arranged at the output end of the test socket, and a plurality of required probes are pre-installed on each test socket through the needle loading mechanism; a prediction mechanism, the prediction mechanism is arranged on the needle loading mechanism, and the prediction mechanism pre-detects whether the probe is normal before the probe of the needle loading mechanism contacts the chip conductor; a probe connection mechanism, the probe connection mechanism is arranged in the needle loading mechanism, and the abnormality of the probe is detected by the prediction mechanism. If the probe body is damaged, the abnormality of the probe body is detected by the prediction mechanism. However, the probe is pre-installed on each test socket in a fixed manner through the needle loading mechanism, and it is impossible to realize the multi-contact heat loss prediction of the semiconductor device under dynamic working conditions, resulting in errors in the acquisition of temperature sensitive parameters of important structural layers inside the semiconductor device, and it is impossible to generate real thermal simulation model experimental data. Summary of the invention

[0004] In view of this, the purpose of the present invention is to propose a semiconductor device heat loss prediction system based on a convolutional neural network to solve the technical problem that the probes in the prior art are pre-installed on each test socket in a fixed manner through a needle mounting mechanism, which cannot realize multi-contact heat loss prediction of semiconductor devices under dynamic working conditions, resulting in errors in obtaining temperature sensitive parameters of important structural layers inside the semiconductor device and inability to generate real thermal simulation model experimental data.

[0005] Based on the above purpose, the present invention provides a semiconductor device heat loss prediction system based on a convolutional neural network, comprising a test chassis and a test platform arranged in the test chassis, wherein a probe is pre-installed on the test platform, a cold plate is arranged on the test platform, and the surface of the cold plate is a plane. The semiconductor device heat loss prediction system also includes:

[0006] A circulating oil cooling box, which is installed on the inner wall of the test box, has a serial interface, and is connected to the cold plate through a circulating water channel;

[0007] A mounting seat, the mounting seat is arranged on the test platform;

[0008] A lifting mounting frame, wherein the lifting mounting frame can be lifted and mounted on the mounting seat through an elastic member;

[0009] A clamping mechanism for clamping a semiconductor device body, wherein the clamping mechanism can be adjusted and mounted on the lifting mounting frame;

[0010] A driving mechanism is used to drive the probe to move in both horizontal and vertical directions.

[0011] Furthermore, the clamping mechanism comprises:

[0012] There are two mounting side panels, each of which is provided with a sliding opening, and a connecting plate is provided between the two mounting side panels. The connecting plate is slidably arranged on the lifting mounting frame, and a positioning groove is provided on the lifting mounting frame. The connecting plate can be adjusted and installed in the positioning groove by a positioning bolt;

[0013] A fixing plate, the fixing plate being mounted on a side end of the mounting side plate;

[0014] A sliding rod, the sliding rod being slidably disposed on the fixing plate;

[0015] A heating channel, the heating channel being installed at the lower end of the sliding rod;

[0016] A spring, wherein the spring is sleeved on the sliding rod, one end of the spring abuts against the lower surface of the fixing plate, and the other end of the spring abuts against the upper surface of the heating channel;

[0017] A clamping head is slidably disposed on both sides of the heating channel, and the clamping head is also slidably disposed in the sliding opening.

[0018] Furthermore, the driving mechanism comprises:

[0019] A substrate, the substrate is fixedly mounted on the test platform;

[0020] A driving motor 1, wherein the driving motor 1 is mounted on a side end surface of the substrate;

[0021] A lead screw 1, one end of which is mounted on the output end of the drive motor 1, and the other end of which is rotatably mounted on the base plate;

[0022] A guide rail 1, wherein the guide rail 1 is detachably mounted on the base plate through a positioning block;

[0023] A lifting plate, the back of which is threadably mounted on the lead screw 1 via a slider 1, and the lifting plate is also slidably mounted on the guide rail 1;

[0024] A second driving motor, wherein the second driving motor is mounted on the lifting plate;

[0025] A second lead screw, one end of which is mounted on the output end of the second drive motor, and the other end of which is rotatably mounted on the lifting plate;

[0026] A second guide rail, wherein the second guide rail is installed on the lifting plate;

[0027] Slider 2, the slider 2 is threadedly mounted on the lead screw 2, the lead screw 2 is also slidably arranged on the guide rail 2, and the probe is mounted on the slider 2.

[0028] Furthermore, the threads on both sides of the lead screw 2 are arranged in opposite directions, and there are two sliders 2. The two sliders 2 are respectively threadedly installed on the opposite threads on both sides of the lead screw 2, so as to drive the probes on the two sliders 2 to move toward or in opposite directions through the driving motor 2.

[0029] Furthermore, a cylinder is hinged on the test box, a box cover is provided on the test box, and an output end of the cylinder is movably connected to an inner wall of the box cover.

[0030] Furthermore, when the output end of the cylinder is in the return stroke contraction state, the box cover is flipped down and closed, and the interior of the test box is in a closed test state.

[0031] Furthermore, when the box cover is in the flipped-down closed state, the box cover will push the lifting mounting frame downward, and the clamping head will be clamped on the semiconductor device body.

[0032] Furthermore, at least two cold plates are provided on the test platform, and surfaces of the two cold plates are in the same plane.

[0033] Furthermore, the heating channel cooperates with the network bridge module of the semiconductor device body, and when the lifting mounting frame is at the lowest point, the network bridge module of the semiconductor device body is connected to the heating channel.

[0034] Furthermore, the test chassis is provided with a test current cable and a voltage measurement line, respectively, and the test current cable and the voltage measurement line are separately inserted into the device terminals of the semiconductor device body.

[0035] The beneficial effects of the present invention are as follows: a semiconductor device heat loss prediction system based on a convolutional neural network of the present invention is adopted. When in use, the semiconductor device body is placed on two cold plates respectively, and the box cover and the test box are closed through the output end control of the cylinder. The inner wall of the box cover presses the lifting mounting frame to move up and down on the mounting seat, and then the semiconductor device body is positioned and clamped by the clamping head on the mounting side plate. At the same time, the device end points of the semiconductor device body are accurately heated through the heating channel to obtain the test temperature of the important structural layer inside the semiconductor device body, and the sliders on both sides are driven by the driving motor one and the driving motor two to move in relative or opposite directions, thereby realizing multi-contact heat loss prediction of the semiconductor device body, avoiding errors in obtaining temperature-sensitive parameters of the important structural layers inside the semiconductor device, and smoothly generating real thermal simulation model experimental data. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 It is a structural schematic diagram of the present invention;

[0038] Figure 2 It is a structural schematic diagram of a part of the structure of the present invention;

[0039] Figure 3 It is a structural schematic diagram of the driving mechanism in the present invention;

[0040] Figure 4 It is a schematic diagram of the assembly of the clamping mechanism, the lifting mounting frame and the cold plate in the present invention;

[0041] Figure 5 It is a structural schematic diagram of the clamping mechanism in the present invention;

[0042] Figure 6 It is a schematic diagram of the assembly of the probe, the second lead screw and the second slider in the present invention.

[0043] The markings in the figure are:

[0044] 1. Test chassis; 2. Test platform; 3. Probe; 4. Cold plate; 5. Circulating oil cooling box; 6. Circulating water channel; 7. Mounting seat; 8. Lifting mounting frame; 9. Mounting side panel; 10. Sliding opening; 11. Connecting plate; 12. Positioning slot; 13. Positioning bolt; 14. Fixing plate; 15. Sliding rod; 16. Heating channel; 17. Spring; 18. Clamping head; 19. Base plate; 20. Drive motor 1; 21. Lead screw 1; 22. Guide rail 1; 23. Positioning block; 24. Lifting plate; 25. Slider 1; 26. Drive motor 2; 27. Lead screw 2; 28. Guide rail 2; 29. ​​Slider 2; 30. Cylinder; 31. Box cover. DETAILED DESCRIPTION

[0045] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0046] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0047] The embodiment of the present invention provides a semiconductor device heat loss prediction system based on convolutional neural network, such as Figure 1-6 As shown, it includes a test chassis 1 and a test platform 2 arranged in the test chassis 1, a probe 3 is pre-installed on the test platform 2, a cold plate 4 is arranged on the test platform 2, and the surface of the cold plate 4 is a plane. The semiconductor device heat loss prediction system also includes:

[0048] A circulating oil cooling box 5, which is installed on the inner wall of the test cabinet 1, has a serial interface and is connected to the cold plate 4 through a circulating water channel 6;

[0049] A mounting seat 7, the mounting seat 7 is arranged on the test platform 2;

[0050] A lifting mounting frame 8, the lifting mounting frame 8 can be lifted and mounted on the mounting seat 7 through an elastic member;

[0051] A clamping mechanism for clamping a semiconductor device body, the clamping mechanism being adjustably mounted on a lifting mounting frame 8;

[0052] A driving mechanism for driving the probe 3 to move in both horizontal and vertical directions.

[0053] In this embodiment, the clamping mechanism includes:

[0054] There are two mounting side plates 9, each of which is provided with a sliding opening 10, and a connecting plate 11 is provided between the two mounting side plates 9. The connecting plate 11 is slidably provided on the lifting mounting frame 8, and a positioning groove 12 is provided on the lifting mounting frame 8. The connecting plate 11 can be adjusted and installed in the positioning groove 12 by means of a positioning bolt 13;

[0055] A fixing plate 14, the fixing plate 14 is mounted on the side end of the mounting side plate 9;

[0056] A sliding rod 15, the sliding rod 15 is slidably disposed on the fixed plate 14;

[0057] A heating channel 16, the heating channel 16 is installed at the lower end of the sliding rod 15;

[0058] A spring 17, wherein the spring 17 is sleeved on the sliding rod 15, one end of the spring 17 abuts against the lower surface of the fixing plate 14, and the other end of the spring 17 abuts against the upper surface of the heating channel 16;

[0059] The clamping head 18 is slidably disposed on both sides of the heating channel 16 , and the clamping head 18 is also slidably disposed in the sliding opening 10 .

[0060] In this embodiment, the driving mechanism includes:

[0061] A substrate 19, the substrate 19 is fixedly mounted on the test platform 2;

[0062] A driving motor 20, the driving motor 20 is mounted on the side end surface of the substrate 19;

[0063] A lead screw 21, one end of which is mounted on the output end of a driving motor 20, and the other end of which is rotatably mounted on a base plate 19;

[0064] A guide rail 22, the guide rail 22 is detachably mounted on the base plate 19 through a positioning block 23;

[0065] A lifting plate 24, the back of which is threadedly mounted on the lead screw 21 through a slider 25, and the lifting plate 24 is also slidably mounted on a guide rail 22;

[0066] A second driving motor 26, the second driving motor 26 is mounted on the lifting plate 24;

[0067] A second lead screw 27, one end of which is mounted on the output end of the second drive motor 26, and the other end of which is rotatably mounted on the lifting plate 24;

[0068] The second guide rail 28 is installed on the lifting plate 24;

[0069] The slider 29 is threadedly mounted on the lead screw 27 , the lead screw 27 is also slidably disposed on the guide rail 28 , and the probe 3 is mounted on the slider 29 .

[0070] In this embodiment, when in use, the semiconductor device body is placed on two cold plates 4 respectively, and the box cover 31 and the test box 1 are closed by controlling the output end of the cylinder 30. The inner wall of the box cover 31 presses the lifting mounting frame 8 to move up and down on the mounting seat 7, and then the semiconductor device body is positioned and clamped by the clamping head 18 on the mounting side plate 9. At the same time, the device end points of the semiconductor device body are accurately heated through the heating channel 16 to obtain the test temperature of the important structural layer inside the semiconductor device body, and the output end of the driving motor 1 20 drives the lead screw 1 21 to rotate, and under the guidance of the guide rail 1 22, the guide rail 2 28 is driven to lift and lower, and the lead screw 27 is driven to rotate by the driving motor 26. Under the guidance of the guide rail 28, the sliders 29 on both sides are driven to move in relative or opposite directions, thereby realizing multi-contact heat loss prediction of the semiconductor device body.

[0071] In this embodiment, the threads on both sides of the lead screw 27 are set in opposite directions, there are two sliders 29, and the two sliders 29 are respectively threadedly installed on the opposite threads on both sides of the lead screw 27, so as to drive the probes 3 on the two sliders 29 to move toward or in opposite directions through the driving motor 26.

[0072] In this embodiment, a cylinder 30 is hinged on the test box 1 , a box cover 31 is provided on the test box 1 , and an output end of the cylinder 30 is movably connected to an inner wall of the box cover 31 .

[0073] In this embodiment, when the output end of the cylinder 30 is in the return stroke contraction state, the box cover 31 is flipped down and closed, and the interior of the test box 1 is in a closed test state.

[0074] In this embodiment, when the box cover 31 is in the flipped-down closed state, the box cover 31 will push the lifting mounting frame 8 to move downward, and the clamping head 18 will clamp on the semiconductor device body.

[0075] In this embodiment, if Figure 1 , Figure 2 As shown, at least two cold plates 4 are provided on the test platform 2 , and the surfaces of the two cold plates 4 are in the same plane.

[0076] In this embodiment, if Figure 4As shown, the heating channel 16 cooperates with the network bridge module of the semiconductor device body. When the lifting mounting frame 8 is at the lowest point, the network bridge module of the semiconductor device body is connected to the heating channel 16.

[0077] In this embodiment, a test current cable and a voltage measurement line are respectively provided on the test chassis 1, and the test current cable and the voltage measurement line are respectively inserted into the device terminals of the semiconductor device body. When a major safety accident occurs, the system will automatically sound an alarm and automatically stop running. The system has a built-in touch-screen computer, which is convenient and simple to operate.

[0078] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention (including the claims) is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0079] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A semiconductor device heat loss prediction system based on a convolutional neural network, comprising a test chassis (1) and a test platform (2) arranged in the test chassis (1), wherein a probe (3) is pre-installed on the test platform (2), characterized in that: A cold plate (4) is arranged on the test platform (2), the surface of the cold plate (4) being a plane, and the semiconductor device heat loss prediction system further comprises: A circulating oil cooling box (5), the circulating oil cooling box (5) being installed on the inner wall of the test machine box (1), the circulating oil cooling box (5) having a serial interface and being connected to the cold plate (4) via a circulating water channel (6); A mounting seat (7), wherein the mounting seat (7) is arranged on the test platform (2); A lifting mounting frame (8), wherein the lifting mounting frame (8) can be lifted and mounted on the mounting seat (7) via an elastic member; A clamping mechanism for clamping a semiconductor device body, the clamping mechanism being adjustably mounted on the lifting mounting frame (8); A driving mechanism for driving the probe (3) to move in both horizontal and vertical directions.

2. A semiconductor device heat loss prediction system based on convolutional neural network according to claim 1, characterized in that: The clamping mechanism comprises: A mounting side plate (9), wherein there are two mounting side plates (9), each of which is provided with a sliding opening (10), a connecting plate (11) is provided between the two mounting side plates (9), the connecting plate (11) is slidably arranged on the lifting mounting frame (8), a positioning groove (12) is provided on the lifting mounting frame (8), and the connecting plate (11) is adjustably installed in the positioning groove (12) by means of a positioning bolt (13); A fixing plate (14), the fixing plate (14) being mounted on a side end of the mounting side plate (9); A sliding rod (15), wherein the sliding rod (15) is slidably disposed on the fixing plate (14); A heating channel (16), wherein the heating channel (16) is installed at the lower end of the sliding rod (15); a spring (17), wherein the spring (17) is sleeved on the sliding rod (15), one end of the spring (17) abuts against the lower surface of the fixing plate (14), and the other end of the spring (17) abuts against the upper surface of the heating channel (16); A clamping head (18), wherein the clamping head (18) is slidably disposed on both sides of the heating channel (16), and the clamping head (18) is also slidably disposed in the sliding opening (10).

3. The semiconductor device heat loss prediction system based on convolutional neural network according to claim 1, characterized in that: The driving mechanism comprises: A substrate (19), wherein the substrate (19) is fixedly mounted on the test platform (2); A driving motor (20), wherein the driving motor (20) is mounted on a side end surface of the substrate (19); A lead screw (21), one end of which is mounted on the output end of the drive motor (20), and the other end of which is rotatably mounted on the base plate (19); A guide rail (22), wherein the guide rail (22) is detachably mounted on the base plate (19) via a positioning block (23); A lifting plate (24), the back of which is threadably mounted on the lead screw (21) via a slider (25), and the lifting plate (24) is also slidably mounted on the guide rail (22); A second driving motor (26), wherein the second driving motor (26) is mounted on the lifting plate (24); A second lead screw (27), one end of which is mounted on the output end of the second drive motor (26), and the other end of which is rotatably mounted on the lifting plate (24); A second guide rail (28), wherein the second guide rail (28) is installed on the lifting plate (24); Slider 2 (29), the slider 2 (29) is threadedly mounted on the lead screw 2 (27), the lead screw 2 (27) is also slidably arranged on the guide rail 2 (28), and the probe (3) is mounted on the slider 2 (29).

4. A semiconductor device heat loss prediction system based on convolutional neural network according to claim 3, characterized in that: The threads on both sides of the lead screw (27) are arranged in opposite directions. There are two sliders (29), and the two sliders (29) are respectively threadedly mounted on the opposite threads on both sides of the lead screw (27) so as to drive the probes (3) on the two sliders (29) to move in opposite directions through the driving motor (26).

5. The semiconductor device heat loss prediction system based on convolutional neural network according to claim 2, characterized in that: The test box (1) is hinged with a cylinder (30), the test box (1) is provided with a box cover (31), and the output end of the cylinder (30) is movably connected to the inner wall of the box cover (31).

6. A semiconductor device heat loss prediction system based on convolutional neural network according to claim 5, characterized in that: When the output end of the cylinder (30) is in a retracted state during the return stroke, the box cover (31) is flipped down and closed, and the interior of the test box (1) is in a closed test state.

7. A semiconductor device heat loss prediction system based on convolutional neural network according to claim 6, characterized in that: When the box cover (31) is in a flipped-down closed state, the box cover (31) pushes the lifting and lowering mounting frame (8) to move downward, and the clamping head (18) is clamped on the semiconductor device body.

8. The semiconductor device heat loss prediction system based on convolutional neural network according to claim 1, characterized in that: At least two cold plates (4) are provided on the test platform (2), and surfaces of the two cold plates (4) are in the same plane.

9. The semiconductor device heat loss prediction system based on convolutional neural network according to claim 2, characterized in that: The heating channel (16) cooperates with the network bridge module of the semiconductor device body; when the lifting mounting frame (8) is at the lowest point, the network bridge module of the semiconductor device body is connected to the heating channel (16).

10. A semiconductor device heat loss prediction based on convolutional neural network according to claim 1 The system is characterized in that The test box (1) is provided with a test current cable and a voltage measurement line, respectively. The test current cable and the voltage measurement line are separately inserted into the device terminals of the semiconductor device body.

Images