Method and system for detecting aging of diodes in a semiconductor device

Abstract

The application discloses a kind of aging detection method and system of diode in semiconductor device, it is related to diode detection relevant technical field, method includes: obtaining the diode access circuit of target semiconductor device, the state self-checking and determination of each diode are executed by micro-self-checker;If the self-checking result is abnormal, trigger inquiry verification test;Including: generating inquiry instruction and sending to micro-self-checker, the same synchronous operating condition parameter is integrated, and the host device is imported in multithreading parallel;Gradient consistency mutual authentication is executed, and aging detection result is determined, if gradient is inconsistent, it is determined as individual aging failure, if gradient is consistent, it is determined as collective drift failure.Solve the technical problems that cannot be efficiently and accurately distinguished between individual aging failure and collective performance drift in redundant / parallel diode circuits in the prior art, resulting in insufficient reliability and real-time performance of aging detection, achieve the technical effects of precise differentiation of fault modes, improve the reliability and real-time performance of diode aging detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of diode detection, and particularly relates to a diode aging detection method and system in a semiconductor device. BACKGROUND

[0002] As the most basic semiconductor element, the performance and reliability of a diode directly affect the stability and service life of the entire circuit system. However, during long-term operation, due to factors such as electrical stress, thermal stress, and material degradation, the diode is prone to aging phenomena such as an increase in forward voltage drop and an increase in reverse leakage current, which in turn leads to a decrease in device performance and even functional failure. Aging problems are particularly prominent in high-temperature, high-humidity, or high-switching-frequency application environments and can cause system failure. Traditional diode aging detection mainly relies on periodic shutdown detection or external test equipment, such as using a multimeter or an oscilloscope to measure parameters. This not only has low efficiency but also makes it difficult to achieve real-time monitoring and early warning. In addition, in a circuit in which multiple diodes are connected in redundancy or parallel, there is a lack of effective processing of the coordinated aging of multiple diodes in the redundant or parallel structure. Due to the mutual influence between diodes, individual aging can be masked, and thus it is difficult to accurately distinguish between individual aging and collective performance drift, resulting in insufficient reliability, real-time performance, and parallelism of the detection results.

[0003] Therefore, in the related art, there is a technical problem that individual aging failure and collective performance drift cannot be efficiently and accurately distinguished in a redundant / parallel diode circuit, resulting in insufficient reliability and real-time performance of the aging detection. SUMMARY

[0004] The present application provides a diode aging detection method and system in a semiconductor device, which solves the technical problem that individual aging failure and collective performance drift cannot be efficiently and accurately distinguished in a redundant / parallel diode circuit, resulting in insufficient reliability and real-time performance of the aging detection in the prior art, and achieves the technical effects of accurately distinguishing fault modes and improving the reliability and real-time performance of diode aging detection.

[0005] The application provides a method for detecting aging of diodes in a semiconductor device, the method comprising: obtaining diode access circuits of a target semiconductor device, performing state self-checking and determination of each diode by deploying micro self-checkers, and determining self-checking results, wherein the diode access circuits are redundantly accessed or in parallel accessed by a plurality of diodes; if the self-checking result is abnormal, triggering inquiry verification determination, wherein each micro self-checker independently performs self-checking and determination; wherein the inquiry verification determination step comprises: a first micro self-checker generates an inquiry instruction, sends it to micro self-checkers of at least two adjacent diodes, synchronizes time stamping of synchronous operation parameters, and imports an upper device in multiple threads in parallel, wherein the first micro self-checker is a micro self-checker with abnormal self-checking, and corresponds to a first diode; the upper device performs gradient consistency mutual verification under a normalized condition to determine the aging detection result of the first diode, wherein if the gradients are inconsistent, it is determined as individual aging failure, and if the gradients are consistent, it is determined as collective drift failure.

[0006] In a possible implementation, the method for detecting aging of diodes in a semiconductor device further performs the following processing: determining diode specifications and aging trend curves for each diode accessed, wherein the aging trend curve is a standardized curve, and each node represents an operation condition of an aging stage; and constructing micro self-checkers and upper devices according to the diode specifications and the aging trend curves, wherein the micro self-checkers and the upper devices are deployed in an integrated circuit controller.

[0007] In a possible implementation, the method for detecting aging of diodes in a semiconductor device further performs the following processing: determining a plurality of mapping groups by mapping the diode specifications and the aging trend curves, wherein each diode accessed by a circuit corresponds to a mapping group; performing dynamic self-checking training under a time sequence for a first mapping group, determining a first micro self-checker, and taking the first micro self-checker as an agent of a first diode; traversing the plurality of mapping groups until a construction of an Nth micro self-checker is completed, and taking the Nth micro self-checker as an agent of an Nth diode; and embedding and deploying the first micro self-checker to the Nth micro self-checker in an integrated circuit controller.

[0008] In a possible implementation, the method for detecting aging of diodes in a semiconductor device further performs the following processing: juxtaposing aging trend curves of each diode, performing normalization processing, and establishing a time sequence curve mapping as a detection baseline; taking at least two diode operation parameter groups as inputs, taking mutual verification based on the detection baseline as a determination mode, and constructing the upper device, wherein the upper device is deployed in an integrated circuit controller, and the determination result of the micro self-checker is a trigger condition.

[0009] In a possible implementation, the method for detecting aging of a diode in a semiconductor device further performs the following processing: the first diode collects real-time operating parameters during operation of the semiconductor device, and the parameters are imported into the first micro-self-checker; node matching and parameter deviation determination are performed in the first aging trend curve, and if there is a parameter deviation, a deviation gradient vector is determined and an inquiry instruction is generated.

[0010] In a possible implementation, the method for detecting aging of a diode in a semiconductor device further performs the following processing: the diodes adjacent to the first diode are taken as inquiry targets, and the inquiry instruction is sent to the micro-self-checkers of the adjacent diodes, wherein the number of the adjacent diodes being inquired is at least two; the micro-self-checkers corresponding to the adjacent diodes receive the inquiry instruction, the second micro-self-checker sorts a second gradient vector under a synchronous time stamp constraint, and the third micro-self-checker sorts a third gradient vector under the synchronous time stamp constraint, wherein the second micro-self-checker and the third micro-self-checker are intelligent agents corresponding to the adjacent diodes.

[0011] In a possible implementation, the method for detecting aging of a diode in a semiconductor device further performs the following processing: according to the interactive connection of the micro-self-checker to the upper device, multi-thread parallel import of the deviation gradient vector, the second gradient vector, and the third gradient vector is performed, and mutual verification under a normalization condition is performed; according to gradient consistency, aging failure of the first diode is determined.

[0012] In a possible implementation, the method for detecting aging of a diode in a semiconductor device further performs the following processing: if the gradients are inconsistent, and the gradient value of the normalized deviation gradient vector is higher than the normalized second gradient vector and the normalized third gradient vector, and a gradient threshold is satisfied, it is determined that a local individual failure occurs, and the first diode is identified as an aging failure diode; if the gradients are consistent, it is determined that a collective drift failure occurs at a whole-machine level, and the first diode is identified as a normal aging diode.

[0013] In a possible implementation, the method for detecting aging of a diode in a semiconductor device further performs the following processing: through identification, if the first diode is a normal aging diode, no response is made; if the first diode is an aging failure diode, failure alarm information is generated according to a circuit access position and a diode code, and the information is displayed for alarm on a terminal interface.

[0014] The application also provides an aging detection system for a diode in a semiconductor device, comprising: a self-check result determination module, configured to access a diode circuit of a target semiconductor device, perform state self-checking and determination of each diode by deploying micro self-checkers, and determine a self-check result, wherein the diode circuit is redundantly accessed or accessed in parallel by a plurality of diodes; an inquiry verification determination module, configured to trigger inquiry verification determination if the self-check result is abnormal, wherein each micro self-checker performs self-checking and determination independently; wherein the inquiry verification determination step comprises: an inquiry instruction generation submodule, configured to generate an inquiry instruction by a first micro self-checker, send the inquiry instruction to micro self-checkers of at least two adjacent diodes, synchronize synchronous time stamping synchronous operating parameters, and import the synchronous operating parameters into an upper device in multiple threads in parallel, wherein the first micro self-checker is a micro self-checker with abnormal self-checking, and corresponds to a first diode; and an aging detection result determination submodule, configured to perform gradient consistency mutual verification under a normalization condition by the upper device, and determine an aging detection result of the first diode, wherein if the gradients are inconsistent, it is determined that individual aging failure occurs, and if the gradients are consistent, it is determined that collective drift failure occurs.

[0015] The application provides a diode aging detection method and system in a semiconductor device, which access a diode circuit of a target semiconductor device, deploy micro self-checkers to perform state self-checking and determination of each diode, and trigger inquiry verification determination if the self-check result is abnormal. The method comprises: generating an inquiry instruction and sending the inquiry instruction to a micro self-checker, synchronizing synchronous operating parameters, and importing the synchronous operating parameters into an upper device in multiple threads in parallel; performing gradient consistency mutual verification, determining an aging detection result, determining individual aging failure if the gradients are inconsistent, and determining collective drift failure if the gradients are consistent. The technical problems that the prior art cannot efficiently and accurately distinguish individual aging failure from collective performance drift in a redundant / parallel diode circuit, and the aging detection reliability and real-time performance are insufficient are solved, and the technical effects of accurately distinguishing fault modes and improving the reliability and real-time performance of diode aging detection are achieved. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly introduced below. In the present application, a flowchart is used to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, according to the needs, various steps can be processed in reverse order or simultaneously. At the same time, other operations can be added to these processes, or one or more steps of the operations can be removed from these processes.

[0017] Figure 1 A diode aging detection method flowchart in a semiconductor device is provided for the embodiments of the present application.

[0018] Figure 2A structure diagram of a semiconductor device diode aging detection system provided by an embodiment of the present application is shown.

[0019] The reference signs are explained as follows: a self-check result determination module 10, an inquiry verification determination module 20, an inquiry instruction generation sub-module 21, and an aging detection result determination sub-module 22. DETAILED DESCRIPTION

[0020] The above description is only a summary of the technical solutions of the present application. In order to make the technical solutions of the present application more clear, the embodiments of the present application can be implemented according to the content of the description, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application are described as follows.

[0021] In order to make the purposes, technical solutions and advantages of the present application more clear, the present application will be further described in detail below with reference to the drawings, and the described embodiments should not be regarded as limiting the present application. All other embodiments obtained by those skilled in the art without making creative efforts belong to the scope of protection of the present application.

[0022] In the following description, "some embodiments" are related to a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The term "first\second" is only to distinguish similar objects, and does not represent a specific order of the objects. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art in the technical field of the present application. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0023] An aging detection method for diodes in a semiconductor device is provided by an embodiment of the present application, as shown in the method comprising: Figure 1

[0024] In step S100, the diode access circuit of the target semiconductor device is obtained, the state self-check and determination of each diode are performed through the deployment of the micro self-checker, and the self-check result is determined, wherein the diode access circuit is a plurality of diodes redundantly accessed or connected in parallel.

[0025] ​Preferably, the target semiconductor device can be an IGBT power module, a rectifier bridge stack, or a specific integrated circuit chip, etc. The diode access circuit of the target semiconductor device is obtained, i.e. the connection mode of the diodes in the device circuit is obtained, wherein the diode access circuit is a plurality of diode redundant access or parallel access. Specifically, the diode access circuit uses a plurality of diodes, redundant access means that a plurality of diodes assume the same functional role in the circuit to improve system reliability, so that even if one diode fails, other diodes can still maintain the basic function of the circuit. The connection points in the circuit can be the same, or there can be a logical backup relationship. Parallel access means that the anode and cathode of a plurality of diodes are directly connected together, the purpose is to share large current, reduce the on-state loss and thermal stress of a single diode, and improve the overall current handling capacity. Then, a micro-self-checker is assigned to each diode in the circuit, i.e. a micro-detection unit, such as a small embedded program, a logic circuit unit, or a micro-hardware unit integrating a sensor and processing function; then the state self-checking and determination of each diode is performed by the micro-self-checker. Each micro-self-checker independently, continuously or periodically monitors the key operating parameters of the diode corresponding to the micro-self-checker, including forward conduction voltage drop, reverse leakage current, junction temperature, etc. Each micro-self-checker compares the monitored operating parameters with the normal parameter range of the corresponding diode preset in the micro-self-checker, and outputs the binary determination result of normal or abnormal as the self-checking result and outputs it.

[0026] Further, step S100 further comprises step S110 of determining the diode specification and aging trend curve for each diode accessed, wherein the aging trend curve is a standardized curve, and each node represents the operating condition of an aging stage; and step S120 of constructing the micro-self-checker and the upper device according to the diode specification and the aging trend curve, wherein the micro-self-checker and the upper device are deployed in the integrated circuit controller.

[0027] Preferably, the diode specification is determined for each diode accessed, that is, by referring to the data manual of the diode, the specific electrical parameters and physical characteristics of each diode in the circuit are obtained and determined, including but not limited to the rated forward current, reverse breakdown voltage, typical value / maximum value of forward voltage drop, thermal resistance, etc., as the reference value for judging whether the current state of the diode is normal; the diode aging trend curve under the specified standard test conditions of ambient temperature, working current, switching frequency, etc. is obtained, and each node in the aging trend curve usually corresponds to a time point or a cumulative stress level, representing the operating condition of the aging stage, that is, under this time point / stress level, the expected key operating parameter condition of the diode, usually the change amount of forward voltage drop or the growth value of reverse leakage current, for example, a diode aging trend curve may be described as follows: after working for 1000 hours under the condition of 125°C junction temperature and rated current, the forward voltage drop is expected to increase by 5%; after working for 2000 hours, the forward voltage drop is expected to increase by 10%, depicting the expected path of the gradual drift of the performance parameters of the diode over time.

[0028] Preferably, according to the diode specification and the aging trend curve, independent detection logic is created for each diode, specifically, the normal parameter range and the expected aging degree of the diode standard condition are configured, that is, according to the aging trend curve, the allowable drift amount of the operating parameter within the normal aging range under the current operating time or stress is determined, and then a micro self-checker is constructed, which compares the parameters of the bound diode in real time with the dynamic and life expectancy changing judgment threshold to make a preliminary judgment of normality or abnormality; according to the diode specification and the aging trend curve, the central processing logic responsible for collaborative verification is created, specifically, the aging trend curves of all diodes are aligned and normalized to make the parameter changes between different diodes comparable, and the gradient consistency mutual verification rule is configured, that is, how to judge whether the data difference of multiple micro self-checkers is caused by the individual failure of a certain diode or by the environmental stress or overall life loss experienced by all diodes; wherein, the micro self-checker and the upper device are deployed in the integrated circuit control, which means that the judgment logic of the micro self-checker and the verification logic of the upper device are in the form of embedded software or hardware logic circuit, integrated in the integrated circuit control, wherein the integrated circuit control refers to a centralized, integrated circuit form control unit, which may include a control chip integrated inside the semiconductor device itself or a microcontroller, an application specific integrated circuit or a programmable logic device closely connected with the diode in the same package or on the same PCB, thereby realizing on-chip detection without the need for external test equipment, realizing miniaturization, embeddedness, real-time and high reliability of diode aging detection of semiconductor devices.

[0029] Further, step S120 further comprises step S121 of determining a plurality of mapping groups by mapping the diode specifications and the aging trend curves, wherein each diode accessed by the circuit corresponds to a mapping group; step S122 of performing dynamic self-check training under time series for the first mapping group to determine a first micro self-checker, and taking the first micro self-checker as an agent of the first diode; step S123 of traversing the plurality of mapping groups until the construction of an Nth micro self-checker as an agent of an Nth diode is completed; and step S124 of embedding and deploying the first micro self-checker to the integrated circuit central control.

[0030] Preferably, for each diode accessed by the circuit, the diode is bound with a corresponding diode specification and aging trend curve to form a corresponding mapping group, i.e., each diode accessed by the circuit corresponds to a mapping group. Then, for the first mapping group, dynamic self-check training under time series is performed, wherein the first mapping group represents mapping data of any diode, a prediction model is constructed based on a long short-term memory network, the first mapping group is trained by using the prediction model, the state of the diode at different time points in the entire life cycle is simulated, how to judge whether the parameter value conforms to the normal range predicted by the aging curve in the first mapping group according to the real-time collected operating parameters combined with the currently running time or stress level is learned, a judgment rule set is generated as the core logic of the first micro self-checker, the first micro self-checker is determined to be generated after the self-check training is completed, and the first micro self-checker is taken as an agent of the first diode for diode state self-checking and judgment. Then, the plurality of mapping groups are traversed to construct a second micro self-checker, a third micro self-checker, and so on until the construction of an Nth micro self-checker is completed, wherein N is a positive integer representing the total number of diodes accessed by the circuit, and then the agents of N diodes are determined. Finally, the first micro self-checker to the Nth micro self-checker are embedded and deployed in the integrated circuit central control. Specifically, the logic of each micro self-checker can be implemented as a small piece of hardware logic circuit on the chip, and finally works cooperatively in the central control chip to perform diode state self-checking and judgment.

[0031] Further, step S120 further comprises step S125 of juxtaposing the aging trend curves of the diodes, performing normalization processing, and establishing a time series curve mapping as a detection baseline; and step S126 of taking a group of operating condition parameters of any at least two diodes as input, constructing the upper device based on mutual verification of the detection baseline as a detection method, and deploying the upper device in the integrated circuit central control with the detection result of the micro self-checker as a trigger condition.

[0032] Preferably, the aging trend curves of all diodes in the collection circuit are compared and analyzed, and the initial parameters of different diodes may be slightly different due to manufacturing tolerances. Normalization is performed, including converting the parameter vertical axis of each aging trend curve into a relative change rate or percentage change, for example, converting a 0.2V increase in forward voltage drop to a 10% increase in forward voltage drop. Then, all diode aging trend curves are fused to establish a time series curve mapping, that is, the horizontal axis is aligned to generate a standardized reference curve representing the overall expected aging behavior of the batch of diodes, which is used as a detection baseline to create a unified judgment standard for the upper device, for judging whether the aging state of any diode deviates from the normal track of the group. With at least two diode operating parameter groups as input, the operating parameter group refers to the data preprocessed by the micro-self-checker, which may include normalized parameter values, deviation values of the parameter relative to the detection baseline, current running time, etc. Then, based on the detection baseline mutual verification, the specific detection method is as follows: when receiving multiple diode parameter groups, map the data to the detection baseline, check whether the gradient of the data points on the detection baseline is consistent, and at the same time, mutual verification is performed, that is, the operating states of multiple diodes are compared, and the deviation degree of the diodes is analyzed. If the deviation degree of a diode is much larger than that of other diodes, it means that the diode has individual aging failure problem. If multiple input diodes increase synchronously and the amplitude is similar, it means that the collective drift failure is caused by environmental factors or collective life consumption. Then, the upper device is constructed. The upper device and the micro-self-checker are embedded in the integrated circuit of the central control, and coexist in the same chip, ensuring high-speed and reliable internal communication and data exchange. The upper device takes the detection result of the micro-self-checker as the trigger condition. When normally operating, the upper device is in standby or low-power state. When any micro-self-checker locally determines that the state of the diode it is responsible for is abnormal, the micro-self-checker encodes the detection result as a signal and sends it to the upper device, triggering the upper device to activate from standby state and start executing cooperative verification, avoiding unnecessary power consumption and calculation resource consumption caused by continuous operation of the upper device.

[0033] Step S200, if the self-checking result is abnormal, trigger inquiry verification detection, wherein each micro-self-checker independently performs self-checking and judgment.

[0034] Preferably, if the self-checking result is abnormal, indicating that the forward voltage drop of the corresponding diode collected by any micro-self-checker in the detection period exceeds the normal aging range allowed at that time point, the inquiry verification test is triggered, that is, the inquiry instruction is automatically generated to initiate a verification request, so as to avoid unnecessary maintenance actions caused by false positives at a single point. Each micro-self-checker is relatively independent in self-checking and judgment. Specifically, each diode only collects the bound diode parameters and judges only according to the collected diode parameters. Each micro-self-checker has an independent logic unit or program thread to process its task, which is not affected by others, thereby achieving accurate judgment of faults and improving the reliability of self-checking and judgment.

[0035] The inquiry verification test step includes:

[0036] In step S210, the first micro-self-checker generates an inquiry instruction and sends it to the micro-self-checkers of at least two adjacent diodes, synchronizes the synchronous operating parameters with the synchronous time stamp, and imports the upper device in multiple threads in parallel. The first micro-self-checker is the micro-self-checker with abnormal self-checking, corresponding to the first diode.

[0037] Step S210 further includes step S211 of collecting real-time operating parameters of the first diode in real time as the semiconductor device operates and importing the first micro-self-checker; and step S212 of performing node matching and parameter deviation judgment in the first aging trend curve. If there is a parameter deviation, a deviation gradient vector is determined and an inquiry instruction is generated.

[0038] Preferably, the first micro-self-checker generates an inquiry instruction. The first micro-self-checker is the micro-self-checker with abnormal self-checking, corresponding to the first diode. Specifically, all micro-self-checkers collect real-time operating parameters of the corresponding diodes of the semiconductor device at the same time, with the same sampling rate and filtering algorithm, to avoid misjudgment caused by asynchronous collection. The real-time operating parameters refer to physical quantity data generated by the diode during operation, which can reflect the health status of the diode, mainly including forward conduction voltage drop, reverse leakage current, junction temperature or shell temperature measured by a thermistor, and possibly including power module parameter changes caused by solder layer fatigue and binding wire shedding. Then, node matching is performed in the first aging trend curve, that is, the state of the diode at the current time is located to the corresponding position on the pre-stored aging trend curve, and parameter deviation judgment is performed, including comparing and calculating the difference between the real-time parameter value just collected and the corresponding theoretical reference value on the curve. If the difference does not satisfy the pre-set allowable deviation threshold, it is determined that there is a parameter deviation, and the deviation value and the change rate of the deviation are calculated. The deviation value, the change rate of the deviation, the time stamp and the corresponding diode ID are constructed into a deviation gradient vector. Finally, the deviation gradient vector is packaged into a standard format data packet as the core content of the inquiry instruction to request the neighbor nodes to perform collaborative verification.

[0039] Further, step S210 further comprises step S213, taking the adjacent diodes of the first diode as the inquiry target, sending the inquiry instruction to the micro-self-checkers of the adjacent diodes, wherein the number of the adjacent diodes being inquired is at least two; step S214, the micro-self-checkers corresponding to the adjacent diodes receive the inquiry instruction, the second micro-self-checker arranges the second gradient vector under the constraint of the synchronous time stamp, and the third micro-self-checker arranges the third gradient vector under the constraint of the synchronous time stamp, wherein the second micro-self-checker and the third micro-self-checker are intelligent agents corresponding to the adjacent diodes.

[0040] Preferably, the adjacent diodes of the first diode are taken as the inquiry target, wherein the adjacent diodes refer to the diodes closest to the first diode in physical layout or circuit connection, for example, the diodes in parallel in the same power module or the diodes next to each other on the PCB board, which may share similar working environments such as temperature and current stress, and the first micro-self-checker sends the inquiry instruction to the adjacent diodes through the data communication link inside the integrated circuit, wherein the number of the adjacent diodes being inquired is at least two, so as to facilitate consistency verification; the adjacent diodes are in the most similar working environment with the first diode, if their states are normal, the abnormality of the first diode is most likely to be individual aging failure; if they also show similar abnormality, it is possible to be collective drift failure.

[0041] Preferably, the micro-self-checkers corresponding to the adjacent diodes receive the inquiry instruction, the second micro-self-checker and the third micro-self-checker continuously monitor the communication link, the second micro-self-checker parses the key information from the received inquiry instruction, especially the synchronous time stamp, and then retrieves the operating parameters of the second diode collected at the time stamp from its local cache or register, and uses the retrieved real-time parameters to execute the same calculation process as the first micro-self-checker, including node matching, parameter deviation judgment, calculation of deviation value and gradient, and generates the second gradient vector representing the health status of the second diode at the time; similarly, the third micro-self-checker generates the third gradient vector representing the state of the third diode, wherein the synchronous time stamp constraint ensures that the first, second and third gradient vectors describe the states of all diodes at the same time; the second micro-self-checker and the third micro-self-checker are intelligent agents corresponding to the adjacent diodes.

[0042] Further, step S210 further comprises step S215, according to the interactive connection of the micro-self-checker to the upper device, performing multi-thread parallel import of the deviation gradient vector, the second gradient vector and the third gradient vector, and performing mutual verification under the normalization condition; step S216, according to the gradient consistency, performing aging failure judgment on the first diode.

[0043] Preferably, the interactive connection of the micro-self-checker to the host device refers to the physical communication path or other high-speed data link between the micro-self-checker and the host device, which ensures that data can be transmitted quickly. The specifications of the accessed diodes are different, and the corresponding aging determination is different. For example, the corresponding aging standards under the same gradient may be different, and then the multi-thread parallel import of the deviated gradient vector, the second gradient vector, and the third gradient vector is performed. That is, the deviated gradient vector, the second gradient vector, and the third gradient vector are imported in multi-thread parallel mode through the shared communication link. The host device can simultaneously receive and process data from multiple micro-self-checkers, improving processing efficiency and response speed. The mutual verification under the normalization condition is performed, including dividing the absolute deviation value in each gradient vector by the initial rated value of the corresponding diode, converting it into a relative change rate, and then performing mutual verification, that is, the host device performs comparison to obtain the gradient consistency verification result. According to the gradient consistency, the aging failure of the first diode is determined. If the gradients are inconsistent, that is, the normalized gradient value of the first diode is greater than the normalized gradient values of the second and third diodes, it indicates that the behavior of the first diode deviates from the group trend, and it is determined to be an individual aging failure. If the gradients are consistent, the normalized gradient values of the first, second, and third diodes are very close, indicating that the behaviors of all diodes are synchronized, and it is determined that it is not a single diode problem, but a collective drift failure caused by a common cause.

[0044] In step S220, the host device performs gradient consistency mutual verification under the normalization condition to determine the aging detection result of the first diode. If the gradients are inconsistent, it is determined to be an individual aging failure, and if the gradients are consistent, it is determined to be a collective drift failure.

[0045] Step S220 further includes step S221, if the gradients are inconsistent and the gradient value of the normalized deviated gradient vector is higher than the normalized second gradient vector and the third gradient vector, and satisfies the gradient threshold, it is judged to be a local individual failure, and the first diode is identified as an aging failure diode. Step S222, if the gradients are consistent, it is judged to be a collective drift failure at the whole machine level, and the first diode is identified as a normal aging diode.

[0046] Preferably, if the gradients are inconsistent and the normalized gradient value of the first diode is significantly greater than the normalized gradient values of the second and third diodes, and simultaneously satisfies the gradient threshold, it indicates that the first diode is a local individual aging failure, wherein the gradient threshold is a preset critical value for preventing false judgment caused by noise interference and ensuring the reliability of the judgment. Then the first diode is identified as an aging failure diode. If the gradients are consistent, that is, the normalized gradient values of the first, second, and third diodes are very close, indicating that the behaviors of all diodes are synchronized, and it is determined to be a collective drift failure, for example, the overall life consumption caused by the general increase of environmental temperature and the increase of system load. Then the first diode is identified as a normal aging diode. Finally, the aging detection result of the first diode is determined.

[0047] Further, step S220 further includes step S223, if it is a normal aging diode, no response is made by performing identification recognition; step S224, if it is an aging failure diode, failure alarm information is generated by using the circuit access position and the diode code, and the terminal interface is displayed for alarm.

[0048] Preferably, after the host completes the gradient consistency mutual authentication, it is concluded that the diode set drifts, the state identification of the first diode is updated from self-test exception to normal aging diode, and identification recognition is performed on it. When the identification of the first diode is identified as a normal aging diode, the internal logic executes the instruction of no response, preventing unnecessary fault alarms due to normal and synchronous performance degradation. After the host identifies the identification of the first diode as an aging failure diode, failure alarm information is generated by using the circuit access position and the diode code, and at least includes the failure diode, the corresponding diode code, the specific position in the circuit, and the fault time, the failure parameter, the severity level, etc. Finally, it is sent to the user interactive terminal device through the communication interface, and after receiving and analyzing the failure alarm information, it is displayed and alarmed in the terminal interface in the form of pop-up window, sound prompt, and red mark in the list, thereby realizing accurate differentiation of fault modes, improving the reliability and real-time performance of diode aging detection, and realizing predictive and accurate maintenance of semiconductor devices diodes.

[0049] In the foregoing, with reference to Figure 1 A diode aging detection method in a semiconductor device according to an embodiment of the application is described in detail. Next, with reference to Figure 2 A diode aging detection system in a semiconductor device according to an embodiment of the application is described.

[0050] The diode aging detection system in a semiconductor device according to the embodiment of the application is used to solve the technical problem that in the prior art, individual aging failure and collective performance drift cannot be efficiently and accurately distinguished in a redundant / parallel diode circuit, resulting in insufficient reliability and real-time performance of aging detection. The technical effect of realizing accurate differentiation of fault modes and improving the reliability and real-time performance of diode aging detection is achieved. As shown in Figure 2 The diode aging detection system in a semiconductor device includes a self-test result determination module 10, an inquiry verification determination module 20, an inquiry instruction generation sub-module 21, and an aging detection result determination sub-module 22.

[0051] The self-checking result determination module 10 is configured to obtain a diode access circuit of a target semiconductor device, perform state self-checking and determination of each diode by deploying a micro self-checker, and determine a self-checking result, wherein the diode access circuit is a plurality of diodes redundantly accessed or connected in parallel; the query verification and determination module 20 is configured to trigger query verification and determination if the self-checking result is abnormal, wherein each micro self-checker independently performs self-checking and determination; wherein the query verification and determination step includes: the query instruction generation submodule 21 is configured to generate a query instruction by a first micro self-checker, send the query instruction to micro self-checkers of at least two adjacent diodes, synchronize time stamping of synchronous operation parameters, and import the synchronous operation parameters into an upper device in a multi-thread parallel manner, wherein the first micro self-checker is a micro self-checker with abnormal self-checking result, and corresponds to a first diode; the aging detection result determination submodule 22 is configured to perform gradient consistency mutual verification under a normalization condition by the upper device, and determine an aging detection result of the first diode, wherein if the gradients are inconsistent, it is determined that the individual is aging failure, and if the gradients are consistent, it is determined that the collective is drifting failure.

[0052] Next, the specific configuration of the self-checking result determination module 10 will be described in detail. The self-checking result determination module 10 further includes: determining diode specifications and aging trend curves for each diode accessed, wherein the aging trend curve is a standardized curve, and each node represents an operation condition of an aging stage; constructing the micro self-checker and the upper device according to the diode specifications and the aging trend curves, wherein the micro self-checker and the upper device are deployed in an integrated circuit controller.

[0053] Next, the specific configuration of the self-checking result determination module 10 will be described in detail. The self-checking result determination module 10 further includes: determining a plurality of mapping groups by mapping the diode specifications and the aging trend curves, wherein each diode accessed by a circuit corresponds to a mapping group; performing dynamic self-checking training under a time sequence for a first mapping group, determining a first micro self-checker, and taking the first micro self-checker as an agent of a first diode; traversing the plurality of mapping groups until the construction of an Nth micro self-checker is completed, and taking the Nth micro self-checker as an agent of an Nth diode; and embedding and deploying the first micro self-checker to the Nth micro self-checker in the integrated circuit controller.

[0054] Next, the specific configuration of the self-checking result determination module 10 will be described in detail. The self-checking result determination module 10 further includes: parallelizing the aging trend curves of each diode, performing normalization processing, and establishing a time sequence curve mapping as a detection baseline; taking at least two diode operation parameter groups as input, constructing the upper device based on mutual verification based on the detection baseline as a determination method, wherein the upper device is deployed in an integrated circuit controller, and the determination result of the micro self-checker is taken as a trigger condition.

[0055] Below, the specific configuration of the query instruction generation submodule 21 will be described in detail. The query instruction generation submodule 21 further comprises: collecting real-time operating parameters by the first diode during the operation of the semiconductor device, and importing the first micro-self-checker; performing node matching and parameter deviation determination in the first aging trend curve, determining the deviation gradient vector and generating the query instruction if there is a parameter deviation.

[0056] Below, the specific configuration of the query instruction generation submodule 21 will be described in detail. The query instruction generation submodule 21 further comprises: taking the adjacent diode of the first diode as the query target, and sending the query instruction to the micro-self-checker of the adjacent diode, wherein the number of the adjacent diodes being queried is at least two; the micro-self-checker corresponding to the adjacent diode receives the query instruction, the second micro-self-checker arranges the second gradient vector under the constraint of the synchronous time stamp, and the third micro-self-checker arranges the third gradient vector under the constraint of the synchronous time stamp, wherein the second micro-self-checker and the third micro-self-checker are intelligent agents corresponding to the adjacent diode.

[0057] Below, the specific configuration of the aging detection result determination submodule 22 will be described in detail. The aging detection result determination submodule 22 further comprises: performing multi-thread parallel import of the deviation gradient vector, the second gradient vector and the third gradient vector according to the interactive connection of the micro-self-checker to the upper device, and performing mutual verification under the normalization condition; performing aging failure determination on the first diode according to the gradient consistency.

[0058] Below, the specific configuration of the aging detection result determination submodule 22 will be described in detail. The aging detection result determination submodule 22 further comprises: if the gradients are inconsistent, and the gradient value of the normalized deviation gradient vector is higher than the normalized second gradient vector and the third gradient vector, and satisfies the gradient threshold, judging as local individual failure, and identifying the first diode as an aging failure diode; if the gradients are consistent, judging as a whole drift failure at the machine level, and identifying the first diode as a normal aging diode.

[0059] Below, the specific configuration of the aging detection result determination submodule 22 will be described in detail. The aging detection result determination submodule 22 further comprises: through the identification, if it is a normal aging diode, no response is made; if it is an aging failure diode, failure warning information is generated in the circuit access position and the diode code, and the display warning is performed on the terminal interface.

[0060] The aging detection system of the diode in the semiconductor device provided by the embodiment of the present application can execute the aging detection method of the diode in the semiconductor device provided by any embodiment of the present application, has the function modules and beneficial effects corresponding to the execution method.

[0061] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or the server, the various units and modules included are only divided according to the functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific name of each functional unit is only for the convenience of mutual differentiation, and is not used to limit the protection scope of the present application.

[0062] The above detailed description does not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method of detecting aging of a diode in a semiconductor device, characterized by, The method comprises: Obtaining a diode access circuit of a target semiconductor device, performing state self-checking and determination of each diode by deploying a micro self-checker, and determining a self-checking result, wherein the diode access circuit is a plurality of diodes redundantly accessed or connected in parallel; If the self-checking result is abnormal, triggering an inquiry verification determination, wherein each micro self-checker independently performs self-checking and determination; The inquiry verification determination step comprises: A first micro self-checker generates an inquiry instruction, sends it to the micro self-checkers of at least two adjacent diodes, synchronizes the synchronous operation parameters of the time stamp, and imports the upper device in multiple threads in parallel, wherein the first micro self-checker is the micro self-checker of the first diode which is self-checked abnormally; The upper device performs gradient consistency mutual verification under the normalization condition to determine the aging detection result of the first diode, wherein if the gradient is inconsistent, it is determined as individual aging failure, and if the gradient is consistent, it is determined as collective drift failure.

2. A method of detecting aging of a diode in a semiconductor device according to claim 1, wherein Deploying a micro self-checker comprises: For each accessed diode, determining a diode specification and an aging trend curve, wherein the aging trend curve is a standardized curve, and each node represents the operation condition of an aging stage; According to the diode specification and the aging trend curve, the micro self-checker and the upper device are constructed, wherein the micro self-checker and the upper device are deployed in an integrated circuit controller.

3. A method of detecting aging of a diode in a semiconductor device according to claim 2, wherein According to the diode specification and the aging trend curve, the micro self-checker is constructed, comprising: By mapping the diode specification and the aging trend curve, a plurality of mapping groups are determined, wherein each diode accessed by the circuit corresponds to a mapping group; For the first mapping group, dynamic self-checking training under time series is performed to determine a first micro self-checker, and the first micro self-checker is taken as an agent of the first diode; The plurality of mapping groups are traversed until the construction of an Nth micro self-checker is completed, which is taken as an agent of an Nth diode; The first micro self-checker to the Nth micro self-checker are embedded and deployed in the integrated circuit controller.

4. A method of detecting aging of a diode in a semiconductor device according to claim 2, wherein According to the diode specification and the aging trend curve, the upper device is constructed, comprising: Parallel the aging trend curves of each diode, perform normalization processing and establish time series curve mapping as a detection baseline; Taking at least two diode operation parameter groups as input, constructing the upper device based on mutual verification of the detection baseline as a determination method, wherein the upper device is deployed in an integrated circuit controller, and the determination result of the micro self-checker is taken as a trigger condition.

5. The method of claim 1, wherein the semiconductor device is a diode. The first micro self-checker generates an inquiry instruction, comprising: ​ With the operation of the semiconductor device, the first diode performs real-time operation parameter collection and imports the first micro self-checker; In the first aging trend curve, node matching and parameter deviation determination are performed, and if there is a parameter deviation, a deviation gradient vector is determined and an inquiry instruction is generated.

6. A method of detecting the aging of a diode in a semiconductor device according to claim 5, wherein Synchronizing the synchronous operation parameters of the time stamp, comprising: Taking the adjacent diodes of the first diode as inquiry targets, the inquiry instruction is sent to the micro self-checkers of the adjacent diodes, wherein the number of the adjacent diodes being inquired is at least two. The micro-self-checker corresponding to the adjacent diode pair receives the query instruction, the second micro-self-checker sorts the second gradient vector under the synchronous timestamp constraint, and the third micro-self-checker sorts the third gradient vector under the synchronous timestamp constraint, wherein the second micro-self-checker and the third micro-self-checker are intelligent agents corresponding to the adjacent diode pair.

7. A method of detecting degradation of a diode in a semiconductor device according to claim 6, wherein According to the interactive connection of the micro-self-checker to the upper device, the multi-thread parallel import of the deviation gradient vector, the second gradient vector and the third gradient vector is performed, and the mutual verification under the normalization condition is performed. According to the gradient consistency, the aging failure of the first diode is determined.

8. A method of detecting degradation of a diode in a semiconductor device according to claim 7, wherein If the gradients are inconsistent, and the gradient value of the normalized deviation gradient vector is higher than the normalized second gradient vector and the third gradient vector, and the gradient threshold is satisfied, it is judged that the local individual fails, and the first diode is identified as an aging failure diode. If the gradients are consistent, it is judged that the whole machine level collective drift failure occurs, and the first diode is identified as a normal aging diode.

9. The method of claim 1, wherein the diode is in a semiconductor device. Through the identification, if it is a normal aging diode, no response is made. If it is an aging failure diode, failure alarm information is generated at the circuit access position and the diode code, and the terminal interface is displayed for alarm.

10. An aging detection system for diodes in semiconductor devices, characterized in that, The system is used to implement the aging detection method of the diode in the semiconductor device according to any one of claims 1 to 9, and the system comprises: A self-checking result determination module is configured to obtain a diode access circuit of a target semiconductor device, deploy a micro-self-checker, perform state self-checking and determination of each diode, and determine a self-checking result, wherein the diode access circuit is a plurality of diodes redundantly accessed or connected in parallel. An inquiry verification determination module is configured to trigger inquiry verification determination if the self-checking result is abnormal, wherein each micro-self-checker independently performs self-checking and determination. The inquiry verification determination step comprises: An inquiry instruction generation submodule is configured to generate an inquiry instruction by a first micro-self-checker, send the inquiry instruction to micro-self-checkers of at least two adjacent diodes, synchronize operation parameters of the same synchronous timestamp, and import the upper device in multi-thread parallel mode, wherein the first micro-self-checker is a micro-self-checker corresponding to the first diode. An aging detection result determination submodule is configured to perform gradient consistency mutual verification under the normalization condition by the upper device, and determine the aging detection result of the first diode, wherein if the gradients are inconsistent, it is judged that the individual is aging failure, and if the gradients are consistent, it is judged that the collective drift failure occurs.

Images