Processor, method for processing, and program
The processing device and method convert transient thermal data into a structure function using logarithmic differentiation and Bayesian deconvolution, effectively addressing the challenge of calculating structure functions from thermal measurements, thereby improving thermal design in electronic circuits.
Patent Information
- Application Number
- JP2023109123
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-07-03
- Publication Date
- 2025-01-17
AI Technical Summary
Existing methods do not provide a clear process for deconvolution to accurately calculate the structure function from transient thermal measurements.
A processing device and method that converts transient thermal change data into a structure function through differentiation with respect to logarithmic time, Bayesian deconvolution, and conversion into RC data to derive thermal resistance and capacity pairs, using a Foster and Cauer RC model.
Enables accurate calculation of the structure function from transient thermal data, enhancing thermal design accuracy in electronic circuits and modules.
Smart Images
Figure 2025007603000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a processing device, a processing method, and a program. [Background technology]
[0002] Thermal design is becoming increasingly important in the development of electronic circuits and modules. In thermal design, it is necessary to understand the thermal characteristics of the target product. However, measuring the thermal characteristics accurately takes time and effort.
[0003] Patent Document 1 describes estimating the structure function of a target product through simulation and calculation. [Prior art documents] [Non-patent literature]
[0004] [Non-Patent Document 1] Luo Yafei, "Definition of isothermal surface resistance by transient heat measurement and structure function and its application," Toyama Prefectural University, dissertation, March 17, 2018 Summary of the Invention [Problem to be solved by the invention]
[0005] However, Non-Patent Document 1 does not disclose how to perform the deconvolution calculation required to calculate the structure function from the temperature response of transient thermal measurement.
[0006] The present invention provides a technique for accurately calculating a structure function of an object from data of transient thermal changes. [Means for solving the problem]
[0007] According to one embodiment of the present invention, there are provided the following processing device, processing method, and program.
[0008] [1] A conversion unit that converts transient thermal change data of an object into a structure function of the object, The conversion unit is A first process is carried out to differentiate the transient thermal change data with respect to logarithmic time; A second process is carried out to convert the result of the first process into a thermal resistance spectrum using Bayesian deconvolution; A third process is performed to convert the thermal resistance spectrum into RC data representing a plurality of pairs of thermal resistance and heat capacity; A fourth process is carried out to derive the structure function using the RC data. Processing unit. [2] In the processing device according to [1], The transient thermal change data is a result of a simulation using a finite element method. Processing unit. [3] In the processing device according to [2], A simulation unit that executes the simulation and generates the transient thermal change data is further provided. Processing unit. [4] In the processing device according to any one of [1] to [3], The object has at least a heat source and a heat dissipation member. Processing unit. [5] In the processing device according to [4], The heat source is a semiconductor chip. Processing unit. [6] One or more computers: A first process is carried out to differentiate the transient thermal change data of the object with respect to logarithmic time; A second process is carried out to convert the result of the first process into a thermal resistance spectrum using Bayesian deconvolution; A third process is performed to convert the thermal resistance spectrum into RC data representing a plurality of pairs of thermal resistance and heat capacity; A fourth process is carried out to derive a structure function of the object using the RC data. Processing methods. [7] A computer is caused to function as a conversion means for converting transient thermal change data of an object into a structure function of the object; The conversion means is A first process is carried out to differentiate the transient thermal change data with respect to logarithmic time; A second process is carried out to convert the result of the first process into a thermal resistance spectrum using Bayesian deconvolution; A third process is performed to convert the thermal resistance spectrum into RC data representing a plurality of pairs of thermal resistance and heat capacity; A fourth process is carried out to derive the structure function using the RC data. program. Effect of the Invention
[0009] According to the present invention, it is possible to provide a technique for calculating the structure function of an object with high accuracy from data of transient thermal changes. [Brief description of the drawings]
[0010] [Figure 1] 1 is a diagram showing an overview of a processing apparatus according to an embodiment; [Diagram 2] FIG. 1 is a diagram illustrating a structure of an object. [Diagram 3] FIG. 1 is a diagram showing a Foster-type RC model. [Figure 4] FIG. 1 is a diagram showing a Cauer type RC model. [Diagram 5] FIG. 2 is a diagram illustrating a functional configuration of a processing device according to an embodiment. [Figure 6] FIG. 1 illustrates an example of a computer for implementing the processing device. [Figure 7] 1 is a flowchart illustrating the flow of a processing method according to an embodiment. [Figure 8] FIG. 1 is a diagram showing a model of an object in a simulation. [Figure 9] FIG. 9 is an enlarged view of the area surrounded by the dashed line in FIG. 8. [Figure 10] FIG. 13 shows the thickness of each element in a laminate in a model. [Figure 11] 4 is a table showing physical properties of each element of an object according to an embodiment. [Figure 12]4 is a graph showing estimated transient thermal change data according to the first embodiment. [Figure 13] 4 is a graph showing both estimated transient heat change data and actually measured transient heat change data according to the first embodiment. [Figure 14] 4 is a graph showing logarithmic time derivative data according to Example 1. [Figure 15] 4 is a graph showing a thermal resistance spectrum according to Example 1. [Figure 16] 1 is a graph plotting a number of pairs (RFoster,i, CFoster,i) according to Example 1. [Figure 17] 1 is a graph plotting a number of pairs (RCauer,i, CCauer,i) according to Example 1. [Figure 18] 4 is a graph showing an estimated structure function according to Example 1. [Figure 19] 2 is a graph showing both an actually measured structure function according to Example 1 and an estimated structure function according to Example 1. [Figure 20] 13 is a graph showing both an actually measured structure function according to Example 2 and an estimated structure function according to Example 2. [Figure 21] 13 is a graph showing both an actually measured structure function according to Example 3 and an estimated structure function according to Example 3. [Figure 22] 13 is a graph showing both the measured structure function according to Example 4 and the estimated structure function according to Example 4. [Figure 23] 13 is a graph showing both the measured structure function according to Example 5 and the estimated structure function according to Example 5. [Figure 24] 13 is a graph showing both the measured structure function according to Example 6 and the estimated structure function according to Example 6. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0011] Hereinafter, embodiments of the present invention will be described with reference to the drawings. In all the drawings, the same components are given the same reference numerals and the description will be omitted as appropriate.
[0012] FIG. 1 is a diagram showing an overview of a processing device 10 according to an embodiment. The processing device 10 according to this embodiment includes a conversion unit 130. The conversion unit 130 converts the transient heat change data of the object into a structure function of the object. The conversion unit 130 executes a first process, a second process, a third process, and a fourth process. The first process is a process of differentiating the transient heat change data with respect to logarithmic time. The second process is a process of converting the result of the first process into a thermal resistance spectrum using Bayesian deconvolution. The third process is a process of converting the thermal resistance spectrum into RC data indicating a plurality of pairs of thermal resistance and heat capacity. The fourth process is a process of deriving a structure function using the RC data. The processing device 10 according to the embodiment will be described in detail below.
[0013] <Object> According to the processing device 10 of this embodiment, the structure function of an object can be calculated from the transient thermal change data of the object. The object has at least a heat source and a heat dissipation member. The object is not particularly limited, but may be, for example, an electronic circuit, a semiconductor module, etc. The semiconductor module includes, for example, a heat dissipation member, grease, a substrate, and a semiconductor chip. The semiconductor chip is, for example, a power semiconductor chip. A semiconductor module including a power semiconductor chip is sometimes called a power semiconductor module. Since the power semiconductor chip generates a large amount of heat, it is important to perform thermal design so that the heat is dissipated appropriately. In other words, it is important to accurately derive the structure function of the power semiconductor module.
[0014] 2 is a diagram illustrating the structure of the target object 20. In the example of this figure, the target object 20 includes a semiconductor chip 21, solder 22, an electrode 23, an insulating substrate 24, an electrode 25, grease 26, and a heat dissipation member 27. In the example of this figure, the heat source is the semiconductor chip 21. The electrode is, for example, copper. The insulating substrate is, for example, ceramics. Examples of ceramics include silicon nitride, alumina, and aluminum nitride. However, the configuration of the target object 20 is not limited to this example.
[0015] <Transient thermal change data> The transient heat change data is data that indicates a temperature change of an object in a non-steady process. Specifically, the transient heat change data is data that indicates a temperature change (transient heat response) of a heat source over time when a sudden change ΔP occurs in the amount of heat generated by the heat source. The sudden change ΔP in the amount of heat generated may be an increase or decrease in the amount of heat generated. In other words, the transient heat change data may be data of a heating process or data of a cooling process. The transient heat response depends on the temperature characteristics of the object 20.
[0016] The transient thermal change data may be data in a table or graph showing the relationship between time t and temperature or temperature difference. Here, "temperature" refers to the temperature T(t) of the heat source, and the temperature difference refers to the temperature T(t) of the heat source and the ambient temperature T amb The difference between (T(t)-T amb )
[0017] The transient heat change data may be actual measurement data obtained by transient heat measurement, or may be the result of a simulation. Examples of simulation methods for obtaining the transient heat change data include the finite element method and the finite volume method. When the transient heat change data is the result of a simulation, the structure function of the object 20 can be obtained without performing an actual measurement. When the transient heat change data is actual measurement data, for example, the processing device 10 may be a part of an evaluation device for thermal characteristics. The transient heat change data is obtained as a measurement result performed by the evaluation device, and the processing device 10 can convert it into a structure function and output it.
[0018] <Structural Function> The structure function indicates the relationship between thermal resistance (cumulative thermal resistance) and heat capacity (cumulative heat capacity). The thermal characteristics of the object 20 can be expressed by the structure function. The structure function derived by the conversion unit 130 may be a differential structure function or a cumulative structure function. The differential structure function is obtained by differentiating the cumulative structure function once. An example in which the conversion unit 130 derives the cumulative structure function will be described below.
[0019] The principle of deriving the structure function is explained below. First, a thermal resistance spectrum is calculated from the transient thermal change data. Then, the thermal resistance spectrum is discretized to calculate multiple pairs of thermal resistance and heat capacity in the Foster type RC model. Furthermore, the multiple pairs of thermal resistance and heat capacity in the Foster type RC model are converted into multiple pairs of thermal resistance and heat capacity in the Cauer type RC model. Then, the cumulative value of thermal resistance in the Cauer type RC model and the cumulative value of heat capacity in the Cauer type RC model are derived. This will be explained in detail below.
[0020] FIG. 3 is a diagram showing a Foster-type RC model. First, assuming such a Foster-type RC model, the temperature difference (T(t)-T amb ) [K] is expressed by the following equation (1).
[0021]
number
[0022] Here, P0 is the heat generation amount [W] of the heat source, which corresponds to the above ΔP. i is the time constant [s], and τ i =R i C i holds true (i = 1, 2, n) (n is a positive integer). R i is the thermal resistance [K / W], C i is the heat capacity [J / K]. By performing Taylor expansion, etc., equation (1) can be transformed into equation (2).
[0023]
number
[0024] Here, τ is the time constant [s] and R(ζ) is the thermal resistance spectrum. If z = ln(t) and ζ = ln(τ), then equation (3) can be obtained from equation (2).
[0025]
number
[0026] Equation (3) can be further transformed to obtain equation (4). Furthermore, equations (5) and (6) hold true using the convolution operator (X surrounded by a circle).
[0027]
number
number
number
[0028] Here, w(z) = exp[z - exp(z)].
[0029] Therefore, the transient thermal change data is differentiated with respect to logarithmic time to obtain dT(z) / dz, and the obtained dT(z) / dz is applied to equation (6) to obtain the thermal resistance R(ζ) for each ζ, i.e., the thermal resistance spectrum.
[0030] Incidentally, dT / dz can be obtained using the relationship in the following equation (7).
[0031]
number
[0032] Here, when the deconvolution operation of equation (6) is performed using Bayesian deconvolution, equation (8) is obtained.
[0033]
number
[0034] In addition, w ki =exp[z k -z i -exp(z k -zi )] and R i =R(z i ) R i (m) means the thermal resistance spectrum after m repeated calculations. In addition, the following equations (9) and (10) hold. From equation (8), each value of the thermal resistance spectrum is R i The thermal resistance spectrum R after M repeated calculations is obtained. i (M) is used as R(ζ). Here, M is not particularly limited, but is, for example, 50 to 200. i (0) Elements R0, R1, R i can be set to all 1, for example. However, R i (0) The elements of can be any value other than 0.
[0035]
number
number
[0036] Next, the thermal resistance R in the Foster type RC model Foster and heat capacity C Foster Multiple pairs of (R Foster,i ,C Foster,i ) (i = 1, 2, . . . n) (n is a positive integer). Foster,i ,C Foster,i ) is obtained by discretizing the thermal resistance spectrum. Specifically, multiple sets (R Foster,i ,C Foster,i ) is obtained. In equation (11), ζ is a logarithmic time constant.
[0037]
number
[0038] Next, the thermal resistance R in the Foster-type RC model Foster and heat capacity C Foster Multiple pairs of (R Foster,i ,C Foster,i ) is the thermal resistance R in the Cauer RC model. Cauer and heat capacity C Cauer Multiple pairs of (R Cauer,i ,C Cauer,i ) (i=1,2,···n) (n is a positive integer).
[0039] The transient response in the Foster-type RC model shown in FIG. 3 is expressed by the following equation (12). Here, ΔT(t) is T(t)-T amb means...
[0040]
number
[0041] Taking the Laplace transform of equation (12), we obtain equation (13), where s is a complex number.
[0042]
number
[0043] 4 is a diagram showing the Cauer RC model. The transient response in the Cauer RC model is expressed by the following equation (14).
[0044]
number
[0045] Therefore, from equations (13) and (14), equation (15) is obtained.
[0046]
number
[0047] Using equation (15), the thermal resistance R in the Foster-type RC model is Foster and heat capacity C Foster Multiple pairs of (R Foster,i ,C Foster,i ) is the thermal resistance R in the Cauer RC model. Cauer and heat capacity C Cauer Multiple pairs of (R Cauer,i ,C Cauer,i ) can be transformed into multiple pairs (R Foster,i ,C Foster,i ) and solving equation (15), the thermal resistance R Cauer and heat capacity C Cauer Multiple pairs of (R Cauer,i ,C Cauer,i ) can be derived in order.
[0048] Multiple thermal resistances R Cauer,i Accumulative value of σR Cauer,i (i.e., R Cauer,1 From R Cauer,i The cumulative value up to the point where the heat capacity C Cauer,i The cumulative value of σC Cauer,i (i.e., C Cauer,1 From C Cauer,i The cumulative value up to is calculated using the following formula (17).
[0049]
number
number
[0050] Thus (σR Cauer,i ,σC Cauer,i ) (i = 1, 2, n) (n is a positive integer) is obtained as the structure function.
[0051] 5 is a diagram illustrating a functional configuration of the processing device 10 according to the present embodiment. In the example of this figure, the processing device 10 further includes a simulation unit 110. The simulation unit 110 executes a simulation to generate transient thermal change data. The functions of the processing device 10 will be described in detail below.
[0052] The simulation unit 110 acquires information necessary for the simulation. The information necessary for the simulation includes, for example, information indicating the structure of the object 20, information indicating the physical properties (density, thermal conductivity, specific heat, etc.) of each component of the object 20, information indicating the ambient temperature, initial temperature, and heat generation conditions of the heat source. For example, the simulation unit 110 acquires information necessary for the simulation input by a user to the processing device 10. As another example, the simulation unit 110 may acquire information necessary for the simulation from another device.
[0053] The simulation unit 110 executes a simulation using information required for the simulation. The simulation unit 110 can execute a simulation using, for example, the finite element method or the finite volume method. The simulation unit 110 generates transient thermal change data as a simulation result.
[0054] The conversion unit 130 acquires the transient heat change data generated by the simulation unit 110. However, the processing device 10 does not have to include the simulation unit 110. The conversion unit 130 may acquire the transient heat change data from another device, or may acquire the transient heat change data by reading out the transient heat change data stored in a storage device accessible from the conversion unit 130. The transient heat change data acquired by the conversion unit 130 may be actual measurement data.
[0055] The conversion unit 130 also acquires other information necessary for converting the transient thermal change data of the object into a structure function. For example, the heat value P0 of the heat source. The conversion unit 130 may further acquire a value indicating the number of iterative calculations M in the Bayesian deconvolution. The conversion unit 130 may acquire the value of M input by the user to the processing device 10, for example. The value of M may be fixed and determined in advance. In that case, the conversion unit 130 does not need to acquire the value of M.
[0056] The conversion unit 130 executes the following first to fourth processes.
[0057] <First process> The first process is a process of differentiating the transient heat change data with respect to logarithmic time. Data obtained by differentiating the transient heat change data with respect to logarithmic time is called logarithmic time differential data. The conversion unit 130 can convert the transient heat change data into logarithmic time differential data using the relationship shown in formula (7). That is, the transient heat change data is differentiated once with respect to time, and the obtained data is further multiplied by t to obtain dT / dz. The logarithmic time differential data is data indicating the relationship between time t and dT / dz. Alternatively, the conversion unit 130 may generate data indicating the relationship between logarithmic time z and dT / dz as the logarithmic time differential data. Note that the conversion unit 130 may perform preprocessing before differentiating the transient heat change data with respect to logarithmic time. For example, the conversion unit 130 may perform processing for reducing noise. In cases where the sampling period of the transient heat change data acquired by the conversion unit 130 is not uniform, the conversion unit 130 may perform processing for making the data uniform. Furthermore, the conversion section 130 may perform other processing on the obtained logarithmic time differential data before performing the next second processing.
[0058] <Second Processing> The second process is a process of converting the result of the first process into a thermal resistance spectrum using Bayesian deconvolution. Specifically, the conversion unit 130 performs an operation such as Bayesian deconvolution on the logarithmic time differential data obtained in the first process to obtain a thermal resistance spectrum. The conversion unit 130 derives a thermal resistance spectrum by applying the logarithmic time differential data to equation (8). Applying the logarithmic time differential data to equation (8) corresponds to performing Bayesian deconvolution. The thermal resistance spectrum is data indicating the relationship between a time constant and a thermal resistance. Alternatively, the thermal resistance spectrum may be data indicating the relationship between a logarithmic time constant and a thermal resistance. The conversion unit 130 may perform other processes on the obtained thermal resistance spectrum before performing the next third process.
[0059] <Third Processing> The third process is a process of converting the thermal resistance spectrum into RC data that indicates multiple pairs of thermal resistance and heat capacity. Specifically, the RC data is the thermal resistance R Cauer and heat capacity C Cauer Multiple pairs of (R Cauer,i ,C Cauer,i ) (i=1,2,...n) (n is a positive integer).
[0060] The third process includes a first step and a second step. In the first step, the thermal resistance spectrum obtained in the second process is converted into the thermal resistance R in the Foster-type RC model. Foster and heat capacity C Foster Multiple pairs of (R Foster,i ,C Foster,i )(i=1,2,···n) (n is a positive integer). In the second step, (R Foster,i ,C Foster,i ) into RC data.
[0061] In the first step, the conversion unit 130 converts the thermal resistance spectrum into (R Foster,i ,C Foster,i )
[0062] In the second step, the conversion unit 130 further calculates (R Foster,i ,C Foster,i ), (R Cauer,i ,C Cauer,i ), that is, converting it into RC data. Note that conversion unit 130 may perform other processing on the obtained RC data before performing the following fourth processing.
[0063] <Fourth Process> The fourth process is a process of deriving a structure function using the RC data. The conversion unit 130 derives a structure function for a plurality of thermal resistances R using the RC data obtained in the third process, equation (16), and equation (17). Cauer,i Accumulative value of σR Cauer,i and multiple heat capacities C Cauer,i The cumulative value of σC Cauer,i Thus, the cumulative value σR Cauer,i and cumulative value σC Cauer,i Multiple pairs of (σR Cauer,i ,σC Cauer,i ) (i = 1, 2,..., n) (n is a positive integer), i.e., data indicating the structure function, is obtained.
[0064] The conversion unit 130 outputs data indicating the derived structure function. There is no particular limitation on the method by which the conversion unit 130 outputs the data indicating the structure function. For example, the conversion unit 130 may output the data indicating the structure function to another device, or may store the data in a storage device accessible by the conversion unit 130.
[0065] The following describes the hardware configuration of the processing device 10. Each functional unit of the processing device 10 (the simulation unit 110 and the conversion unit 130) is realized by a combination of hardware and software (e.g., a combination of an electronic circuit and a program for controlling it).
[0066] 6 is a diagram illustrating a computer 1000 for implementing the processing device 10. The computer 1000 is any computer. For example, the computer 1000 is a system on chip (SoC), a personal computer (PC), a server machine, a tablet terminal, or a smartphone. The computer 1000 may be a dedicated computer designed to implement the processing device 10, or may be a general-purpose computer. The processing device 10 may be implemented by one computer 1000, or may be implemented by a combination of multiple computers 1000.
[0067] The computer 1000 includes a bus 1020, a processor 1040, a memory 1060, a storage device 1080, an input / output interface 1100, and a network interface 1120. The bus 1020 is a data transmission path for the processor 1040, the memory 1060, the storage device 1080, the input / output interface 1100, and the network interface 1120 to transmit and receive data to and from each other. However, the method of connecting the processor 1040 and the like to each other is not limited to bus connection. The processor 1040 is various processors such as a central processing unit (CPU), a graphics processing unit (GPU), or a field-programmable gate array (FPGA). The memory 1060 is a main storage device realized using a random access memory (RAM) or the like. The storage device 1080 is an auxiliary storage device realized using a hard disk, a solid state drive (SSD), a memory card, or a read only memory (ROM) or the like.
[0068] The input / output interface 1100 is an interface for connecting the computer 1000 and an input / output device. For example, an input device such as a keyboard and an output device such as a display are connected to the input / output interface 1100. The input / output interface 1100 may be connected to the input device or the output device by wireless connection or by wired connection. For example, the simulation unit 110 can receive input of information required for simulation via the input / output interface 1100.
[0069] The network interface 1120 is an interface for connecting the computer 1000 to a network. This communication network is, for example, a LAN (Local Area Network) or a WAN (Wide Area Network). The network interface 1120 may be connected to the network wirelessly or by wire. The processing device 10 can communicate with other devices via the input / output interface 1100 or the network interface 1120.
[0070] The storage device 1080 stores program modules that realize the various functional components of the processing device 10. The processor 1040 reads these program modules into the memory 1060 and executes them to realize the functions corresponding to the program modules.
[0071] FIG. 7 is a flowchart illustrating the flow of the processing method according to this embodiment. In the processing method according to this embodiment, one or more computers execute a first process S10, execute a second process S20, execute a third process S30, and execute a fourth process S40. The first process S10 is a process of differentiating the transient thermal change data of the object with respect to logarithmic time. The second process S20 is a process of converting the result of the first process into a thermal resistance spectrum using Bayesian deconvolution. The third process S30 is a process of converting the thermal resistance spectrum into RC data indicating a plurality of pairs of thermal resistance and heat capacity. The fourth process S40 is a process of deriving a structure function of the object using the RC data.
[0072] The processing method according to the present embodiment can be executed by the processing apparatus 10 according to the present embodiment.
[0073] Next, the operation and effect of this embodiment will be described. According to this embodiment, the conversion unit 130 converts the result of the first process into a thermal resistance spectrum using Bayesian deconvolution. Therefore, the structure function of the target structure can be calculated with high accuracy from the data of the transient thermal change. EXAMPLES
[0074] The present embodiment will be described in detail below with reference to examples. However, the present embodiment is not limited to the description of these examples.
[0075] <Example 1> The processing method according to the embodiment was used to derive the structure function of the object.
[0076] Fig. 8 is a diagram showing a model of an object in a simulation. Fig. 9 is an enlarged view of the area surrounded by a dashed line in Fig. 8. In the model, the length of each structure is set to one-fourth the length of the actual object. Fig. 10 shows the thickness of each element in the laminate 300 in the model. The size of the heat dissipation member 37 in the model is as shown in Fig. 8.
[0077] The target object according to Example 1 had a structure in which grease 36, electrode 35, insulating substrate 34, electrode 33, solder 32, and semiconductor chip 31 were laminated in this order on heat dissipation member 37. Of these, the portion in which grease 36, electrode 35, insulating substrate 34, electrode 33, solder 32, and semiconductor chip 31 were laminated is called laminate 300. Laminate 300 was placed at a corner of heat dissipation member 37.
[0078] In the actual object of Example 1, the semiconductor chip 31 was a Si-IGBT (Insulated Gate Bipolar Transistor) chip, the grease 36 was a general heat dissipation grease (Grease Oil Compound G-747, manufactured by Shin-Etsu Chemical Co., Ltd.), the solder 32 was Sn-Pb solder, the electrodes 33 and 35 were copper, and the insulating substrate 34 was silicon nitride ceramics.
[0079] FIG. 11 is a table showing the physical properties of each element of the object according to the example. The values shown in FIG. 11 were used in the simulation. In the simulation, the semiconductor chip 31 was used as the heat source, and the internal heat generation P0 was set to 16.56 W. The initial temperatures of the atmosphere and the object were set to 25° C. As convection conditions, the convection heat transfer coefficient at the bottom surface of the heat dissipation member 37 was set to 1000 W / (m 2 ·K).
[0080] Using the above model and conditions, a simulation was performed using the finite element method to obtain estimated transient heat change data. Specifically, the temperature change at the center of the upper part of the semiconductor chip 31 was plotted to obtain the estimated transient heat change data. The obtained estimated transient heat change data is shown in FIG.
[0081] Meanwhile, the transient thermal change of the object according to Example 1 that was actually prepared was measured using a transient thermal resistance evaluation device (T3Ster built-in power cycle tester, model number PWT1500A, manufactured by Siemens) to obtain measured transient thermal change data according to Example 1. In addition, a structure function was also obtained by calculation within this transient thermal resistance evaluation device. This structure function is called the measured structure function according to Example 1.
[0082] 13 is a graph showing both the estimated transient heat change data and the actually measured transient heat change data according to the first embodiment. In this figure, the vertical axis represents the temperature T(t) of the heat source and the ambient temperature T amb The difference between (T(t)-T amb As shown in Figure 13, the simulation enabled accurate estimation of the transient thermal change data.
[0083] The first process described in the embodiment was performed on the estimated transient thermal change data according to Example 1 to obtain logarithmic time differential data according to Example 1. FIG.
[0084] The second process described in the embodiment was performed on the several hourly differential data of Example 1 to obtain the thermal resistance spectrum of Example 1. FIG.
[0085] The first step of the third process described in the embodiment is performed on the thermal resistance spectrum of Example 1, and the thermal resistance R Foster and heat capacity C Foster Multiple pairs of (R Foster,i ,C Foster,i ) was obtained. FIG. 16 shows the results of multiple sets (R Foster,i ,C Foster,i ) is plotted in the graph.
[0086] Furthermore, the second step of the third process described in the embodiment is performed to obtain a plurality of sets (R Foster,i ,C Foster,i ) is the thermal resistance R in the Cauer type RC model. Cauer and heat capacity C Cauer Multiple pairs of (R Cauer,i ,C Cauer,i ) was converted. FIG. 17 shows the multiple sets (R Cauer,i ,C Cauer,i ) is plotted in the graph.
[0087] A plurality of sets (R Cauer,i ,C Cauer,i ) was subjected to the fourth process described in the embodiment to obtain a structure function. This structure function is called an estimated structure function according to Example 1. FIG. 18 is a graph showing the estimated structure function according to Example 1.
[0088] Fig. 19 is a graph showing both the measured structure function according to Example 1 and the estimated structure function according to Example 1. As shown in Fig. 19, in Example 1, the structure function could be estimated with high accuracy.
[0089] <Example 2> In Example 2, the measured structure function and the estimated structure function were obtained in the same manner as in Example 1, except that a high-reliability heat dissipation grease (Grease Oil Compound G-779, manufactured by Shin-Etsu Chemical Co., Ltd.) was used as the grease 36 in the subject instead of a general heat dissipation grease.
[0090] Fig. 20 is a graph showing both the measured structure function according to Example 2 and the estimated structure function according to Example 2. As shown in Fig. 20, the structure function was estimated with high accuracy in Example 2.
[0091] <Example 3> In Example 3, a measured structure function and an estimated structure function were obtained in the same manner as in Example 1, except that high thermal conductive silicon nitride ceramics was used instead of silicon nitride ceramics as the insulating substrate 34 in the target object.
[0092] Fig. 21 is a graph showing both the measured structure function according to Example 3 and the estimated structure function according to Example 3. As shown in Fig. 21, in Example 3, the structure function could be estimated with high accuracy.
[0093] <Example 4> In Example 4, the measured structure function and the estimated structure function were obtained in the same manner as in Example 1, except that a high-reliability heat-dissipating grease (Grease Oil Compound G-779, manufactured by Shin-Etsu Chemical Co., Ltd.) was used as the grease 36 in the subject instead of general heat-dissipating grease, and high thermal conductivity silicon nitride ceramics was used as the insulating substrate 34 instead of silicon nitride ceramics.
[0094] Fig. 22 is a graph showing both the measured structure function according to Example 4 and the estimated structure function according to Example 4. As shown in Fig. 22, the structure function was estimated with high accuracy in Example 4.
[0095] <Example 5> In Example 5, a measured structure function and an estimated structure function were obtained in the same manner as in Example 1, except that alumina ceramics was used instead of silicon nitride ceramics as the insulating substrate 34 in the target object.
[0096] Fig. 23 is a graph showing both the measured structure function according to Example 5 and the estimated structure function according to Example 5. As shown in Fig. 23, in Example 5, the structure function could be estimated with high accuracy.
[0097] <Example 6> In Example 6, the measured structure function and estimated structure function were obtained in the same manner as in Example 1, except that a high-reliability heat dissipation grease (Grease Oil Compound G-779, manufactured by Shin-Etsu Chemical Co., Ltd.) was used as the grease 36 in the subject instead of general heat dissipation grease, and alumina ceramics was used as the insulating substrate 34 instead of silicon nitride ceramics.
[0098] Fig. 24 is a graph showing both the measured structure function according to Example 6 and the estimated structure function according to Example 6. As shown in Fig. 24, the structure function in Example 6 could be estimated with high accuracy.
[0099] <Comparative Example> Using the estimated transient thermal change data according to Example 1, an attempt was made to derive a structure function. However, in the comparative example, in processing the logarithmic time differential data, instead of performing deconvolution using Bayesian deconvolution, deconvolution was performed using a combination of Fourier transform, division, and inverse Fourier transform. As a result, a valid thermal resistance spectrum curve could not be obtained. This is believed to be due to the strong influence of noise on the results of the deconvolution calculation.
[0100] Although the embodiments of the present invention have been described above with reference to the drawings, these are merely examples of the present invention, and various configurations other than those described above may be adopted. For example, the order of steps performed in each embodiment is not limited to the order described. In each embodiment, the order of steps may be changed as long as it does not cause any problems in terms of content. In addition, the above-described embodiments may be combined as long as the contents are not contradictory. [Explanation of symbols]
[0101] 10 Processing equipment 20 Object 21,31 Semiconductor chips 22,32 Solder 23,25,33,35 electrode 24,34 Insulating substrate 26,36 Grease 27,37 Heat dissipation material 110 Simulation Department 130 Conversion unit 300 Laminate 1000 calculator 1020 Bus 1040 Processor 1060 Memory 1080 Storage Device 1100 Input / Output Interface 1120 Network Interface
Claims
1. A conversion unit converts the transient thermal change data of the object into a structure function of the object, The conversion unit is A first process is carried out to differentiate the transient thermal change data with respect to logarithmic time; A second process is carried out to convert the result of the first process into a thermal resistance spectrum using Bayesian deconvolution; A third process is performed to convert the thermal resistance spectrum into RC data representing a plurality of pairs of thermal resistance and heat capacity; A fourth process is carried out to derive the structure function using the RC data. Processing unit.
2. 2. The processing apparatus according to claim 1, The transient thermal change data is a result of a simulation using a finite element method. Processing unit.
3. 3. The processing apparatus according to claim 2, A simulation unit that executes the simulation and generates the transient thermal change data is further provided. Processing unit.
4. 3. The processing apparatus according to claim 1, The object has at least a heat source and a heat dissipation member. Processing unit.
5. 5. The processing apparatus according to claim 4, The heat source is a semiconductor chip. Processing unit.
6. One or more computers A first process is carried out to differentiate the transient thermal change data of the object with respect to logarithmic time; A second process is carried out to convert the result of the first process into a thermal resistance spectrum using Bayesian deconvolution; A third process is performed to convert the thermal resistance spectrum into RC data representing a plurality of pairs of thermal resistance and heat capacity; A fourth process is carried out to derive a structure function of the object using the RC data. Processing methods.
7. causing the computer to function as a conversion means for converting the transient thermal change data of the object into a structure function of the object; The conversion means is A first process is carried out to differentiate the transient thermal change data with respect to logarithmic time; A second process is carried out to convert the result of the first process into a thermal resistance spectrum using Bayesian deconvolution; A third process is performed to convert the thermal resistance spectrum into RC data representing a plurality of pairs of thermal resistance and heat capacity; A fourth process is carried out to derive the structure function using the RC data. program.