JP2013015542A - Method for predicting lifetime of solder joint part, device for predicting lifetime of solder joint part, and electronic apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and device for predicting the lifetime of a solder joint part, which accurately evaluates the lifetime of a joint part in consideration of the reduction of rigidity due to advance of damage in individual solder joint parts, and to provide an electronic apparatus.SOLUTION: The method for predicting the lifetime of a solder joint part includes the steps of: referring to history information of a temperature of a measurement target including a solder joint part; examining at least one of physical quantities of an amplitude of temperature variation, the number of cycles, an average temperature, and a period by cycle counting in accordance with the history information of the temperature; calculating a distortion range from at least one of the physical quantities examined by cycle counting by using a preliminarily generated response curved surface; and calculating a distortion range increase rate by referring to a preliminarily obtained damage value and a distortion variation history of the distortion range in accordance with the distortion range.

Description

  Embodiments described herein relate generally to a solder joint life prediction method, a solder joint life prediction apparatus, and an electronic apparatus.

  There are various types of defects under use of electronic devices. Among them, in particular, defects in joints such as solder joints occur frequently and are one of the troublesome troubles in that once they occur, the operation is seriously affected. The solder joint is distorted by various loads such as a thermal load caused by power ON / OFF and an external mechanical load. Even if the strain is such that it does not break at a time, strain amplitude may accumulate and metal fatigue may occur due to repeated loading. As a technique for estimating the life until breakage due to such a fatigue phenomenon, a structural health monitoring technique for electronic devices is known.

  In order to evaluate the fatigue life of metals including solder joints with high accuracy, it is important to correctly estimate the value of strain generated at the evaluation point. However, in the life prediction of BGA solder joints, fluctuations in strain caused by the progress of damage to the solder joints, including cracks, are often not considered. In other words, the same strain prediction method as before the progress of damage is often used even after damage due to cracks progresses in the solder joint and the amount of strain fluctuation increases due to a decrease in rigidity. Considering the time and effort required to create a damage model (model for predicting damage values of solder joints from temperature fluctuations) and simplification of the algorithm at the time of mounting, the accuracy of life prediction itself can be achieved even in a method that does not consider damage. Inferior but practical predictions are possible. However, when strain amplitude accumulates in the solder joints, cracks propagate in the solder and the rigidity of the solder joints decreases, so the stress that each solder joint takes on as the temperature fluctuates. The ratio also changes. Therefore, when predicting the lifetime of the signal bumps further inside than the dummy bumps existing outside, an error occurs between the actual lifetime and the progress of damage is not considered. The main cause of the life error here is that a difference occurs between the strain value predicted for each bump and the actual strain due to the change in rigidity.

JP 2010-73795 A

  The first embodiment provides a method for predicting the life of a solder joint that evaluates the life of the joint with higher accuracy in consideration of the decrease in rigidity caused by the progress of damage to individual solder joints.

  The second embodiment is intended to provide a solder joint life prediction apparatus that realizes the first embodiment.

  An object of the third embodiment is to provide an electronic apparatus having a configuration capable of efficiently detecting breakage.

  According to the first embodiment, the step of referring to the temperature history information of the measurement target having a solder joint, and the temperature fluctuation amplitude, the number of cycles, the average temperature, and the cycle by the cycle count from the temperature history information. A step of examining at least one physical quantity; a step of calculating a strain range from at least one of the physical quantities examined by the cycle count using a response surface prepared in advance; and And a step of calculating a strain range increase rate with reference to a damage value obtained in advance and a strain fluctuation history of the strain range.

  According to the second embodiment, the first storage unit that stores the history information of the temperature of the measurement target having a solder joint, and the physical quantity of at least one of the amplitude of the temperature fluctuation, the number of cycles, the average temperature, and the period. A second storage unit storing a response surface for determining a strain range from the third storage unit storing a damage value history information, and a strain fluctuation history for determining an increase rate of the strain range from the damage value Is stored, a first control unit that obtains temperature history information with reference to the first storage unit, and a temperature fluctuation amplitude and cycle by cycle count from the temperature history information. A second control unit that examines at least one physical quantity of number, average temperature, and period; and at least one of physical quantities and physical quantities examined by the cycle count with reference to the second storage unit A strain range increase rate is calculated from the strain range calculated by the third control unit with reference to the third control unit that calculates the strain range from the third control unit, the third storage unit, and the fourth storage unit. And a fourth control unit. A solder joint life prediction device is provided.

  According to the third embodiment, an electronic component, a mounting substrate, the electronic device and the mounting substrate are mechanically joined, and an electric signal is exchanged between the electronic device and the mounting substrate. A first joint, a second joint that mechanically joins the electronic device and the mounting substrate and does not mediate an electrical signal exchange between the electronic device and the mounting substrate; 3rd joining formed between 1 joining part and the said 2nd, mechanically joining the said electronic device and the said mounting substrate, and monitoring the joining state of the said mounting substrate and the said electronic component And an electronic device characterized by having a unit.

The process flow figure concerning a 1st embodiment. The block diagram which concerns on 2nd Embodiment. The conceptual diagram which shows the structure of 2nd Embodiment. FIG. 4 is an enlarged view of a solder joint according to a range of A1 in FIG. 3. The enlarged view of the solder joint which concerns on the range of A2 of FIG. The conceptual diagram which shows the relationship between the position of a solder joint part, a temperature cycle, and a strain range. The conceptual diagram which shows the relationship between the number of temperature cycles and a strain range. Results of strain history fluctuation history obtained by FEM numerical analysis. The conceptual diagram which shows the load state of a theoretical fixed temperature range and a fixed time. The conceptual diagram which shows an actual temperature fluctuation history. The conceptual diagram which shows the power law until solder reaches a fracture | rupture. The conceptual diagram which shows the relationship between a temperature amplitude and its frequency. The conceptual diagram which shows the relationship between a strain range and its frequency. The figure which shows the relationship between a damage value and the increase rate of a strain range. Flow chart of creation process of strain fluctuation history database. The block diagram of the apparatus which produces a distortion fluctuation history database. The process flow figure showing the modification of a 1st embodiment. The block diagram of the apparatus which performs an error correction routine. The example of the lifetime prediction correction which concerns on the modification of 1st Embodiment. The conceptual diagram which shows the structure of 3rd Embodiment. The conceptual diagram which shows the modification of 3rd Embodiment.

  Hereinafter, embodiments will be described with reference to the drawings.

(First embodiment)
A solder life prediction method will be described as a first embodiment with reference to FIGS.

(Step 001-002)
When a life update event is detected, the first control unit (update event detection unit) 11 refers to the temperature information in the second storage unit (temperature history database) 22 stored up to that point.

  The life update event is stored in the firmware stored in the first storage unit (update event storage unit) 21 by an instruction program for triggering the event at regular intervals, and the update event detection unit 11 stores the instruction program. It can be realized by executing.

  The temperature history database 22 stores past temperature history. The temperature of the substrate 1 is measured by the detection unit 7 and stored in the temperature history database 22.

(Step 003)
The second control unit (cycle count investigation unit) 12 refers to the temperature history database 22 and examines information such as temperature amplitude, number of cycles, average temperature, and cycle by cycle count.

  The temperature history database 22 stores the temperature measured by the detection unit 7 and the time when the temperature was measured as time series data.

(Step 004)
The third control unit (temperature amplitude-strain range conversion unit) 13 is a third storage unit (response surface database) based on information such as temperature amplitude, number of cycles, average temperature, and period obtained by the cycle count investigation unit 12. 23, the temperature amplitude ΔT is converted into a strain range Δε.

  The response surface database 23 stores response surfaces that are functions of temperature fluctuation amplitude, number of cycles, average temperature, period and strain range. By using this response curved surface, the strain range can be obtained from the amplitude of the temperature fluctuation, the number of cycles, the average temperature, and the period. This response surface is stored in advance in the third storage unit (response surface database) 23.

(Step 005)
The fourth control unit (strain range increase rate calculation unit) 14 refers to the fourth storage unit (damage value database) 24 and the fifth storage unit (strain fluctuation history database) 25, and converts the converted strain range Δε. Is used to calculate the rate of increase of the strain range caused by the progress of the damage accumulated in the solder joint 71 when the current event is updated.

  The damage value database 24 stores the damage value and the time when the damage value was obtained as time series data. The strain fluctuation history database 25 stores a function of a damage value and a strain range increase rate.

  For example, when the damage value D1 at a certain time t1 is obtained from the damage value database 24, the strain range increase rate α is obtained from the damage value D1. For example, a new strain range Δε2 = α · Δε1 is obtained from the strain range Δε1 using the strain range increase rate α.

(Step 006)
The fifth control unit (strain range recounting unit) 15 counts the damage value again in consideration of the increase rate of the strain range calculated in step 005. Then, the newly obtained damage value is stored in the fourth storage unit (damage value database) 24 together with the time when the current event is updated.

  Taking the above example, a new damage value D2 can be obtained from the newly obtained strain range Δε2 using equations 1 and 2 described later. The newly obtained damage value D2 is stored in the damage value database together with the time t2.

(Step 007)
The sixth control unit (damage value determination unit) 16 refers to the seventh storage unit (threshold value database) 27 and determines whether or not the new damage value is greater than or equal to a specific threshold value. When the new damage value is equal to or greater than a specific threshold value, the damage value determination unit 26 is allowed to transmit a signal instructing to take a predetermined action.

  Possible actions include automatically backing up data assuming that a failure is near, telling the user about the possibility of failure, and telling the server that manages the device that the life is near. It is done.

  The specific threshold value is appropriately set according to the type of product or the use of the product. For example, when high reliability such as a power plant or medical equipment is required, the threshold value can be set strictly. On the other hand, when it is applied to a server of a network system in which a mechanism for compensating for a failure is prepared, the threshold can be set loosely.

(Action, function, technical significance)
Here, assuming a mounting board having a solder joint as a product to be measured, and an example of a method for predicting fatigue life of a BGA for a temperature fluctuation load in a use environment after this product is shipped The technical significance of the first embodiment will be further described.

  The first embodiment is characterized in that the life prediction of the solder joint portion is performed in consideration of the effect that the rigidity is lowered due to the damage of each solder joint portion.

  A schematic diagram of the progress of damage at the solder joint is shown in FIG. This figure shows that when the BGA is subjected to a temperature cycle load for the dummy bump 1 (A-1), the dummy bump 2 (B-2), and the signal bump 3 (C-3) among the solder joints shown in FIG. FIG. 4 schematically shows the variation of the strain range due to the progress of damage. Note that the dummy bump 1, the dummy bump 2, and the signal bump 3 indicate a region A2 surrounded by a dashed line in the upper right corner of the component 4-7 shown in FIG. And area | region A2 is detail drawing of area | region A1 enclosed with the dashed-dotted line of the components 4-7 of the electronic device shown in FIG. In this specification, in FIG. 5, the solder joints are denoted as A, B, C,... From right to left, 1, 2, 3,. May be indicated.

  As the temperature cycle ΔT progresses, the strain range of the outermost dummy bump 1 first increases. Next, the strain range increases in the order of the dummy bump 2 and finally the signal bump 3. This phenomenon occurs because cracks in the solder joint progress due to accumulation of strain, and the rigidity of the solder joint decreases. Therefore, by calculating the strain fluctuation history of the dummy bumps in advance by numerical analysis when calculating the strain from temperature fluctuation information, it is possible to calculate the strain taking into account the decrease in rigidity, and expect high life accuracy. .

  FIG. 7 shows an example of strain fluctuation history of dummy bumps (A-2 to E-2) and signal bumps shown in FIG. 4 from numerical analysis by FEM. In this case, it is observed that the strain variation increases in order from the outer dummy bump, and it is understood that the result shows the same tendency as in FIG.

  On the other hand, FIG. 8 shows a case where the strain amplitude is evaluated by a conventional evaluation method. In this case, the strain amplitude is determined regardless of the progress or fracture of damage such as a crack in the solder joint.

(Preparing the database)
Based on these considerations, the first embodiment can be implemented as follows.

  First, a database representing a strain change history is prepared in advance for each solder joint for which the life needs to be predicted in order to take into account the decrease in rigidity that occurs as the temperature cycle progresses.

  Here, in a numerical analysis such as a finite element method assuming a state in which a temperature cycle as shown in FIG. 9 acts, by applying a method that considers a decrease in rigidity due to damage of the solder joint, FIG. Such a relationship between the number of temperature cycles and the strain range Δε can be obtained. However, the relationship between the number of temperature cycles and the strain range Δε in FIG. 6 is the value of temperature width (ΔT in FIG. 9), holding time (Δt2, Δt4 in FIG. 9), rise time, and fall time (Δt1, Δt3 in FIG. 9). Fluctuates depending on. Further, the temperature history experienced by the product in the market is not a cycle of a constant temperature width and a constant time as shown in FIG. 9, but is subjected to more complicated temperature fluctuations as shown in FIG. For such complicated temperature fluctuations, a database for estimating a correct strain range in consideration of damage is required.

Therefore, as information stored in the database, the relationship between the damage value of each solder joint that requires life prediction and the increase in strain range is used. The damage value D is an index indicating the degree of fatigue when a load having a different amplitude is applied. When D = 1 is reached, the service life is reached and breakage occurs. FIG. 11 shows a life prediction method when the amplitude of the constant strain range Δε ₀ is repeated. When the amplitude of the constant strain range Δε ₀ is repeated, the number of repetitions N ₀ until the fracture is derived according to an exponential rule (in particular, in the case of solder, an exponential rule called Coffin-Manson rule or Basquin rule). here,
α, β: constants determined by the material Nf: crack initiation cycle N ₀ is the number of crack initiation cycles assuming that the strain of Δε ₀ is repeated N: the number of cycles in which Δε ₀ is actually loaded,
Δε: Strain range D is a damage value under N-cycle load.

  On the other hand, FIGS. 1 and 12 to 14 show a life prediction method when a plurality of amplitudes in different strain ranges (temperature ranges) are applied.

  The temperature history data (FIG. 10) is converted into temperature amplitude data by a method called cycle count (step 0003 in FIG. 1). In addition, the number of cycles and the period of the temperature history are simultaneously investigated. An example of temperature amplitude data is shown in FIG. See, for example, ASTM E 1049-85, “Standard Practices for cycle counting in fatigue analysis”, ASTM Standards, Vol. 03.01 (Reapproved 1997), Philadelphia, 1999.

Further, the temperature amplitude data is converted into a strain range generated in the solder joint by a response curved surface prepared in advance (step 0004 in FIG. 1). An example converted to the strain range is shown in FIG. Here, the response curved surface is an equation that is created by numerical analysis assuming cases of various temperature ranges in advance and can predict the strain generated in the solder joint with high accuracy. At the time of creating this response surface, it is assumed that each solder joint has a constant rigidity regardless of the progress cycle in order to avoid complication, and the decrease in rigidity due to the progress of damage is not considered. Allow that. The life when amplitudes of different strain ranges Δε _k act can be obtained according to a linear cumulative damage law (also referred to as Miner law), and the sum D of reciprocals of the fracture life N _k derived by each strain range can be obtained (formulae) 1, Formula 2). When the damage value D reaches 1, the solder joint portion breaks. By defining D as a damage value, it is possible to represent the degree of progress of damage in the solder joint until the fatigue life is reached.

An example of a relationship diagram in which the horizontal axis represents the damage value D and the vertical axis represents the increase rate of the strain range Δε for the solder joints (A-1, A-2, B-1 in FIG. 5). As shown in FIG. Here, the increase rate of the strain range Δε is a damage value in which the strain range generated in the solder joint when a certain temperature amplitude is applied in the initial state before damage progresses is defined as the initial strain range Δε _0. The strain range generated by applying the same temperature amplitude when D is D is a value indicating the rate of increase with respect to the strain range Δε _{0 in} the initial state, and is obtained by numerical analysis such as FEM in consideration of the progress of damage. be able to. For the specific method, refer to, for example, the Japan Society of Mechanical Engineers Proceedings (A), Volume 73, No. 736 (2007-12) "Damage path simulation of solder bump joint"

  This relationship is stored in advance in the strain fluctuation history database 25 by numerical analysis or the like and is compared with the damage value D obtained from the temperature history under the use environment of the product, thereby knowing the increasing rate of the strain range Δε due to the progress of damage. be able to. 15 to 16 show a procedure for creating a strain fluctuation history database and a block diagram of an apparatus for creating a strain fluctuation history database.

(Procedure for creating strain fluctuation history database)
(Steps 011 to 012)
The eleventh control unit (cycle determination unit) 31 determines a temperature range, a temperature holding time, a temperature rise time, a temperature fall time, and a temperature repetition cycle number.

(Step 013)
The twelfth control unit (numerical analysis unit) 32 performs numerical analysis in consideration of the progress of damage. Specifically, the temperature conditions specified by the FEM are set, and when the strain range generated in the solder reaches the fatigue life number specified by the Coffin-Manson rule or the Basquin rule, Simulate a crack by removing or drastically reducing stiffness and consider the progression of damage. For details, see the Japan Society of Mechanical Engineers Proceedings (A), Volume 73, Issue 736 (2007-12).

(Step 014)
A thirteenth control unit (strain variation history calculation unit) 33 calculates a strain variation history for each temperature cycle. The strain fluctuation history calculation unit 33 stores the obtained strain fluctuation history for each temperature cycle in the fifth storage unit (strain fluctuation history information database) 25.

(Step 015)
The fourteenth control unit (damage value calculation unit) 34 refers to the eleventh storage unit (solder fatigue database) 26 and calculates (estimates) the damage value D for each temperature cycle. Further, the damage value calculation unit 34 stores the increase rate of the strain range Δε in each temperature cycle and the calculated damage value D corresponding thereto in the fifth storage unit (strain variation history information database) 25.

  The 015th step can be performed at an arbitrary opportunity such as real time or a predetermined timing. As the determined timing, for example, periodically or when the power is turned on and / or off.

(effect)
Here, as a first advantage of using the damage value D to know the increase amount of the strain range Δε, only the damage value is considered without considering the applied temperature range, holding time, rise time, and fall time. It is mentioned that the increase rate of the strain range can be known by referring to. That is, the present invention can be easily applied not only to a simple temperature fluctuation such as a constant temperature amplitude but also to a complicated temperature fluctuation in an actual use environment. The second advantage is that the relationship between the damage value D and the increasing rate of the strain range Δε is robust, and the relationship is unlikely to change greatly depending on the applied temperature range and holding time.

  Because of these advantages, the relationship between the damage value and strain range due to typical temperature history is examined in advance by numerical analysis, and the rate of increase in strain range for various temperature histories can be known easily and with relatively high accuracy. Is possible.

  In order to estimate the strain range increase rate with higher accuracy, numerical analysis using the temperature range, holding time, ascending and descending time as parameters is held in the database, and multiple relationships are retained depending on the added temperature range. It is also possible to adopt a method of applying various relational expressions.

  In the first embodiment, a method is used in which the temperature variation amplitude, the number of cycles, the average temperature, and the period are checked from the temperature history information by the cycle count, and the strain range is calculated using a response surface prepared in advance. Although shown, it is not limited to this. Depending on the usage environment of the object to be measured, estimation may be possible without using all physical quantities of temperature amplitude, number of cycles, and period. Specifically, there are cases where estimation is possible only by the amplitude and the number of cycles.

  Therefore, it is possible to check the physical quantity of at least one of the amplitude of the temperature fluctuation, the number of cycles, the average temperature, and the period by the cycle count from the temperature history information, and a response surface corresponding to the investigated physical quantity is prepared. If the strain range can be calculated from at least one of the physical quantities examined by the cycle count, the life prediction of the solder joint may be possible.

(Modification of the first embodiment)
A modification of the first embodiment will be described. In this modification and the first embodiment, the procedure for predicting the life is the same, but when a break of the dummy bump is detected, an error between the expected life of the dummy bump and the actual life is calculated. The difference is that a routine for correcting the error is incorporated. FIG. 17 shows a life prediction method incorporating an error correction routine. Moreover, FIG. 18 shows the structure added to 1st Embodiment among the structures required in order to perform this lifetime prediction method.

(Steps 021 to 022)
A dummy bump break detection event is provided before the life prediction method (step S025) of the first embodiment.

  The twenty-first control unit (break event detecting unit) 51 determines whether or not the dummy bump is broken (step 022). When it is determined that the dummy bump is not broken, the first embodiment (steps 001 to 009) is executed (FIG. 1).

(Step 023)
If it is determined that the dummy bump is broken, the twenty-second control unit (damage value correcting unit) 52 refers to the damage value database 24 and reads a damage value obtained in advance. And
The damage value correcting unit 52 compares a damage value given in advance with a newly obtained damage value. Further, when a difference is found in the damage value in the comparison result, the damage value correcting unit 52 corrects the life prediction value of the solder joint portion that is not broken (step 023).

  A method for correcting the predicted life value will be described with reference to an example shown in FIG. The damage value D should be broken when it reaches 1, but it is assumed that the breakage of the solder joint is detected when D = 0.75 as shown in FIG. In this case, it is determined that a damage value corresponding to ΔD = 0.25 (= 1−0.75) is added for some reason that is not predicted, and the damage value of the remaining unjoined solder joint is D = Divide by 0.75. That is, 0.533 obtained by dividing the damage value 0.4 by 0.75 is set as a new damage value.

  Further, the damage value correcting unit 52 stores the calculated corrected damage value 0.533 in the twenty-first storage unit (dummy bump situation database) 61. Solder joint breakage information and damage values stored in the 21st storage unit (dummy bump status database) 61 can be hosted using information transmission means such as SMART (Self-Monitoring, Analysis and Reporting Technology) as necessary. Provided on the side. Thereafter, the first embodiment (steps 001 to 009) is executed (FIG. 1).

(Second Embodiment)
A second embodiment will be described with reference to FIGS. The second embodiment is an embodiment of a solder joint life prediction apparatus.

  FIG. 3 is a schematic diagram showing an example of the second embodiment in which the semiconductor memory 4-1 to 4-8, the capacitor 6, the measurement unit 7, and the control unit 10 are mounted on the mounting substrate 1. The semiconductor memories 4-1 to 4-8 are connected to the mounting substrate via joints such as solder. The detector 7 measures the electrical characteristics of each junction of the semiconductor memories 4-1 to 4-8, and observes the connection state. The detection unit 7 measures the temperature of the mounting substrate 1. A part of the semiconductor memory can be used as the storage unit 20.

  FIG. 2 is a block diagram according to the second embodiment. 2nd Embodiment has the detection part 7, the control part 10, and the memory | storage part 20 as a component. These are connected to each other via a signal line L that mediates the exchange of information and signals. The signal line can be wired, wireless, or a mixture of these.

  Corresponding to the first embodiment, the control unit 10 includes an update event detection unit 11, a cycle count investigation unit 12, a temperature-strain range conversion unit 13, a strain range increase rate calculation unit 14, a strain range recount unit 15, a damage A value determination unit 16 is included. The storage unit 20 includes an update event storage unit 21, a temperature history database 22, a response surface database 23, a damage value database 24, a strain variation history database 25, a solder fatigue database 26, and a threshold database 27.

  Corresponding to the modification of the first embodiment, the control unit 10 may further include a fracture event detection unit 31 and a damage value correction unit 32, and the storage unit 20 may further include a dummy bump situation database 61.

  As a specific aspect, for example, a CPU can be used for the control unit 10. Further, a semiconductor memory can be used for the storage unit 20. The storage unit is not limited to this, and any storage medium that can store information and programs can be used, and an LSI such as a NAND semiconductor memory, HDD, or ROM can be used.

  The detection unit 7 (break event detection unit 31) can include, for example, a circuit that measures the resistance value of the solder joint 71, and / or a circuit that measures impedance. The detection part 7 can detect a fracture | rupture by measuring the electrical property of a junction part. For example, when the joint portion breaks, the electric resistance suddenly increases, so that the breakage can be detected. Further, in order to measure the temperature, the detection unit 7 can be provided with a thermocouple. When the circuit for measuring the resistance value of the solder joint, the circuit for measuring the impedance is an analog circuit, and the output of the thermocouple is an analog signal, the detection unit 7 may include an A / D converter. By converting analog signals into digital signals, the control unit 10 and the storage unit 20 can easily handle these signals.

  Since the operation, operation, and effect of the second embodiment have been described in the first embodiment and its modifications, the description thereof will be omitted.

(Third embodiment)
A third embodiment will be described. The third embodiment is devised in the way of forming a signal line for detecting breakage. This will be described with reference to FIG.

  In FIG. 20, a plurality of solder joints 71 are arranged in the rectangular semiconductor memory 4. The silicon chip 73 included in the semiconductor memory 4 is slightly smaller than the outermost periphery of the semiconductor memory 4 and is formed inside the semiconductor memory 4.

  The solder joint portion 71 has a dummy bump region (second joint portion, B-1), a break detection region (third joint portion, B-2), a signal line region (first portion) in the direction of the rectangular line target axis. , B-3), a fracture detection region (third joint, B-4), and a dummy bump region (second joint, B-5) are formed in this order. That is, the break detection area is interposed between the signal line area and the dummy bump area.

  The dummy bump areas (B-1, B-5) are areas surrounded by a two-dot broken line in FIG. The dummy bumps in this region mechanically join the semiconductor memory 4 and the mounting substrate 1 which are electronic devices, but do not mediate exchange of electrical signals between the semiconductor memory 4 and the mounting substrate 1.

  The signal line region (B-3) is a region surrounded by a one-dot broken line in FIG. The solder joints in this region not only mechanically join the semiconductor memory 4 and the mounting substrate 1 which are electronic devices, but also mediate exchange of electrical signals between the semiconductor memory 4 and the mounting substrate 1.

  The fracture detection areas (B-2, B-4) are areas surrounded by a broken line in FIG. The solder joint in this region mechanically joins the semiconductor memory 4 that is an electronic device and the mounting substrate 1, but does not mediate the exchange of electrical signals between the semiconductor memory 4 and the mounting substrate 1. However, the signal line 72 for detecting breakage is formed using a part of the solder joint portion existing in this region. The signal line 72 is connected to the detection unit 7 and can monitor the bonding state between the semiconductor memory 4 and the mounting substrate 1. In the example shown in FIG. 20, when two adjacent solder joints in the outermost detection area of the rupture detection region are set as one set, four sets facing each other are connected in series with one electric circuit. Depending on the situation, it is possible to have only two places or only one place in an even number relationship. At this time, if it is selected so as to include a region where the temperature change is most severe among the four corners, it becomes possible to detect the breakage of the solder joint portion at an early stage.

  As described above, the advantage of providing a break detection area between the dummy bump area and the signal line area is that the break detection area is relatively less affected by the variation than the dummy bump area, and the life can be predicted with high accuracy. The point becomes. Since the dummy bump area exists around the periphery of the package, it breaks earlier than the break detection area and the signal area. Physically, this phenomenon is subject to a greater mechanical load on the solder joints on the outside of the package, and the solder joints on the inner side of the package average the remaining loads other than those on the outer solder joints. Is equivalent to At this time, the strain range generated in each solder joint on the inside of the package is averaged compared to the outside (each solder joint on the inner side bears an equivalent load on average). Variation is reduced. This variation tendency is experimentally verified by a load test such as a temperature cycle test.

  The advantage that the variation in the breakage detection area is small also works as an advantage when the damage value of the signal bump is corrected in step 023 in the first embodiment. In other words, since the variation in the solder joints for detecting breakage is small, the variation in the damage value after correction is reduced in other solder joints for which the damage value is corrected by detecting the breakage. Prediction becomes possible. If priority is given to early detection, it is desirable to detect fracture in the dummy bump area outside the package, but from the viewpoint of life prediction accuracy, it is possible to predict with higher accuracy by detecting fracture at the solder joint inside the dummy bump area. become. By appropriately selecting a solder joint for detecting breakage, it is possible to balance early detection with high-precision life prediction.

(Modification of the third embodiment)
As a modification of the third embodiment, the life evaluation can be performed properly by combining the second embodiment with the first embodiment by a chain for detecting breakage. This is shown in FIG.

  Each semiconductor memory (4-1, 4-2,..., 4-N) in FIG. 21 is connected to the detection unit 7 by one stroke writing through a break detection signal line 72. Thereby, the solder joint part of the some semiconductor memory 4 can be monitored in the detection part 7 of one place.

  According to at least one embodiment described above, it is possible to create a test schedule that enables efficient test item implementation in consideration of the load that the diagnosis itself places on the device.

  These embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention and are also included in the invention described in the claims and the equivalents thereof.

1 ... Mounting board (circuit board)
2 ... Connector 3 ... Boss holes 4, 4-1 to 4-8 ... Measurement target (semiconductor memory)
6: Capacitor 7: Detection unit (break detection unit, temperature measurement unit)
10: Control unit 11: First control unit (update event detection unit)
12 ... 2nd control part (cycle count investigation part)
13 ... 3rd control part (temperature-strain range conversion part)
14 ... 4th control part (strain range increase rate calculation part)
15 ... 5th control part (strain range recount part)
16 ... 6th control part (damage value judgment part)
31 ... 11th control part (cycle determination part)
32 ... 12th control part (numerical analysis part)
33 ... 13th control part (strain fluctuation history calculation part)
34 ... 14th control part (damage value calculation part)
51 ... 21st control part (break event detection part)
52 ... 22nd control part (damage value correction part)
20, 20-1, 20-2 ... storage unit 21 ... first storage unit (update event storage unit)
22 ... 2nd memory | storage part (temperature history database)
23 ... 3rd memory | storage part (response curved surface database)
24 ... 4th memory | storage part (damage value database)
25 ... 5th memory | storage part (strain fluctuation history database)
26 ・ ・ ・ Sixth storage section (solder fatigue database)
27: Seventh storage unit (threshold value database)
61 ... 21st memory | storage part (dummy bump situation database)
71 ・ ・ ・ Solder joint (solder bump)
72 ... Break detection signal line 73 ... Si chip 81 ... Host equipment B-1, B-5 ... Dummy bump area B-2, B-4 ... Break detection area B-3 ..Signal line area

Claims (4)

Electronic components,
A mounting board;
A first joint that mechanically joins the electronic device and the mounting substrate, and mediates exchange of electrical signals between the electronic device and the mounting substrate;
A second joint that mechanically joins the electronic device and the mounting substrate and does not mediate the exchange of electrical signals between the electronic device and the mounting substrate;
A third part formed between the first joint and the second, mechanically joins the electronic device and the mounting board, and monitors a joining state between the mounting board and the electronic component. A joint of
An electronic device comprising:   2. The third joint portion according to claim 1, wherein two adjacent solder joint portions are set as one set, and at least one of four pairs facing each other is connected in series. Electronics.   The plurality of electronic components are mounted on the mounting board, the electronic component has a plurality of the third joint portions, and the third joint portions are all connected in series. The electronic device according to claim 1 or 2. A first storage unit that stores history information of the temperature of the measurement target having a solder joint;
A second storage unit storing a response curved surface for obtaining a strain range from at least one physical quantity of temperature fluctuation amplitude, number of cycles, average temperature, and period;
A third storage unit storing history information of damage values;
A fourth storage unit storing a strain fluctuation history for obtaining an increase rate of the strain range from the damage value;
A first control unit that refers to the first storage unit and acquires temperature history information;
A second control unit that examines at least one physical quantity of the amplitude of the temperature fluctuation, the number of cycles, the average temperature, and the period by a cycle count from the history information of the temperature;
A third control unit that refers to the second storage unit and calculates a strain range from at least one of the physical quantities examined by the cycle count;
A fourth control unit that refers to the third storage unit and the fourth storage unit and calculates a strain range increase rate from the strain range calculated by the third control unit;
The electronic device according to any one of claims 1 to 3, further comprising a life prediction device for a solder joint including:

Images