Manufacturing method of artificial intelligence visual identification module with three-dimensional heat dissipation structure

Abstract

The invention discloses a manufacturing method of an artificial intelligence visual identification module with a three-dimensional heat dissipation structure, and belongs to the field of semiconductor devices. Firstly, a flexible insulating heat-conducting sheet is prepared through a hot pressing process, and two sides of the flexible insulating heat-conducting sheet are coated with heat-conducting bonding layers to form a three-dimensional heat dissipation structure; and secondly, for an FPGA chip heat source on the middle-layer acceleration module, heat is transmitted to the heat dissipation end of the shell in the three-dimensional direction through a three-dimensional heat dissipation structure. According to the method, by optimizing the chip layout and the heat dissipation path, the heat dissipation problem of the high-power-density chip in a compact space is solved, and the method has the advantages of being light in weight, high in heat dissipation efficiency, high in integration degree and the like.

Description

Technical Field

[0001] The present invention belongs to the field of semiconductor devices and relates to a manufacturing method of an artificial intelligence vision recognition module with a three-dimensional heat dissipation structure. Background Art

[0002] With the improvement of the computing power of artificial intelligence vision recognition chips (such as GPUs, TPUs, FPGAs), their power density has increased significantly, and traditional heat dissipation solutions are difficult to effectively dissipate heat in a compact space. Such chips need to operate at high load for a long time in scenarios such as industrial quality inspection, autonomous driving, and medical imaging, and their heat generation problem has become the core bottleneck restricting the reliability and lifespan of the system. Existing technologies mostly rely on thermal conductive pads between module layers for indirect heat dissipation, but the distance between the chip and the heat dissipation end is far and the space span is large, resulting in excessive local temperature rise.

[0003] Limited by the compact space, the applicability of heat dissipation technologies such as liquid cooling plates is poor inside the module. The IDC report in 2024 pointed out that the annual failure rate of artificial intelligence Internet of Things (AIoT) devices due to insufficient heat dissipation increases by 12%. Therefore, there is an urgent need for a new heat dissipation method that can directly target the chip heat source and adapt to three-dimensional space heat transfer to break through the contradiction between high power density and miniaturized co-design.

[0004] The three-dimensional heat dissipation structure manufactured by the present invention has insulation, flexibility, and high thermal conductivity. The present invention proposes a three-dimensional heat dissipation integration solution directly for the chip. By directly contacting the flexible heat dissipation structure with the chip, the heat conduction path is optimized, and the chip operating temperature is reduced. The method provided by the present invention has the characteristics of simple process, low cost, energy saving, safety, etc. The present invention provides a method for preparing an artificial intelligence vision recognition module with a three-dimensional heat dissipation structure. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to overcome the deficiencies of the prior art, ensure the appropriate operating temperature of the internal chip of the artificial intelligence vision recognition module, and realize the production of an artificial intelligence vision recognition module with good heat dissipation performance.

[0006] The present invention proposes a manufacturing method of an artificial intelligence vision recognition module with a three-dimensional heat dissipation structure, and the detailed manufacturing steps are as follows: A manufacturing method of an artificial intelligence vision recognition module with a three-dimensional heat dissipation structure, characterized in that the main components include: a main control module, an acceleration module, a storage module, a box body, fixed columns, fastening devices, and a three-dimensional heat dissipation structure, and its manufacturing steps include: S01. Preparation of a three-dimensional heat dissipation structure: Cut a heat-conducting layer and an insulating layer with an adhesive layer on one or both sides so that the heat-conducting layer and the insulating layer with the adhesive layer have the same shape, and the area of the heat-conducting layer is smaller than that of the insulating layer with the adhesive layer. Set the insulating layer as A, the heat-conducting layer as B, and the adhesive layer as C. Stack them in any one of the sequences of ACBCA, ACBCACBCA, and ACBCACBCACBCA, and the edge of the heat-conducting layer does not exceed the insulating layer. Obtain an insulating heat-conducting sheet through thermocompression bonding. Screen-print heat-conducting adhesive layers on both sides of the insulating heat-conducting sheet, and protect the heat-conducting adhesive layers with release films or release papers on both sides. Then heat and cure to obtain a three-dimensional heat dissipation structure; S02. Chip processing and integration: Encapsulate the chip, and the surface roughness Ra after encapsulation ranges from 1.0 to 2.0 μm. Then arrange the main control chip, acceleration chip, and storage chip on different modules respectively; S03. Fixing of the main control and acceleration modules: Fix the lower-layer main control module on the fixing posts through fastening devices. Then fix the middle-layer acceleration module on the fixing posts through fastening devices and keep a safe distance from the lower-layer main control module; S04. Installing the three-dimensional heat dissipation structure at the heat source end: Peel off the two release films of the three-dimensional heat dissipation structure made in step S01, and apply pressure to bond it to the heat source chip of the middle-layer acceleration module; S05. Fixing of the storage module: Fix the upper-layer storage module on the fixing posts through fastening devices so that the upper-layer storage module keeps a safe distance from the middle-layer acceleration module; S06. Mounting the three-dimensional heat dissipation structure at the heat dissipation end: Bend the two sides of the three-dimensional heat dissipation structure upward by 180 degrees, apply pressure to bond it above the upper-layer storage module, and place a heat-conducting pad above the bent three-dimensional heat dissipation structure to obtain an assembled visual recognition module; S07. Overall assembly: Assemble the four sides and the bottom of the outer shell, place a heat-conducting pad at the bottom inside the outer shell, and put the visual recognition module assembled in step S06 into the outer shell. Connect the outer shell to the fixing posts, and connect the outer shell cover to the fixing posts and the outer shell respectively so that the outer shell cover and the heat-conducting pad in step S06 are in close contact to obtain an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure.

[0007] Further, it is characterized in that the insulating strength of the insulating layer in step S01 is greater than 10 kV / mm, and it contains one or more of polyester film, polyimide film, polypropylene film, and polycarbonate film, and the thickness range is 0.05 to 0.2 mm. The thickness of the adhesive layer coated on one or both surfaces of the insulating layer in step S01 is 0.02 to 0.05 mm.

[0008] Further, it is characterized in that in the step S01, the heat conduction layer is a graphite film, and the inside of the graphite film contains one or more of metal powder, silicon powder, boron nitride powder, aluminum nitride powder, silicon nitride powder, silicon carbide powder, diamond powder. The shape of the above powders is spherical or ellipsoidal, the particle size is 5-20 μm, the longitudinal thermal conductivity is greater than 6 W / (m·K), the transverse thermal conductivity is greater than 600 W / (m·K), the thickness is 0.01 mm-3 mm, the surface is provided with a large number of through holes, the diameter of the through holes is less than 1 mm, and the density range of the through holes is 5-10 pieces / cm 2 ; the hot pressing temperature in the step S01 is 80-120 °C, the hot pressing pressure is 0.2-2 MPa, and the hot pressing time is 120-300 s; Further, it is characterized in that the thermal conductivity of the thermal conductive adhesive layer in the step S01 is 3-8 W / (m·K), the thickness is 0.5-1 mm, and the peel strength is 4-8 N / mm; the thermal conductive adhesive layer in the step S01 contains one or more of a resin matrix and a thermal conductive filler. The resin matrix includes one or more of acrylic resin and silicone resin. The thermal conductive filler includes one or more of silicon dioxide, aluminum oxide, magnesium oxide, zinc oxide, silicon, boron nitride, aluminum nitride, silicon nitride, silicon carbide, diamond, carbon nanotubes, graphene, graphite nanosheets, and metal particles. The particle size range of the thermal conductive filler is 0.01-50 μm. The heating and curing temperature in the step S02 is 60-120 °C, and the heating time is 10-200 min. The three-dimensional heat dissipation structure in the step S01 is flexible, can be bent by more than 180°, and the thermal conductivity decreases by less than 5% after bending.

[0009] Further, it is characterized in that the main control module in the step S03 has a main control chip, including one or more of a graphics processing unit (GPU), a tensor processing unit (TPU), a digital signal processor (DSP), a system-on-chip (SoC), a microcontroller unit (MCU), and a vision processing unit (VPU). The power density of the main control chip is 5-50 W / cm 2 ; the safety distance in the step S03 is 10-40 mm.

[0010] Further, it is characterized in that the acceleration module in the step S03 has a field programmable gate array (FPGA) chip to accelerate image preprocessing and real-time inference; the high-power consumption units (such as DSP and high-speed transceivers) inside the FPGA chip are designed and distributed at the edge of the chip to accelerate the heat transfer; the power density range of the FPGA chip is 5-30 W / cm 2 .

[0011] Further, it is characterized in that the storage module in the step S05 includes one or more of dynamic random access memory (DRAM), static random access memory (SRAM), Flash memory, programmable ROM (PROM), erasable PROM (EPROM), and electrically erasable PROM (EEPROM), and the safety distance is 10 - 40 mm.

[0012] Further, it is characterized in that the material of the fixing posts in the steps S03, S05, and S07 is one or more of aluminum and its alloys, titanium and its alloys, and carbon fiber composite materials. The materials of the outer shell and the outer shell cover plate in the step S07 are one or more of aluminum and its alloys, titanium and its alloys, and carbon fiber composite materials, and the outer shell and the outer shell cover plate are provided with air cooling or liquid cooling measures on the outside.

[0013] Further, it is characterized in that the thermal conductivity of the thermal conductive pad in the steps S06 and S07 is 3.5 - 8 W / (m·K), the Shore hardness is 30 - 50, and the thickness is 2 - 5 mm.

[0014] An artificial intelligence vision recognition module with a three-dimensional heat dissipation structure, characterized in that the artificial intelligence vision recognition module with a three-dimensional heat dissipation structure is prepared by the above method and has the function of transferring the heat of the middle-layer acceleration module chip in three-dimensional directions to reduce the chip operating temperature. Compared with the prior art, the advantages of the present invention are as follows: (1) For the artificial intelligence vision recognition module prepared by this method, the heat of the FPGA acceleration chip on the middle-layer acceleration module can be transferred to the cold end of the outer shell through the three-dimensional heat dissipation structure, reducing the operating temperature of the FPGA chip and improving the image processing efficiency; (2) For the artificial intelligence vision recognition module prepared by this method, the heat of the main control chip on the lower-layer main control module can be directly transferred to the cold end of the outer shell through the thermal conductive pad, reducing the operating temperature of the main control chip; (3) The method provided by the present invention has the characteristics of simple process and low cost. Description of the Drawings

[0015] Attached Figure 1 is the overall manufacturing flow chart of the present invention.

[0016] Attached Figure 2 is the exploded view of the structure of the present invention.

[0017] Attached Figure 3 is the three-dimensional structure schematic diagram of the present invention in the application scenario.

[0018] Attached Figure 4 is the exploded view of the three-dimensional heat dissipation structure of the present invention.

[0019] Attached Figure 5It is a comparison chart of the operating temperatures of the middle - layer FPGA chips of the examples and the comparative examples under the same power. Detailed implementation manners

[0020] In the following description, only some exemplary embodiments are simply described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the spirit or scope of the present invention. Therefore, the drawings and description are considered to be exemplary in nature rather than restrictive.

[0021] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention and are not used to limit the present invention.

[0022] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. They are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation of the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality" means two or more, unless otherwise specifically defined.

[0023] The following disclosure provides many different embodiments or examples for implementing the present invention. To simplify the disclosure of the present invention, specific examples are described below. Of course, they are only examples and are not intended to limit the present invention. In addition, the present invention may repeat reference numerals and / or reference letters in different examples. This repetition is for the purpose of simplification and clarity and does not itself indicate the relationship between the various embodiments and / or arrangements discussed. In addition, the present invention provides examples of various specific processes and materials, but those of ordinary skill in the art can be aware of the application of other processes and / or the use of other materials.

[0024] Embodiment

[0025] The embodiment specifically elaborates on the application of an artificial - intelligence vision - recognition module with a three - dimensional heat - dissipation structure.

[0026] Refer to the attached Figure 2 With the attachedFigure 3 The fabricated artificial intelligence vision recognition module with a three-dimensional heat dissipation structure mainly includes the following parts: a housing cover plate 101, a thermal pad 102, a three-dimensional heat dissipation structure 103, an upper-layer storage module 104, a middle-layer acceleration module 105, a lower-layer main control module 106, fixing columns 107, fastening devices 108, and a housing 109; Refer to the appendix Figure 4 The fabricated three-dimensional heat dissipation structure mainly includes the following parts: a thermal conductive adhesive layer 201, an insulating layer 202, an adhesive layer 203, and a thermal conductive layer 204; S01. Prepare the three-dimensional heat dissipation structure: Select a polyimide film with a thickness of 0.2 mm and an insulation strength of 12 kV / mm as the insulating layer 202, on the surface of which an epoxy resin adhesive layer 203 with a thickness of 0.02 mm is coated. Select a graphite film with a longitudinal thermal conductivity of 85 W / (m•K) and a transverse thermal conductivity of 800 W / (m•K) as the thermal conductive layer 204. Aluminum nitride spheres with a particle size of 10 μm are added inside the graphite film, and the surface through-hole diameter is 0.8 mm and the density is 8 per cm 2 , Cut the thermal conductive layer 204 into a rectangle of 35×170 mm, and cut the insulating layer 202 with the adhesive layer 203 into a rectangle of 40×180 mm. Set the insulating layer 202 as A, the thermal conductive layer 204 as B, and the adhesive layer 203 as C. Stack them in the order of ACBCA, and the edge of the thermal conductive layer 204 does not exceed the insulating layer 202. Hot press and bond them at 120 °C and 1 Mpa pressure for 300 s to obtain an insulating and thermal conductive sheet. Add alumina powder with a particle size of 5 μm and silicon carbide powder with a particle size of 10 μm to the silicone resin to make a thermal conductive adhesive. Coat the two sides of the insulating and thermal conductive sheet made in step S01 by screen printing to form a thermal conductive adhesive layer 201 with a thickness of 0.5 mm, a thermal conductivity of 5 W / (m·K), and a peel strength of 6 N / mm, and protect the thermal conductive adhesive layer with PET release film on both sides. Then heat and cure at 100 °C for 30 min to obtain the three-dimensional heat dissipation structure 103; S02. Fabricate each module: Solder components such as resistors and capacitors to different modules through the reflow soldering process. Process the upper surface of the FPGA chip with a Ra value in the range of 1.4~1.6 μm, which is directly in contact with the three-dimensional heat dissipation structure 103. Reflow solder the main control chip based on the ARM architecture on the lower-layer main control module 106, the FPGA chip on the middle-layer acceleration module 105, and the EEPROM and DRAM memories on the upper-layer storage module 104. The main control chip and the FPGA chip use AXI interconnection. The size of the three modules is 100×100 mm. Drill a circular limiting hole at each of the four corners of the module, the center of the circle is located on the diagonal of the square, 5 mm away from the boundary line, and the radius of the circle is 2.5 mm; S03, Fixing the main control and acceleration modules: Fix the lower - layer main control module 106 with a main control chip based on the ARM architecture on the surrounding fixing posts 107 through the fastening device 108. Subsequently, fix the middle - layer acceleration module 105 with an FPGA chip on the fixing posts 107 through the fastening device 108, and keep a safety distance of 20 mm from the lower - layer module; S04, Installing the three - dimensional heat - dissipation structure at the heat - source end: Peel off the two - layer PET release film of the three - dimensional heat - dissipation structure 103 made in step S01, and apply a force of 0.1 MPa to bond it to the heat - source FPGA chip in the middle of the middle - layer acceleration module 105. The size of the FPGA chip is 35×35 mm, and the heat - generation power density is 10 W / cm 2 ; S05, Fixing the upper - layer module: Fix the upper - layer storage module 104 with EEPROM and DRAM on the fixing posts 107 through the fastening device 108, and keep a safety distance of 20 mm between the upper - layer storage module 104 and the middle - layer acceleration module 105; S06, Mounting the three - dimensional heat - dissipation structure at the heat - dissipation end: Bend the two sides of the three - dimensional heat - dissipation structure 103 upwards by 180 degrees, apply a force of 0.1 MPa to bond it above the upper - layer storage module 104, and place a heat - conducting pad 102 with a thermal conductivity of 5 W / (m•K), a Shore hardness of 45, and a thickness of 5 mm above the bent three - dimensional heat - dissipation structure 103 to obtain the assembled vision recognition module; S07, Overall assembly: Assemble the four sides and the bottom of the outer shell 109. Place a heat - conducting pad with a thermal conductivity of 8 W / (m•K), a Shore hardness of 45, and a thickness of 3 mm at the inner bottom of the outer shell 109, and put the vision recognition module assembled in step S07 into the outer shell 109. Connect the outer shell 109 to the fixing posts 107, and connect the outer - shell cover plate 101 to the fixing posts 107 and the outer shell 109 respectively, so that the outer - shell cover plate 101 and the heat - conducting pad 102 in step S07 are in close contact to obtain the artificial - intelligence vision recognition module with a three - dimensional heat - dissipation structure.

[0027] When the artificial - intelligence vision recognition module works, the heat generated by the FPGA chip is transferred longitudinally to the heat - conducting layer of the three - dimensional heat - dissipation structure, and then transferred along the heat - conducting layer in the three - dimensional space direction to the upper heat - conducting pad, and then transferred from the heat - conducting pad to the outer - shell cover plate, so as to achieve the heat - dissipation effect. The main - control module is close to the cold end of the outer shell, and a high - thermal - conductivity heat - conducting pad can be directly used for heat transfer to achieve the heat - dissipation effect. The system formed in this way can dissipate heat efficiently in the three - dimensional space, effectively reducing the temperatures of the main - control and acceleration chips.

[0028] Although the present invention has been described with reference to exemplary embodiments, it should be understood that the terms used are descriptive and exemplary rather than restrictive. Since the present invention can be embodied in many forms without departing from the spirit or essence of the invention, it should be understood that the above embodiments are not limited to any of the foregoing details.

[0029] Comparative example Except for not installing the three-dimensional heat dissipation structure and the heat conduction pads on the upper storage module, other structures and processes are the same as those in the embodiment.

Claims

1. A manufacturing method of an artificial intelligence vision recognition module with a three-dimensional heat dissipation structure, characterized in that, The main components include: a main control module, an acceleration module, a storage module, a box body, fixed columns, fastening devices, and a three-dimensional heat dissipation structure. Its manufacturing steps include: S01, preparing the three-dimensional heat dissipation structure: cutting a heat-conducting layer and an insulating layer with an adhesive layer on one or both sides, making the shapes of the heat-conducting layer and the insulating layer with the adhesive layer the same, and the area of the heat-conducting layer being smaller than that of the insulating layer with the adhesive layer. Set the insulating layer as A, the heat-conducting layer as B, and the adhesive layer as C. Stack them in any one of the orders of ACBCA, ACBCACBCA, and ACBCACBCACBCA, and the edge of the heat-conducting layer does not exceed the insulating layer. Obtain an insulating heat-conducting sheet through thermal compression bonding. Screen-print heat-conducting adhesive layers on both sides of the insulating heat-conducting sheet, and protect the heat-conducting adhesive layers with release films or release papers on both sides. Then heat and cure to obtain the three-dimensional heat dissipation structure; S02, chip processing and integration: encapsulate the chip, and the surface roughness Ra range after encapsulation is 1.0~2.0 μm. Then arrange the main control chip, acceleration chip, and storage chip on different modules respectively; S03, fixing the main control and acceleration modules: fix the lower-layer main control module on the fixed column through the fastening device. Then fix the middle-layer acceleration module on the fixed column through the fastening device and keep a safe distance from the lower-layer main control module; S04, installing the three-dimensional heat dissipation structure at the heat source end: peel off the two release films of the three-dimensional heat dissipation structure made in step S01, and apply pressure to bond it on the heat source chip of the middle-layer acceleration module; S05, fixing the storage module: fix the upper-layer storage module on the fixed column through the fastening device, and keep a safe distance between the upper-layer storage module and the middle-layer acceleration module; S06, mounting the three-dimensional heat dissipation structure at the heat dissipation end: bend the two sides of the three-dimensional heat dissipation structure upward by 180 degrees, apply pressure to bond it above the upper-layer storage module, and place a heat-conducting pad above the bent three-dimensional heat dissipation structure to obtain the assembled visual recognition module; S07, overall assembly: assemble the four sides and the bottom surface of the outer shell. Place a heat-conducting pad at the bottom inside the outer shell, and put the visual recognition module assembled in step S06 into the outer shell. Connect the outer shell to the fixed column, and connect the outer shell cover plate to the fixed column and the outer shell respectively, so that the outer shell cover plate and the heat-conducting pad in step S06 are in close contact to obtain an artificial intelligence visual recognition module with a three-dimensional heat dissipation structure.

2. The manufacturing method according to claim 1, characterized in that, The insulating strength of the insulating layer in step S01 is greater than 10 kV / mm, and it contains one or more of polyester film, polyimide film, polypropylene film, and polycarbonate film, and the thickness range is 0.05~0.2 mm. The thickness of the adhesive layer coated on one or both surfaces of the insulating layer in step S01 is 0.02~0.05 mm.

3. The manufacturing method according to claim 1, characterized in that, In the step S01, the heat-conducting layer is a graphite film, and the interior of the graphite film contains one or more of metal powder, silicon powder, boron nitride powder, aluminum nitride powder, silicon nitride powder, silicon carbide powder, and diamond powder. The shape of the above powders is spherical or ellipsoidal, the particle size is 5 - 20 μm, the longitudinal thermal conductivity is greater than 6 W / (m·K), the transverse thermal conductivity is greater than 600 W / (m·K), the thickness is 0.01 mm - 3 mm, the surface is provided with a large number of through holes, the diameter of the through holes is less than 1 mm, and the density range of the through holes is 5 - 10 pieces / cm 2 ; the hot pressing temperature in the step S01 refers to 80 - 120 °C, the hot pressing pressure is 0.2 - 2 MPa, and the hot pressing time is 120 - 300 s.

4. The manufacturing method according to claim 1, characterized in that, The thermal conductivity of the thermal conductive adhesive layer in the S01 step is 3-8 W / (m·K), the thickness is 0.5-1 mm, and the peel strength is 4-8 N / mm; the thermal conductive adhesive layer in the S01 step contains one or more of a resin matrix and thermal conductive fillers. The resin matrix includes one or more of acrylic resin and silicone resin. The thermal conductive fillers include one or more of silica, alumina, magnesia, zinc oxide, silicon, boron nitride, aluminum nitride, silicon nitride, silicon carbide, diamond, carbon nanotubes, graphene, graphite nanosheets, and metal particles. The particle size range of the thermal conductive fillers is 0.01-50 μm. The heating and curing temperature in the S02 step is 60-120 °C, and the heating time is 10-200 min; the three-dimensional heat dissipation structure in the S01 step is flexible, can withstand a 180° bend, and the thermal conductivity decreases by less than 5% after 100 bends.

5. The manufacturing method according to claim 1, wherein The main control module in the step S03 has a main control chip, including one or more of a graphics processing unit (GPU), a tensor processing unit (TPU), a digital signal processor (DSP), a system-on-chip (SoC), a microcontroller unit (MCU), and a vision processing unit (VPU); the power density of the main control chip is 5-50 W / cm 2 ; the safety distance in the step S03 is 10-40 mm.

6. The manufacturing method according to claim 1, wherein The acceleration module in the S03 step has a field programmable gate array (FPGA) chip to accelerate image preprocessing and real-time inference; the high-power units (such as DSP and high-speed transceivers) inside the FPGA chip are designed to be distributed at the edge of the chip to accelerate heat dissipation, and the power density range of the FPGA chip is 5~30W / cm 2 .

7. The manufacturing method according to claim 1, wherein The storage module in the S05 step includes one or more of dynamic random access memory (DRAM), static random access memory (SRAM), Flash memory, programmable ROM (PROM), erasable PROM (EPROM), and electrically erasable PROM (EEPROM), and the safety distance is 10-40 mm.

8. The manufacturing method according to claim 1, wherein The material of the fixing posts in the S03, S05, and S07 steps is one or more of aluminum and its alloys, titanium and its alloys, and carbon fiber composite materials. The materials of the outer shell and the outer shell cover in the S07 step are one or more of aluminum and its alloys, titanium and its alloys, and carbon fiber composite materials, and the outer shell and the outer shell cover are provided with air cooling or liquid cooling measures on the outside.

9. The manufacturing method according to claim 1, characterized in that, The thermal conductivity of the thermal conductive pad in the S06 and S07 steps is 3.5-8 W / (m·K), the Shore hardness is 30-50, and the thickness is 2-5 mm.

10. An artificial intelligence vision recognition module with a three-dimensional heat dissipation structure, characterized in that, The artificial intelligence vision recognition module with a three-dimensional heat dissipation structure is prepared by the method described in any one of claims 1-9, and has the function of transferring the heat of the middle-layer acceleration module chip in three-dimensional directions and reducing the chip operating temperature.

Images