JP2004079914A - Electronic component mounting reliability prediction method and its prediction system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To eliminate the need of preparing an FEM model for every reliability evaluation and to easily and quickly predict the reliability service life of a solder joint inside an electronic appliance. <P>SOLUTION: A method is for predicting the mounting reliability of CSP to a mounting board of the electronic appliance provided with the mounting board and composed by respectively attaching the CSP through a solder bump to both opposing surfaces of the substrate. The method comprises a process of obtaining a distortion amount generated at the solder bump corresponding to position relation between the CSP, and preparing a response surface for three-dimensionally displaying the distortion amount corresponding to the position relation between the CSP beforehand; and a process of arbitrarily setting the position relation between the CSP, applying the set position relation to the response surface, and calculating the distortion amount generated at the solder bump. The reliability service life of the solder bump is easily and quickly predicted. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an electronic component mounting reliability prediction method and a prediction system suitable for application to the reliability evaluation of an electronic device in which a CSP (Chip Size Package) is attached via solder bumps to both surfaces of a mounting board. Things. More specifically, this is a method for predicting the mounting reliability of an electronic component on a double-sided mounting type board. In advance, the amount of distortion generated in a solder joint portion is determined in advance in accordance with the positional relationship between the electronic components, and this distortion is determined. By making a response curve or a response surface that displays the quantity corresponding to the positional relationship between electronic components, the reliability life of solder joints in electronic equipment can be easily and quickly predicted. It is.
[0002]
[Prior art]
2. Description of the Related Art In recent years, electronic components such as semiconductor packages have been miniaturized year by year as electronic devices have become smaller and lighter. Therefore, the part of the semiconductor package that is bonded to the printed circuit board is also reduced in size, and the mounting reliability of the semiconductor package is sufficient for thermal stress caused by the difference in the coefficient of thermal expansion between the semiconductor package and the mounting board. Structural design margins are decreasing.
[0003]
To ensure sufficient reliability against thermal stress, various reliability tests are required. In particular, in the case of a semiconductor package called a CSP, which is designed so that its size is almost the same as that of a silicon chip, thermal bumps are applied to the solder bumps that also serve as electrical connections, so a test that guarantees its reliability is important. . For this reason, each semiconductor manufacturer and a set manufacturer using the same put a test board on which a CSP is mounted in a constant temperature chamber, and perform a reliability assurance test that withstands actual use by moving the temperature up and down.
[0004]
In order to guarantee the reliability of the CSP mounted on the double-sided mounting board, the CSP is mounted on the front and back surfaces of the double-sided mounting type test board in the same manner as in the case described above, and the assurance test is performed. The CSP test to guarantee reliability has two stages. The first stage is the reliability confirmation on the test board, but the second stage is the reliability confirmation test in the same state as the actual product. It is.
[0005]
In the case of double-sided mounting, it is known that the mounting reliability of the CSP changes depending on the arrangement of the front and back CSPs on the mounting board. It is difficult to predict the reliability on the mounting board from the data. Depending on the position and size of the CSP disposed on the back surface, even if the first stage test board is OK, the second stage mounting board often does not pass the reliability test.
[0006]
In this case, the effects of thermal strain are reduced by various design factors such as the material and thickness of the mounting board and the arrangement of the electronic components, and the design is changed to commercialize. However, in recent years, product design cycles have become increasingly shorter, and a design change at the confirmation stage on a mounting board leads to a change in specifications once determined. This prolongs the design period of the product, resulting in an increase in development and manufacturing costs.
[0007]
On the other hand, as a method of estimating the life of a solder joint of a semiconductor package by a numerical calculation method, a prediction method based on a fatigue life equation such as the Coffin-Manson rule is known.
[0008]
This prediction method is based on a finite element method (hereinafter, also referred to as an FEM method) in which an amplitude of plastic strain or creep strain of a solder joint generated due to a difference in thermal expansion coefficient between a substrate and a semiconductor package is determined. This is a method of predicting the reliability life of the solder joint by applying the obtained strain amplitude to the fatigue life formula of the Coffin-Manson rule.
[0009]
In this FEM method, the shape and material characteristics (Young's modulus, Poisson's ratio, linear expansion coefficient, etc.) of each component of the CSP are numerically input to a dedicated tool called a Pre-Post processor, and the FEM as shown in FIG. Create a model 70. Then, based on the FEM model 70, when the temperature moves up and down within a certain range, for example, the shape and material characteristics of each component of the CSPs 50A and 50B are changed to the joints such as the solder bumps 75A and 75B. Calculate the effect.
[0010]
The reliability life of the solder bumps 75A and 75B is determined by the shape of the bumps, the number and arrangement of the bumps, the size and thickness of the built-in silicon chip, the thickness of the entire package, the rigidity and linear expansion coefficient of the mold resin, It depends on the material and thickness of the interposer substrate (IP), the size of the package, and the thickness, rigidity, and coefficient of linear expansion of the mounting substrate.
[0011]
[Problems to be solved by the invention]
By the way, according to the method of estimating the mounting reliability of the CSP according to the conventional example, the shape and material characteristics (Young's modulus, Poisson's ratio, linear expansion coefficient, etc.) of each component of the CSPs 50A and 50B are called a Pre-Post processor. The FEM model 70 was created by inputting numerical values into the dedicated tool.
[0012]
Then, based on the FEM model 70, for example, when the temperature moves up and down within a certain range, the shape and material characteristics of each component of the CSPs 50A and 50B are changed to the joints such as the solder bumps 75A and 75B. The effect was calculated.
[0013]
For this reason, in the mounting reliability prediction method according to the conventional example, if the method of defining the model (for example, the fineness of the FEM model, the boundary condition of the calculation, etc.) differs depending on the operator or the like, the parameter variables are numerically input to the same value. However, there is a problem that a different prediction result is derived.
[0014]
That is, in the FEM method, it is necessary to use a dedicated Pre-Post processor to divide the cross section and surface of the mounting board and the solder bumps with a mesh and construct the FEM model 70 while selecting the nodes of the mesh. There is inconvenience that a correct solution cannot be derived unless a person skilled in operation or a specialized analyst with a certain knowledge operates this Pre-Post processor. Therefore, in the prediction method according to the conventional example, the prediction of the mounting reliability of the CSPs 50A and 50B must always be performed by a specialized engineer who can operate the Pre-Post processor.
[0015]
Further, the above-mentioned FEM method repeatedly calculates how much the nodes of the mesh are displaced by thermal stress (stress) or the like. In particular, the calculation of creep strain is obtained by integrating the strain amount along the elapsed time. Since it is necessary to keep going, the calculation process requires a certain amount of time.
[0016]
For this reason, after determining the layout design (floor layout) of the CSPs 50A and 50B, the parameter variables are input to the Pre-Post processor to create the FEM model 70, and it takes a considerable amount of time to predict the mounting reliability. It takes time, and the relationship between the floor layout of the CSPs 50A and 50B and the mounting reliability cannot be grasped in real time.
[0017]
Further, when changing the floor layout of the CSPs 50A and 50B, it is necessary to re-create the above-mentioned FEM model 70. Therefore, it takes more time to predict the mounting reliability of the CSP (hereinafter, also referred to as electronic components). There was a problem that it was necessary.
[0018]
Therefore, the present invention solves such a problem, and it is not necessary to create an FEM model for every reliability evaluation, and the reliability life of a solder joint in an electronic device can be easily shortened. It is an object of the present invention to provide a method for predicting the mounting reliability of an electronic component and a system for predicting the mounting reliability, which can be quickly and quickly predicted.
[0019]
[Means for Solving the Problems]
The above-described problem is to predict the mounting reliability of the electronic component with respect to the board of an electronic device having a wiring board and having the electronic component attached to both opposing surfaces of the board via solder joints. And a response curve displaying in advance the amount of strain occurring in the solder joint in accordance with the positional relationship between the electronic components and displaying the amount of strain in accordance with the positional relationship between the electronic components. Or a step of preparing a response surface, and a step of arbitrarily setting the positional relationship between the electronic components, and applying the set positional relationship to a response curve or a response surface to calculate a distortion amount generated in the solder joint. The problem is solved by a method for predicting mounting reliability of an electronic component, characterized by having:
[0020]
ADVANTAGE OF THE INVENTION According to the mounting reliability prediction method of the electronic component which concerns on this invention, the trouble which creates the FEM model corresponding to an electronic device for every reliability evaluation can be omitted, the reliability life of a solder joint part is easy, and Can be predicted quickly.
[0021]
An electronic component mounting reliability prediction system according to the present invention is directed to an electronic device having a wiring substrate and having electronic components attached to both opposing surfaces of the substrate via solder joints, with respect to the substrate. A system for predicting mounting reliability of an electronic component, comprising: a model storage unit for storing a response curve or a response surface for displaying an amount of strain generated in the solder joint in correspondence with a positional relationship between the electronic components; A position information storage unit for storing arbitrary positional relationship information between the two, and the positional relationship information stored in the position information storage unit is input to a response curve or a response surface stored in the model storage unit, and the position is stored in the solder joint. And an arithmetic processing unit for calculating the amount of distortion generated.
[0022]
According to the electronic component mounting reliability prediction system according to the present invention, a response curve or a response surface that displays the amount of strain generated at the solder joint portion in correspondence with the positional relationship between the electronic components includes an arbitrary value between the electronic components. The positional relationship information is input, and the amount of distortion generated in the solder joint is calculated.
[0023]
Therefore, compared to the conventional method, it is not necessary to create an FEM model corresponding to the electronic device for each reliability evaluation in order to obtain the amount of distortion of the solder joint in the electronic device. The time required can be greatly reduced.
[0024]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a method and a system for predicting the mounting reliability of an electronic component according to an embodiment of the present invention will be described with reference to the drawings.
[0025]
In this embodiment, the mounting reliability of the electronic component on the substrate is estimated for an electronic device having a wiring substrate and having electronic components attached to opposing surfaces of the substrate via solder joints, respectively. At this time, the amount of distortion generated in the solder joint is determined in advance in accordance with the positional relationship between the electronic components, and the response curve or the response surface displaying the amount of distortion in accordance with the positional relationship between the electronic components. In addition, the FEM model does not have to be created for each reliability evaluation, and the reliability life of the solder joint in the electronic device can be easily and quickly predicted. Things.
[0026]
First, an electronic component mounting reliability prediction system according to an embodiment of the present invention will be described.
[0027]
FIG. 1 is a block diagram illustrating a configuration example of an electronic component mounting reliability prediction system 100 according to an embodiment of the present invention. The mounting reliability prediction system 100 includes a mounting substrate (an example of a wiring substrate), and a semiconductor package (an electronic component such as a CSP) is provided on both opposing surfaces of the substrate via solder bumps (an example of a solder joint). Is a system for predicting the mounting reliability between a mounting board and a CSP in an electronic device to which each is attached.
[0028]
As shown in FIG. 1, the mounting reliability prediction system (hereinafter, also simply referred to as a prediction system) 100 includes a parameter storage unit (an example of a position information storage unit) 11, a model storage unit 13, and a fatigue life equation storage. The unit 15 includes an experiment data storage unit 17, a distortion amount storage unit 19, a reliability data storage unit 21, a CPU 23, a ROM 25, a monitor 27, and the like.
[0029]
The parameter storage unit 11 shown in FIG. 1 stores parameter variables such as the positional relationship between the CSPs attached to both sides of the mounting board, the shape and material characteristics of the CSP, and the shape and material characteristics of the solder bumps (hereinafter simply referred to as parameters). Is also input numerically. FIG. 2 is a table showing an example of parameters stored in the parameter storage unit 11 and their setting ranges. Each parameter shown in FIG. 2 is numerically set within the upper and lower limits, and stored in the parameter storage unit 11. The parameter storage unit 11 is configured by, for example, a RAM (Random Access Memory).
[0030]
The model storage unit 13 stores a response surface that three-dimensionally displays plastic strain, creep strain, and the like generated in the solder bumps in accordance with the positional relationship (ΔX, ΔY) between the CSPs mounted on both surfaces of the mounting board. You.
[0031]
FIG. 3 is a table showing an example of the response surface. The X axis in FIG. 3 is the relative displacement (ΔX) of the center point of the CSP 50B in the X direction with respect to the center point of the CSP 50A. The Y-axis in FIG. 3 is the relative displacement (ΔY) of the center point of the CSP 50B in the Y direction with respect to the center point of the CSP 50A. The Z-axis in FIG. 3 represents the amount of distortion (hereinafter, also referred to as nonlinear distortion amplitude) of the solder bumps when one cycle of a TC (Thermal Cycle) test is performed in a temperature range of −25 ° C. to 125 ° C. is there.
[0032]
The response surface shown in FIG. 3 is represented, for example, by the equation (1) in FIG. In equation (1), the variable X _i , X _j Is a general notation of each parameter. The variable Y is the amount of nonlinear distortion amplitude of the solder bump described above, and β ₀ , Β _i , Β _ii , Β _ij ... are constants according to the parameters. In equation (1), the positional relationship between the CSPs (ΔX, ΔY) is X _i , X _j , The amount of nonlinear distortion amplitude of the solder bump corresponding to the positional relationship is calculated. A method of creating the response surface will be described later.
[0033]
The model storage unit 13 illustrated in FIG. 1 is configured by a large-capacity storage device such as a hard disk (HDD). The distortion amount storage unit 19 shown in FIG. 1 stores the amount of nonlinear distortion amplitude calculated from the response surface. The distortion storage unit 19 is constituted by, for example, a RAM.
[0034]
In FIG. 1, the result of a TC test on an experimental board on which two or more CSPs are mounted on both sides via solder bumps, that is, the solder measured on the experimental board, The non-linear distortion amplitude of the bump and the reliability life value (the number of cycles in the TC test) are stored. The experiment data storage unit 17 is configured by, for example, a RAM.
[0035]
The fatigue life equation storage section 15 stores, for example, a fatigue life equation of the Coffin-Manson rule as shown in equation (2) of FIG. In equation (2), Y is the amount of nonlinear distortion amplitude of the solder bump, and ε ₀ Is the fatigue life characteristic. This fatigue life characteristic ε ₀ Is a value that depends on the shape of each solder bump. N is a thermal cycle life (reliable life value of the solder bump: cycle). When both sides of the equation (2) are displayed in logarithm, a logarithmic graph shown in FIG. 4 can be obtained.
[0036]
The horizontal axis in FIG. 4 is the logarithm of N, and the vertical axis is the logarithm of Y. The slope of the graph (line segment) shown in FIG. 4 differs depending on the material properties of the solder bump. For example, in the case of Sn-Pb eutectic solder, it is about -0.5. Also, the height of this graph (line segment) is the fatigue life characteristic ε. ₀ Fluctuates depending on The fatigue life equation storage unit 15 for storing the equation of the Coffin-Manson rule is, for example, an HDD.
[0037]
By the way, the equation (3) in FIG. 5 is a fatigue life equation based on the Coffin-Manson rule like the equation (2) in FIG. In equation (3), N _R Is the experimental result of the TC test on the experimental substrate, and is the reliability life cycle (cycle) of the solder bump. Also, Y _R Is the experimental result of the TC test on the experimental substrate, and is the amount of nonlinear distortion amplitude of the solder bump. That is, N _R And Y _R Is known data, so from equation (3) ₀ Can be obtained. Since the shape of the solder bumps on the experimental board is the same as that on the mounting board, the ε obtained by equation (3) was used. ₀ Is ε in equation (2) ₀ be equivalent to.
[0038]
Therefore, ε obtained from equation (3) ₀ By substituting the non-linear distortion amplitude Y of the solder bump obtained from the equation (1) (response curved surface) into the equation (2), a reliability life value N of the solder bump on the mounting board is obtained. Can be.
[0039]
The reliability life value N is stored in the reliability data storage unit 21 shown in FIG. The reliability data storage unit 21 is, for example, a RAM. The reliability life value N of the solder bump stored in the reliability data storage unit 21 can be output to the outside of the prediction system 100.
[0040]
The prediction system 100 includes a parameter storage unit 11, a model storage unit 13, a fatigue life equation storage unit 15, an experiment data storage unit 17, a strain amount storage unit 19, and a reliability data storage unit 21. The storage unit main body 60 is configured.
[0041]
The CPU 23 shown in FIG. 1 operates so as to read data and calculation formulas from each storage unit constituting the storage unit main body 60, perform a predetermined calculation process, and write the calculation result to the storage unit main body 60. An operation program for the CPU 23 is stored in a ROM (Read Only Memory) 25. Further, on the monitor 27, each data stored in the storage unit main body 60 is arbitrarily displayed.
[0042]
Next, a method of creating the response surface stored in the model storage unit 13 will be described. The response surface shown in FIG. 3 is composed of many points. The points that make up this response surface are the shape, number and arrangement of solder bumps, the size and thickness of the silicon chip incorporated, the size and thickness of the entire package, the rigidity and linear expansion coefficient of the mold resin, It is calculated by the finite element method (FEM method) based on parameters such as the material and thickness of the interposer and the thickness, rigidity, and coefficient of linear expansion of the mounting board.
[0043]
However, if the number of parameters and the variable range of the response surface are increased unnecessarily because the response surface is required to have high accuracy and high range, the number of sampling of the amount of nonlinear distortion amplitude becomes enormous, and the response surface is created. A long time is required.
[0044]
Therefore, in this embodiment, for example, a characteristic determined by a certain width such as 0.3 to 0.4, such as a Poisson's ratio, is fixed at its average value. In addition, since there are many types of polyimide material or FR-4 material for the actual CSP interposer substrate (IP), the material is also limited to two levels (types). Furthermore, the thickness of the IP is, for example, 0.07 mm for the polyimide material and 0.2 mm and 0.4 mm for the FR-4 material as standard specifications. Is limited to two levels.
[0045]
In addition, among parameters constituting the response surface, parameters with low sensitivity relating to the reliability life of the solder bumps are fixed at an average value, and only parameters with high sensitivity are narrowed down. Thereby, the parameters constituting the response surface can be effectively reduced, and the creation of the response surface can be facilitated.
[0046]
Incidentally, the DOE method (Taguchi method) can be used to narrow down the parameters. For example, FIG. 6 shows the result of a sensitivity analysis called L12 with two levels and seven parameters. The horizontal axis in FIG. 6 is the level (two levels are represented by -1, 1), and the vertical axis is the amount of nonlinear distortion amplitude (distortion amount) of the solder bump.
[0047]
As is clear from FIG. 6, the relative position ΔX and the relative position ΔY between the CSPs have high sensitivity regarding the reliability life of the solder bumps, and the Mold resin thickness is low. Therefore, it can be seen that even if the Mold resin thickness is fixed to the average value of general CSP, there is not much difference in the calculation result. Thereby, it is also possible to exclude the Mold resin thickness from the parameters and to fix it at an average value.
[0048]
In this embodiment, parameters constituting the response surface are narrowed down by the above-described method, and as shown in FIG. 2, the parameters constituting the response surface are converted into the shape and material properties of the mounting board, the shape and the material properties of the CSP, The relative position (positional relationship) between the CSPs was taken. In addition, the upper and lower limits of each parameter are limited within the range of common sense in consideration of material properties, shapes, layouts, and the like that can be actually used.
[0049]
Next, the X- and Y-axes of the response surface are defined as the positional relationship between the CSPs (ΔX, ΔY), and the Z-axis is defined as the non-linear distortion amplitude (distortion) of the solder bumps. Create a response surface by connecting points. According to various combinations of the shape and material characteristics of the mounting substrate and the shape and material characteristics of the CSP, response curved surfaces with the positional relationship (ΔX, ΔY) between the CSPs as the X axis and the Y axis are formed.
[0050]
FIG. 7 is an example in which a response curve (regression curve) is created by a fourth-order polynomial for one variable from an analysis result of each point obtained by the finite element method. The horizontal axis in FIG. 7 is, for example, the relative position ΔX between the CSPs. The vertical axis indicates the amount of nonlinear distortion amplitude of the solder bump. The creation of this polynomial is performed by, for example, the least squares method. When the relative position ΔY between the CSPs is fixed, the nonlinear distortion amplitude of the solder bump can be obtained from this response curve. When the relative position between the CSPs increases to two variables (ΔX, ΔY), this response curve becomes a multidimensional response surface (regression surface).
[0051]
Next, a method for predicting the mounting reliability of an electronic component according to the embodiment of the present invention will be described. FIG. 8 is a flowchart showing a method for predicting the mounting reliability of an electronic component. Here, it is assumed that the mounting reliability of the CSP arranged on both sides of the mounting board is predicted using the above-described prediction system 100.
[0052]
First, at step 1 in FIG. 8, the numerical values of the parameters shown in FIG. FIG. 9 is a conceptual diagram showing an example of an input screen on the monitor 27. The operator inputs numerical values of each parameter shown in FIG. 2 on an input screen as shown in FIG.
[0053]
Based on the positional relationship (ΔX, ΔY) input by numerical values, if the overlapping degree of the CSPs can be displayed in real time on the input screen, it is possible to prevent the input error of the positional relationship (ΔX, ΔY), It is convenient.
[0054]
Next, in step 2 of FIG. 8, it is confirmed whether each value of the parameter numerically input on the input screen is correct. If correct, go to step 3. If correction is necessary, the process returns to step 1 and the numerical values of the parameters are re-input. In step 3 of FIG. 8, a predetermined response surface is called from the model storage unit 13 based on the input parameter values and displayed on the monitor 27. From the response surface displayed on the monitor screen, the relationship between the relative position between the CSPs and the amount of nonlinear distortion amplitude of the solder bump can be roughly grasped.
[0055]
Then, in step 4 in FIG. 8, the CPU 23 applies the positional relationship (.DELTA.X, .DELTA.Y) between the CSPs to the response surface to calculate the amount of nonlinear distortion of the solder bump. The calculated nonlinear distortion amplitude of the solder bump is stored in the distortion storage unit 19.
[0056]
In step 5 of FIG. 8, the amount of nonlinear distortion amplitude and the reliability life value of the solder bumps on the test board, not on the mounting board, are stored in the test data storage unit 17. Then, in step 6 of FIG. 8, the amount of nonlinear distortion amplitude (calculated value) of the solder bump on the mounting board, the amount of nonlinear distortion amplitude of the solder bump on the experimental board (experimental value), and the reliability life (experimental value) ) Is substituted into the fatigue life formula of the Coffin-Manson rule.
[0057]
Thereby, the reliability life value of the solder bump on the experimental board is calculated (see the equations (1) to (3) in FIG. 5). These series of calculation processes are performed by the CPU 23. The calculated reliability life value of the solder bump on the mounting board is stored in the reliability data storage unit 21 and displayed on the monitor 27 (step 7). Thus, the method for estimating the mounting reliability of the electronic component shown in FIG. 8 is completed.
[0058]
As described above, according to the electronic component mounting reliability prediction system 100 according to the embodiment of the present invention, since the reliability life solution is derived using the response surface, there is no need to write an FEM model or the like, and the numerical value is reduced. Since the solution can be obtained immediately after the input, the operator does not need specialized knowledge. Therefore, the reliability design can be operated by an electric designer in charge of the floor layout in spite of the mechanical field, and can be used for floor layout planning.
[0059]
Further, in this prediction system 100, it is possible to predict in real time how the CSP is moved on both sides of the mounting board and how the reliability changes (for example, whether the direction is good or bad). It is possible to determine a more optimal arrangement position while taking into account. Since the relationship between the floor layout of the CSP and the mounting reliability can be grasped almost in real time, it is possible to avoid an over-specification of the mounting reliability and conversely a situation in which the mounting reliability is insufficient.
[0060]
In this embodiment, the positional relationship between the CSPs (ΔX, ΔY) is input to the response surface to determine the amount of distortion of the solder bump, and the amount of distortion of the solder bump is input to the fatigue life equation to obtain the reliability of the solder bump. The system and method for calculating the characteristic life value have been described. Conversely, the desired reliability life value is input to the above-mentioned fatigue life formula of the Coffin-Manson rule to determine the strain amount of the solder bump. May be input to the response surface to calculate the positional relationship (ΔX, ΔY) between the CSPs that satisfies the desired reliability life value.
[0061]
FIG. 10 is a block diagram illustrating a configuration example of the mounting reliability prediction system 100 ′. 10, components having the same functions as those in FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted. In FIG. 10, the parameter storage unit 11 stores parameters of the shape and material characteristics of the mounting board and the CSP excluding the positional relationship (ΔX, ΔY) between the CSPs. Further, the experiment data storage unit 17 stores the experiment results (strain amount, reliability life value) on the test board. Further, the reliability data storage unit 21 stores a desired reliability life value.
[0062]
Then, the CPU 23 'inputs the experimental data and the desired reliability life value into the fatigue life equation (see equations (2) and (3) in FIG. 5) and calculates the amount of nonlinear distortion amplitude of the solder bump. Further, the CPU 23 'inputs the calculated nonlinear distortion amplitude of the solder bump into the response surface, and outputs the positional relationship (ΔX, ΔY) between the CSPs that satisfies the desired reliability life value.
[0063]
By displaying the positional relationship between the output CSPs by contour lines or the like on the monitor screen 27, it is possible to determine at a glance which direction and how much each CSP mounted on both sides of the mounting board should be moved to improve the reliability. It is easy to grasp and convenient. It is possible to grasp the arrangement area where the connection reliability of the CSP can be maintained up to the XX cycle, and it is possible to easily teach the floor layout designer where the arrangement is possible. For example, the response surface shown in FIG. 3 is displayed on the monitor screen 27.
[0064]
In addition, if the mounting reliability prediction systems 100 and 100 ′ can be accessed from a remote place via a telephone line, a wired cable such as an optical cable, or a wireless communication, the analysis life is not limited and the reliability life can be improved. This is convenient because calculation processing of the optimal positional relationship (ΔX, ΔY) can be executed. By using the systems 100 and 100 'through a web browser, input and results can be displayed on a homepage screen. In addition, software upgrades and bug fixes can be centrally managed.
[0065]
【The invention's effect】
As described above, according to the method for predicting the mounting reliability of an electronic component according to the present invention, the electronic component is mounted on each of both opposing surfaces of the substrate via a solder joint on each of the opposing surfaces of the substrate. When estimating the mounting reliability of the electronic component on the board of the electronic device, the amount of distortion generated in the solder joint is determined in advance in accordance with the positional relationship between the electronic components, and the amount of distortion is determined. A response curve or a response surface to be displayed corresponding to the positional relationship between the electronic components is created.
[0066]
With this configuration, unlike the conventional method, it is not necessary to create an FEM model corresponding to the electronic device for each reliability evaluation in order to obtain the amount of distortion of the solder joint in the electronic device. It is not necessary to re-create the FEM model according to. Therefore, it is possible to easily and quickly predict the reliability life of the solder joint in the electronic device.
[0067]
Further, according to the electronic component mounting reliability prediction system according to the present invention, an electronic device having a wiring substrate and having the electronic component attached to both opposing surfaces of the substrate via solder joints, A system for predicting the mounting reliability of the electronic component with respect to the substrate, a response curve or a response surface that displays the amount of strain generated at the solder joint in accordance with the positional relationship between the electronic components, It is provided with an arithmetic processing unit for inputting arbitrary positional relationship information and calculating the amount of distortion generated in the solder joint.
[0068]
With this configuration, unlike the conventional method, it is not necessary to create an FEM model corresponding to the electronic device for each reliability evaluation in order to obtain the distortion amount of the solder joint in the electronic device. Therefore, it is possible to easily and quickly predict the reliability life of the solder joint in the electronic device.
[0069]
INDUSTRIAL APPLICABILITY The present invention is very suitable when applied to the reliability evaluation of digital cameras, notebook computers, and the like in which a CSP is attached to both sides of a mounting board via solder bumps.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration example of a mounting reliability prediction system 100 according to an embodiment of the present invention.
FIG. 2 is a table showing an example of parameters and their setting ranges.
FIG. 3 is a table showing an example of a response surface.
FIG. 4 is a logarithmic graph showing a fatigue life formula based on the Coffin-Manson rule.
FIG. 5 is an example of an arithmetic expression used for calculating a distortion amount.
FIG. 6 is a table showing an example of a parameter sensitivity analysis result by the Taguchi method.
FIG. 7 is a table showing an example of a response curve.
FIG. 8 is a flowchart showing a method for predicting mounting reliability.
FIG. 9 is a conceptual diagram illustrating an example of an input screen.
FIG. 10 is a block diagram illustrating a configuration example of a mounting reliability prediction system 100 ′.
FIG. 11 is a conceptual diagram illustrating an example of an FEM model 70.
[Explanation of symbols]
11 ... parameter storage unit (position information storage unit), 13 ... model storage unit, 15 ... fatigue life type storage unit, 17 ... experiment data storage unit, 19 ... distortion amount storage unit, 21: reliability data storage unit, 23: CPU (first and second arithmetic processing units), 23 ': CPU (third and fourth arithmetic processing units), 27: monitor , 50A, 50B ... CSP, 60 ... Storage unit body, 100, 100 '... Mounting reliability prediction system

Claims (12)

A method of predicting the mounting reliability of an electronic component on an electronic device having a wiring substrate and having electronic components attached to opposite surfaces of the substrate via solder joints, respectively,
In advance, the amount of distortion generated in the solder joint is determined in accordance with the positional relationship between the electronic components, and a response curve or a response surface displaying the amount of distortion in association with the positional relationship between the electronic components is calculated. The process of creating,
Arbitrarily setting a positional relationship between the electronic components, and applying the set positional relationship to the response curve or response surface to calculate a strain amount generated in the solder joint. Component mounting reliability prediction method. Applying a strain amount of the solder joint obtained from the response curve or the response surface to a fatigue life equation such as the Coffin-Manson rule to calculate a reliability life value of the solder joint. Item 1. A method for predicting mounting reliability of an electronic component according to Item 1. Applying the strain amount of the solder joint determined from the response curve or the response surface to a fatigue life formula such as the Coffin-Manson rule, and calculating the reliability life value of the solder joint,
Prepare a substrate for the experiment, attach each of the electronic components via solder joints to both opposing surfaces of the substrate,
Determine the strain amount and the reliability life value of the solder joint on the experimental substrate to which the electronic component is attached, and then,
The amount of solder joint distortion and reliability life value determined on the experimental substrate, and the amount of solder joint distortion on the wiring substrate calculated from the response curve or response surface, 3. The reliability prediction of electronic component mounting according to claim 2, wherein the reliability life value of the solder joint on the wiring board is calculated by applying a fatigue life formula such as the Coffin-Manson rule. Method. In the step of determining the amount of strain generated in the solder joint in accordance with the positional relationship between the electronic components,
The method according to claim 1, wherein the amount of distortion is calculated using a finite element method. The parameter variable constituting the response curve or the response surface is a positional relationship between the electronic components, a shape and a material property of the electronic component, and a shape and a material property of the solder joint. 2. The method for predicting mounting reliability of an electronic component according to item 1. A system for predicting the mounting reliability of an electronic component with respect to the substrate, wherein the electronic device includes a wiring substrate and has electronic components attached to opposite sides of the substrate via solder joints, respectively,
A model storage unit that stores a response curve or a response surface that displays the amount of strain generated in the solder joint in association with the positional relationship between the electronic components,
A position information storage unit for storing arbitrary positional relationship information between the electronic components,
An arithmetic processing unit that inputs the positional relationship information stored in the position information storage unit to a response curve or a response surface stored in the model storage unit, and calculates an amount of strain generated in the solder joint. Characteristic electronic component mounting reliability prediction system. When the arithmetic processing unit is a first arithmetic processing unit,
A fatigue life equation storage unit for storing a fatigue life equation such as Coffin-Manson rule;
A second calculation for calculating the reliability life value of the solder joint by inputting the strain amount of the solder joint calculated by the first arithmetic processing unit into a fatigue life formula such as the Coffin-Manson rule. 7. The system according to claim 6, further comprising a processing unit. A third arithmetic processing for calculating a distortion amount of a solder joint by inputting a desired reliability life value of the solder joint into a fatigue life equation such as the Coffin-Manson rule stored in the fatigue life equation storage unit The mounting reliability prediction system according to claim 6, further comprising a unit. A fourth step of inputting the distortion amount of the solder joint calculated by the third processing unit to the response curve or the response surface, and outputting positional relationship information between electronic components satisfying the reliability life value; The mounting reliability prediction system according to claim 8, further comprising an arithmetic processing unit. 8. The mounting reliability according to claim 7, further comprising a monitor unit for displaying on a screen positional relationship information between the electronic components and a reliability life value of the solder joint corresponding to the positional relationship information. Forecasting system. 7. The mounting reliability prediction system according to claim 6, further comprising a monitor unit for displaying the response curved surface on a screen with contour lines. The mounting reliability prediction system according to claim 6, wherein the location information storage unit and the arithmetic processing unit are accessible via a telephone line, a cable such as an optical cable, or wireless.

Images