CN117766014A - method for testing irradiation detection memory chip - Google Patents
method for testing irradiation detection memory chip Download PDFInfo
- Publication number
- CN117766014A CN117766014A CN202410190601.7A CN202410190601A CN117766014A CN 117766014 A CN117766014 A CN 117766014A CN 202410190601 A CN202410190601 A CN 202410190601A CN 117766014 A CN117766014 A CN 117766014A
- Authority
- CN
- China
- Prior art keywords
- parameter
- value
- data
- tested
- sensitivity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000012360 testing method Methods 0.000 title claims abstract description 77
- 238000000034 method Methods 0.000 title claims abstract description 64
- 238000001514 detection method Methods 0.000 title claims description 40
- 230000008569 process Effects 0.000 claims abstract description 42
- 238000012545 processing Methods 0.000 claims abstract description 34
- 230000007306 turnover Effects 0.000 claims abstract description 6
- 238000010998 test method Methods 0.000 claims abstract description 4
- 239000011159 matrix material Substances 0.000 claims description 48
- 230000005855 radiation Effects 0.000 claims description 34
- 238000013528 artificial neural network Methods 0.000 claims description 15
- 230000002159 abnormal effect Effects 0.000 claims description 8
- 238000004140 cleaning Methods 0.000 claims description 6
- 239000003086 colorant Substances 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 description 16
- 238000013507 mapping Methods 0.000 description 16
- 238000012544 monitoring process Methods 0.000 description 12
- 230000035945 sensitivity Effects 0.000 description 12
- 238000013461 design Methods 0.000 description 9
- 230000005540 biological transmission Effects 0.000 description 8
- 238000003860 storage Methods 0.000 description 8
- 230000007334 memory performance Effects 0.000 description 7
- 238000007405 data analysis Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000010219 correlation analysis Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 230000007423 decrease Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000011156 evaluation Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 238000007637 random forest analysis Methods 0.000 description 3
- 239000007787 solid Substances 0.000 description 3
- 238000012706 support-vector machine Methods 0.000 description 3
- 230000003471 anti-radiation Effects 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013480 data collection Methods 0.000 description 2
- 238000007418 data mining Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 238000005065 mining Methods 0.000 description 2
- 238000003909 pattern recognition Methods 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 230000005355 Hall effect Effects 0.000 description 1
- 238000012300 Sequence Analysis Methods 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000000540 analysis of variance Methods 0.000 description 1
- 238000012098 association analyses Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- JJWKPURADFRFRB-UHFFFAOYSA-N carbonyl sulfide Chemical compound O=C=S JJWKPURADFRFRB-UHFFFAOYSA-N 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 230000001010 compromised effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000013506 data mapping Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000006993 memory improvement Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 230000002085 persistent effect Effects 0.000 description 1
- 238000000513 principal component analysis Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000000191 radiation effect Effects 0.000 description 1
- 239000000941 radioactive substance Substances 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000010206 sensitivity analysis Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
- 238000002759 z-score normalization Methods 0.000 description 1
Abstract
The application provides a test method for detecting a memory chip by irradiation, which comprises the steps of arranging a plurality of high-sensitivity sensors on a chip to be tested to obtain parameter quantity and parameter variation quantity at the same time in the test process, wherein the parameter quantity comprises a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value, converting a collected analog signal into a digital signal, and transmitting the digital signal to an external processing system; the external processing system writes in input data to a plurality of identical chips to be detected, acquires read output data of the chips to be detected, records the parameter quantity and the parameter variation quantity at the moment when the data inversion occurs in the chips to be detected, acquires the data inversion with the occurrence frequency higher than a preset frequency value, and corresponds to the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the moment, and is used as a judging standard for judging whether the data inversion occurs in the chips to be detected; when the same parameter quantity and parameter variation quantity appear in other identical chips to be tested, the chips to be tested are marked as data turnover.
Description
Technical Field
the application relates to the field of memory chip detection, in particular to a testing method for detecting a memory chip by irradiation.
Background
With the rapid development of aerospace industry, the memory chip is widely applied to the fields of irradiation environments such as satellites, aerospace and the like. Therefore, the evaluation of the irradiation resistance of the memory chip and the evaluation of the reliability and stability under irradiation are particularly critical. The conventional irradiation resistance detection of the memory chip refers to a process of testing and evaluating the stability and reliability of the memory chip in a radiation environment. The radiation may come from cosmic rays, radioactive substances, or other sources of radiation. In a radiation environment, the memory chip may be subject to radiation-induced bit flipping, failure or corruption, resulting in data loss or memory failure.
the existing anti-radiation memory chip detection technology mainly analyzes and compares the irradiated test data through writing and reading operations of the memory chip, evaluates the anti-radiation capability of the memory chip, and determines the influence degree of radiation on the memory performance.
the method for judging the performance of the memory needs to transmit a large amount of data back in the upper computer, and analysis is complex, so a new mechanism is needed to realize the irradiation-resistant detection scheme of the memory chip.
Disclosure of Invention
The embodiment of the application aims to provide a test method for detecting a memory chip by irradiation, which is used for realizing the technical effect of rapidly judging the performance of the memory chip.
The embodiment of the application provides a testing method for radiation detection of a memory chip, which comprises the following steps: step one, arranging a plurality of high-sensitivity sensors on a chip to be tested to obtain parameter quantity and parameter variation quantity of the high-sensitivity sensors at the same time in the test process, wherein the parameter quantity comprises a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value, and converting analog signals collected by the plurality of high-sensitivity sensors into digital signals and transmitting the digital signals to an external processing system; writing input data into a plurality of identical chips to be tested by the external processing system, acquiring read output data of the chips to be tested, and recording parameter amounts and parameter variation amounts of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value at the moment when the data inversion occurs in the chips to be tested is identified by the external processing system, acquiring data inversion with the occurrence frequency higher than a preset frequency value, and correspondingly acquiring the parameter amounts and parameter variation amounts of the high-sensitivity sensors at the moment, and taking the parameter amounts and the parameter variation amounts as judgment standards for judging whether the data inversion occurs in the chips to be tested; and thirdly, marking the chip to be tested as data turnover when other identical chips to be tested have identical changes of current value, voltage value, resistance value, capacitance value, inductance value, temperature value and magnetic field intensity value.
In the implementation process, a plurality of high-sensitivity sensors are arranged on a chip to be tested, so that parameter quantity and parameter variation quantity of the high-sensitivity sensors at the same time in the test process are obtained in real time, wherein the parameter quantity comprises a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value, and analog signals collected by the plurality of high-sensitivity sensors are converted into digital signals and are sent to an external processing system. The external processing system writes in input data to the chip to be tested, acquires read output data of the chip to be tested, records parameter amounts and parameter variation amounts of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value at the moment when the external processing system recognizes that data inversion occurs in the chip to be tested, acquires data inversion with occurrence frequency higher than a preset frequency value, and establishes a database as a template corresponding to the parameter amounts and parameter variation amounts of the high-sensitivity sensor at the moment, and serves as a judging standard for judging whether the data inversion occurs in the chip to be tested. The frequency value of data inversion is preset, and the data inversion with more occurrence frequency can be obtained. When other chips to be tested have the same current value, voltage value, resistance value, capacitance value, inductance value, temperature value and magnetic field intensity value, the chip to be tested is marked as data turnover. The parameter quantity and the parameter variation quantity when the data of a plurality of identical chips to be tested are overturned are used as the standard, the judging standard of the performance of the chips to be tested is converted into the judgment of the parameter quantity and the parameter variation quantity, the complex operation flow of an upper computer is avoided, and the testing speed of the performance of the chips to be tested is accelerated.
in one possible implementation, a multidimensional matrix mapping is established between the parameter quantity collected by the high-sensitivity sensor and the performance of the chip to be tested, and when the same parameter variation quantity occurs in the multidimensional matrix as that when data inversion occurs, corresponding fault points are recorded.
In the implementation process, the parameter collected by the high-sensitivity sensor and the performance of the chip to be tested are subjected to multi-dimensional matrix mapping, for example: [ target chip, target unit, irradiation test condition, detection time, current value, voltage value, resistance value, capacitance value, inductance value, temperature value, magnetic field intensity value, data inversion ], [ target chip, target unit, irradiation test condition, detection time, current value, voltage value, resistance value, capacitance value, inductance value, temperature value, magnetic field intensity value, data damage ], target chip, namely chip number being detected, target unit, namely chip inside unit number being detected, irradiation test condition, namely current irradiation environment record, mainly refers to intensity and accumulated irradiation duration. The detection time is the accurate time for recording the current detection behavior, and the voltage value is the quantized voltage value collected by the high-sensitivity sensor and the variation with the last detection time. The current value is the quantized current value measured by the high sensitivity sensor and the variation from the last detection time. The resistance value is the quantized resistance value measured by the high sensitivity sensor and the variation from the last detection time. The capacitance value is the quantized capacitance value measured by the high-sensitivity sensor and the variation from the last detection time. The inductance value is the quantized inductance value measured by the high-sensitivity sensor and the variation from the last detection time. The temperature value is the quantized temperature value measured by the high-sensitivity sensor and the variation from the last detection time. The magnetic field intensity value is the quantized magnetic field intensity value measured by the high-sensitivity sensor and the variation from the last detection time. Data flip is typically accompanied by data corruption, with a data flip state prediction default of zero, i.e., no flip, which predicts the probability of a target cell of a target chip being flipped or corrupted under current irradiation conditions based on the aforementioned statistics and various analysis methods. When the matrix is used for data recording, the data flip or damage state prediction is changed into data flip or damage state recording, which records the situation that a target unit of a target chip is flipped or damaged under the current irradiation condition. The overall trend of the data in the multidimensional matrix is observed, such as a pattern in which the performance increases or decreases with an increase in a certain parameter. Fault point positioning: matrix mapping can help to locate failure points quickly, which may be due to test errors or real failures.
In one possible implementation manner, fault modes corresponding to different types of fault points are preset, and values corresponding to the cells in the multidimensional matrix are preset for color coding so as to distinguish fault degrees corresponding to the values of the cells in the multidimensional matrix with different colors.
in the implementation process, the fault modes corresponding to different types of fault points are preset, for example, the fault points may form a line or a region in the multidimensional matrix. The preset color codes correspond to the values of the cells in the multidimensional matrix, for example, the high fault degree is set to be red, and the low fault degree is set to be blue, so that different performance levels or fault degrees can be distinguished conveniently.
In one possible implementation manner, the step two further includes: and (3) cleaning data, and eliminating parameter variation values and noise which are irrelevant and abnormal in the parameter quantity collected by the high-sensitivity sensor.
in the implementation process, the second step further includes: and (3) cleaning data, removing parameter variation values and noise which are irrelevant and abnormal in the parameter quantity collected by the high-sensitivity sensor, further improving the accuracy of collecting the parameter quantity, and ensuring the quality of the collected parameter quantity.
In one possible implementation, the method further includes: and (3) signal processing, namely collecting time series data of the parameter by the high-sensitivity sensor, analyzing the time series data, and establishing a Fourier transform curve image by the frequency characteristic and the time dependence of the parameter so as to predict whether data inversion occurs at a certain moment.
In the implementation process, the method further comprises signal processing, time series data analysis of the parameter is collected for the high-sensitivity sensor, a Fourier transform curve image is built according to the frequency characteristic and the time dependence of the parameter, and the frequency characteristic and the time dependence of parameters (such as a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value, a magnetic field intensity value and the like) are analyzed to predict whether data inversion occurs at a certain moment.
in one possible implementation manner, the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the same time when the data overturn occurs are taken as input data, fault modes corresponding to different types of fault points are taken as reference data, a trained artificial neural network model is obtained, the parameter quantity and the parameter variation quantity collected by the high-sensitivity sensor are input into the artificial neural network model, and the fault modes to be generated are estimated to be taken as output data.
In the implementation process, the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the same time when the data overturn occurs are taken as input data, fault modes corresponding to different types of fault points are taken as reference data, a trained artificial neural network model is obtained, the parameter quantity and the parameter variation quantity collected by the high-sensitivity sensor are input into the artificial neural network model, and the fault modes to be generated are estimated to be taken as output data. By estimating the failure mode to be generated, the failure mode to be generated can be known in advance, and the solution is prepared in advance. The algorithm can construct an artificial neural network model for predicting the performance of the memory chip under a specific radiation environment (such as a support vector machine, a random forest and a neural network).
In one possible implementation, based on the parameter quantity and parameter variation quantity of the high-sensitivity sensor at the same time in the test process, one of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value with the largest variation degree when data inversion occurs is obtained, and irradiation conditions are adjusted to reduce the variation degree of the corresponding parameter quantity.
in the implementation process, based on the parameter quantity and parameter variation quantity of the high-sensitivity sensor at the same time in the test process, one of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value with the largest variation degree is obtained when data overturn occurs, and irradiation conditions are adjusted to reduce the variation degree of the corresponding parameter quantity. By analyzing the sensitivity of the parameters, it is determined which physical parameters have the greatest effect on the performance of the memory chip, and then the test conditions or designs can be adjusted to improve the radiation resistance of the memory chip.
In one possible implementation manner, a threshold interval of the parameter number is preset based on the parameter number and the parameter variation quantity of the high-sensitivity sensor at the same time in the test process, and when the parameter quantity collected by the high-sensitivity sensor exceeds the threshold interval, the external processing system sends out an early warning.
In the implementation process, a threshold interval of the parameter quantity is preset based on the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the same time in the test process, and when the parameter quantity collected by the high-sensitivity sensor exceeds the threshold interval, the external processing system sends out early warning. Through setting up threshold value interval to and the mode of early warning, when parameter variation is too big, represent that the data upset degree that takes place is big, and the fault degree is higher, and external processing system sends the early warning, so that the operating personnel in time looks over.
In one possible implementation, before starting the test, the high-sensitivity sensor is calibrated based on the parameter quantity and the parameter variation quantity at the same time in the test process, so as to improve the accuracy of collecting the parameter quantity by the high-sensitivity sensor.
In the implementation process, before the test is started, the high-sensitivity sensor is calibrated based on the parameter quantity and the parameter variation quantity at the same time in the test process, so that the accuracy of collecting the parameter quantity by the high-sensitivity sensor is improved. In addition, these high sensitivity sensors also need to be maintained and calibrated regularly to ensure their accuracy and reliability throughout the test cycle. By careful design and implementation of high sensitivity sensor deployment, high quality and reliable data can be ensured during radiation testing.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments of the present application will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present application and should not be considered as limiting the scope, and other related drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
fig. 1 is a flowchart of a testing method for radiation detection of a memory chip according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings in the embodiments of the present application.
The embodiment of the application provides a testing method of an irradiation detection memory chip, please refer to fig. 1, which comprises the following steps: step one, arranging a plurality of high-sensitivity sensors on a chip to be tested to obtain parameter quantity and parameter variation quantity of the high-sensitivity sensors at the same time in the test process, wherein the parameter quantity comprises a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value, and converting analog signals collected by the plurality of high-sensitivity sensors into digital signals and transmitting the digital signals to an external processing system; writing input data into a plurality of identical chips to be tested by an external processing system, acquiring read output data of the chips to be tested, recording parameter amounts and parameter variation amounts of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value at the moment when the data inversion occurs in the chips to be tested, acquiring data inversion with the occurrence frequency higher than a preset frequency value, and correspondingly acquiring the parameter amounts and parameter variation amounts of a high-sensitivity sensor at the moment, and taking the parameter amounts and the parameter variation amounts as judging standards of whether the data inversion occurs in the chips to be tested; and thirdly, marking the chip to be tested as data turnover when other identical chips to be tested have identical changes of current value, voltage value, resistance value, capacitance value, inductance value, temperature value and magnetic field intensity value.
In the implementation process, a plurality of high-sensitivity sensors are arranged on a chip to be tested, so that parameter quantity and parameter variation quantity of the high-sensitivity sensors at the same time in the test process are obtained in real time, wherein the parameter quantity comprises a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value, and analog signals collected by the plurality of high-sensitivity sensors are converted into digital signals and are sent to an external processing system. The external processing system writes in input data to the chip to be tested, acquires read output data of the chip to be tested, records parameter amounts and parameter variation amounts of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value at the moment when the external processing system recognizes that data inversion occurs in the chip to be tested, acquires data inversion with occurrence frequency higher than a preset frequency value, and corresponds to the parameter amounts and parameter variation amounts of the high-sensitivity sensor at the moment, and is used as a judging standard for judging whether the data inversion occurs in the chip to be tested. The frequency value of data inversion is preset, and the data inversion with more occurrence frequency can be obtained. When other chips to be tested have the same current value, voltage value, resistance value, capacitance value, inductance value, temperature value and magnetic field intensity value, the chip to be tested is marked as data turnover. The parameter quantity and the parameter variation quantity when the data of a plurality of identical chips to be tested are overturned are used as the standard, the judging standard of the performance of the chips to be tested is converted into the judgment of the parameter quantity and the parameter variation quantity, the complex operation flow of an upper computer is avoided, and the testing speed of the performance of the chips to be tested is accelerated.
In one possible implementation, the parameter number collected by the high-sensitivity sensor and the performance of the chip to be tested are mapped into a multidimensional matrix, and when the same parameter variation occurs in the multidimensional matrix as when the data is overturned, the corresponding fault point is recorded.
In the implementation process, the parameter collected by the high-sensitivity sensor and the performance of the chip to be tested are subjected to multi-dimensional matrix mapping, for example: [ target chip, target unit, irradiation test condition, detection time, current value, voltage value, resistance value, capacitance value, inductance value, temperature value, magnetic field intensity value, data inversion ], [ target chip, target unit, irradiation test condition, detection time, current value, voltage value, resistance value, capacitance value, inductance value, temperature value, magnetic field intensity value, data damage ], target chip, namely chip number being detected, target unit, namely chip inside unit number being detected, irradiation test condition, namely current irradiation environment record, mainly refers to intensity and accumulated irradiation duration. The detection time is the accurate time for recording the current detection behavior, and the voltage value is the quantized voltage value collected by the high-sensitivity sensor and the variation with the last detection time. The current value is the quantized current value measured by the high sensitivity sensor and the variation from the last detection time. The resistance value is the quantized resistance value measured by the high sensitivity sensor and the variation from the last detection time. The capacitance value is the quantized capacitance value measured by the high-sensitivity sensor and the variation from the last detection time. The inductance value is the quantized inductance value measured by the high-sensitivity sensor and the variation from the last detection time. The temperature value is the quantized temperature value measured by the high-sensitivity sensor and the variation from the last detection time. The magnetic field intensity value is the quantized magnetic field intensity value measured by the high-sensitivity sensor and the variation from the last detection time. Data flip is typically accompanied by data corruption, with a data flip state prediction default of zero, i.e., no flip, which predicts the probability of a target cell of a target chip being flipped or corrupted under current irradiation conditions based on the aforementioned statistics and various analysis methods. When the matrix is used for data recording, the data flip or damage state prediction is changed into data flip or damage state recording, which records the situation that a target unit of a target chip is flipped or damaged under the current irradiation condition. The overall trend of the data in the multidimensional matrix is observed, such as a pattern in which the performance increases or decreases with an increase in a certain parameter. Fault point positioning: matrix mapping can help to locate failure points quickly, which may be due to test errors or real failures.
in one possible implementation, the fault modes corresponding to the different types of fault points are preset, and the values of the cells in the multidimensional matrix corresponding to the color codes are preset so as to distinguish the fault degrees corresponding to the values of the cells in the multidimensional matrix of different colors.
in the implementation process, the fault modes corresponding to different types of fault points are preset, for example, the fault points may form a line or a region in the multidimensional matrix. The preset color codes correspond to the values of the cells in the multidimensional matrix, for example, the high fault degree is set to be red, and the low fault degree is set to be blue, so that different performance levels or fault degrees can be distinguished conveniently.
In one possible implementation manner, the step two further includes: and (3) cleaning data, and eliminating parameter variation values and noise which are irrelevant and abnormal in the parameter quantity collected by the high-sensitivity sensor.
in the implementation process, the second step further includes: and (3) cleaning data, removing parameter variation values and noise which are irrelevant and abnormal in the parameter quantity collected by the high-sensitivity sensor, further improving the accuracy of collecting the parameter quantity, and ensuring the quality of the collected parameter quantity.
In one possible implementation, the method further includes: and (3) signal processing, namely collecting time series data of the parameter by the high-sensitivity sensor, analyzing the time series data, and establishing a Fourier transform curve image by the frequency characteristic and the time dependence of the parameter so as to predict whether data inversion occurs at a certain moment.
In the implementation process, the method further comprises signal processing, time series data analysis of the parameter is collected for the high-sensitivity sensor, a Fourier transform curve image is built according to the frequency characteristic and the time dependence of the parameter, and the frequency characteristic and the time dependence of parameters (such as a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value, a magnetic field intensity value and the like) are analyzed to predict whether data inversion occurs at a certain moment.
in one possible implementation manner, the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the same time when the data overturn occurs are taken as input data, fault modes corresponding to different types of fault points are taken as reference data, a trained artificial neural network model is obtained, the parameter quantity and the parameter variation quantity collected by the high-sensitivity sensor are input into the artificial neural network model, and the fault modes to be generated are estimated to be taken as output data.
In the implementation process, the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the same time when the data overturn occurs are taken as input data, fault modes corresponding to different types of fault points are taken as reference data, a trained artificial neural network model is obtained, the parameter quantity and the parameter variation quantity collected by the high-sensitivity sensor are input into the artificial neural network model, and the fault modes to be generated are estimated to be taken as output data. By estimating the failure mode to be generated, the failure mode to be generated can be known in advance, and the solution is prepared in advance. The algorithm can construct an artificial neural network model for predicting the performance of the memory chip under a specific radiation environment (such as a support vector machine, a random forest and a neural network).
In one possible implementation, based on the parameter quantity and parameter variation quantity of the high-sensitivity sensor at the same time in the test process, one of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value with the largest variation degree when data inversion occurs is obtained, and irradiation conditions are adjusted to reduce the variation degree of the corresponding parameter quantity.
in the implementation process, based on the parameter quantity and parameter variation quantity of the high-sensitivity sensor at the same time in the test process, one of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value with the largest variation degree is obtained when data overturn occurs, and irradiation conditions are adjusted to reduce the variation degree of the corresponding parameter quantity. By analyzing the sensitivity of the parameters, it is determined which physical parameters have the greatest effect on the performance of the memory chip, and then the test conditions or designs can be adjusted to improve the radiation resistance of the memory chip.
In one possible implementation, the threshold interval of the parameter number is preset based on the parameter number and the parameter variation quantity of the high-sensitivity sensor at the same time in the test process, and when the parameter quantity collected by the high-sensitivity sensor exceeds the threshold interval, the external processing system sends out early warning.
In the implementation process, a threshold interval of the parameter quantity is preset based on the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the same time in the test process, and when the parameter quantity collected by the high-sensitivity sensor exceeds the threshold interval, an external processing system sends out early warning. Through setting up threshold value interval to and the mode of early warning, when parameter variation is too big, represent that the data upset degree that takes place is big, and the fault degree is higher, and external processing system sends the early warning, so that the operating personnel in time looks over.
In one possible implementation, before starting the test, the high-sensitivity sensor is calibrated based on the parameter quantity and the parameter variation quantity at the same time in the test process, so as to improve the accuracy of collecting the parameter quantity by the high-sensitivity sensor.
In the implementation process, before the test is started, the high-sensitivity sensor is calibrated based on the parameter quantity and the parameter variation quantity at the same time in the test process, so that the accuracy of collecting the parameter quantity by the high-sensitivity sensor is improved. In addition, these high sensitivity sensors also need to be maintained and calibrated regularly to ensure their accuracy and reliability throughout the test cycle. By careful design and implementation of high sensitivity sensor deployment, high quality and reliable data can be ensured during radiation testing.
In the irradiation test process of the memory chip, a radiation-resistant high-sensitivity sensor is arranged to detect the current, voltage, temperature and magnetic field of the chip, and a radiation-resistant storage unit is arranged to store and record test data, and the correlation is generated through the external environment and the detection result. Furthermore, the association relationship forms matrix mapping for searching in a large range in the set.
The steps of the method generally comprise the following key processes: sensor deployment and monitoring: a series of high sensitivity sensors are deployed in the memory chip for monitoring a plurality of physical parameters of the chip in real time. In radiation detection of memory chips, the deployment and monitoring of sensors is a critical step, as they are directly responsible for collecting physical parameter data affecting the memory performance. The following are the detailed steps and gist of sensor deployment and monitoring: 1. an appropriate sensor is selected. First, it is necessary to select sensors that can accurately measure parameters such as current, voltage, temperature, and magnetic field. These sensors must have high sensitivity and high accuracy to operate stably in a radiation environment. For different physical quantities, different types of sensors may be required, for example: a current sensor: the method is used for monitoring the current change of the chip during operation. A voltage sensor: for detecting voltage changes in the supply line or in a specific area. Temperature sensor: for measuring the temperature change of the chip in operation. Magnetic field sensor (e.g., hall effect sensor): for detecting changes in the magnetic field caused by radiation. 2. Layout and deployment of sensors. The layout of the sensor should be designed according to the structure and test requirements of the memory chip. It must be ensured that the sensor is able to cover critical measurement points and has minimal impact on the normal functioning of the chip. Possible layout schemes include: uniform layout: the sensors are uniformly distributed on the chip to obtain comprehensive data. Key layout: more sensors are arranged in critical parts of the chip or areas that are expected to be more affected by radiation. 3. And a data acquisition system. The data collected by the sensors needs to be recorded and processed by an efficient data acquisition system. This typically includes: analog/digital converter: the analog signal collected by the sensor is converted to a digital signal. A data recorder: the digital signal output by the recording sensor is usually stored in a memory with high radiation resistance. Communication interface: the collected data is sent to an external processing system for further analysis. 4. And (5) monitoring and feeding back in real time. During the test, the sensor data needs to be monitored in real time in order to immediately find any anomalies. The monitoring system should include: real-time data analysis: the sensor data is processed and interpreted on the fly. And (3) an alarm system: if abnormal data are detected, early warning can be sent out in time. Feedback loop: the test conditions or the intervention operation are automatically adjusted according to the monitoring result to prevent damage. 5. Environmental adaptability. Considering the harshness of the radiation environment, the sensor and the related circuits thereof need to have good environmental adaptability, including high temperature resistance, electromagnetic interference resistance and the like. 6. Calibration and maintenance. The sensor should be calibrated strictly before being put into use to ensure the accuracy of the data. In addition, periodic maintenance and calibration of these sensors is also required to ensure their accuracy and reliability throughout the test cycle. By careful design and implementation of sensor deployment and monitoring strategies, high quality and reliable data can be ensured during radiation testing, providing a solid foundation for further analysis and improvement of memory performance.
Data collection and recording: during irradiation, the data changes detected by the sensor are recorded by the radiation-resistant storage unit for subsequent analysis. Data collection and recording is an important component of sensor monitoring systems that ensures that all sensor data monitored can be effectively captured, recorded and used for subsequent analysis. The following are the detailed steps and gist: 1. and (5) data acquisition. Acquisition frequency: and setting a proper data acquisition frequency according to the test requirement. For rapidly changing physical parameters, such as voltage or current spikes, high frequency acquisition may be required. Whereas for relatively stable parameters, such as temperature, the acquisition frequency may be relatively low. Synchronous acquisition: in order to accurately analyze the correlation between parameters, data acquisition by multiple sensors needs to be performed simultaneously. 2. And (5) formatting data. Digitization: the acquired data typically needs to be converted to digital signals by an analog-to-digital converter for computer processing and analysis. Encoding: the digital signals are encoded in a format for storage and transmission. The encoding format requires sufficient resolution to ensure the accuracy of the data. 3. And (5) data storage. Local caching: inside the chip, the acquired data may be temporarily stored using a radiation-resistant storage unit. Persistent storage: data needs to be moved periodically from the cache to a more secure long-term storage medium, such as a radiation-resistant solid state drive or remote server. Data integrity. Error detection and correction: ensuring that the data during storage and transmission is not compromised by radiation or other external influences. And (3) checksum: a checksum is calculated for each set of data to verify that the data is complete and error free when read. 4. And (5) data transmission. Interface and communication: reliable data transmission channels are set, such as data transmission using radiation-resistant enhanced communication interfaces and protocols. Redundant transmission: to ensure that critical data is not lost, a data redundancy transmission strategy is employed. 5. And (5) data analysis preparation. Timestamp: a time stamp is added to each data sample to facilitate time-sequential association and tracking of the data during subsequent analysis. Pretreatment: the data is cleaned, removing significant noise and irrelevant information, so that valuable data is retained in the data set. 6. The data is secure. Backup: the collected data is backed up periodically to prevent accidental loss. Encryption: to prevent unauthorized access to the data during transmission or storage, the data is encrypted. 7. User interface and access. Visualization tool: an intuitive user interface is provided for monitoring data in real time, helping a user track test progress and timely discover anomalies. And (3) data access: the user is granted permission to access and download the stored data for detailed analysis. Through the steps, the data generated in the test process can be accurately and completely recorded, and solid data support is provided for subsequent data analysis, performance evaluation and fault diagnosis.
Correlation analysis: the relationship between the sensor's data changes and memory performance is analyzed. In particular, weak signal changes accompanied by data inversion are of concern. Correlation analysis is a core step of data processing aimed at finding correlations between physical parameter changes and performance manifestations within a memory chip. From this analysis, it can be identified which specific physical changes may cause data flipping or memory corruption in the radiation environment. The following are the detailed steps and gist of the association analysis: 1. data preparation. Data synchronization: ensuring that the collected data has the same point of time reference facilitates analysis of the timing relationship between the parameters. Data cleaning: outliers and noise are removed, and data quality is guaranteed, which is important for improving accuracy of correlation analysis. 2. And (5) quantitatively analyzing. The statistical method comprises the following steps: statistical methods (e.g., regression analysis, analysis of variance, etc.) are used to explore quantitative relationships between different physical parameters and memory performance. And (3) data mining: data mining techniques (e.g., cluster analysis, principal component analysis, etc.) are used to identify patterns and trends in the data. 3. Association rule mining. Pattern recognition: frequently occurring patterns of parameters are found in a large amount of test data, which patterns may be predictive of changes in memory performance. Association rule application: association rule mining algorithms are used to find meaningful associations between different physical parameters. 4. And (5) time sequence analysis. Abnormality detection: time series data of physical parameters are analyzed to predict potential data rollover or corruption by identifying abnormal patterns. And (3) signal processing: signal processing techniques (e.g., fourier transforms) are applied to analyze the frequency characteristics and time dependence of physical parameters (e.g., current, voltage). 5. Machine learning. Prediction model: machine learning algorithms (e.g., support vector machine, random forest, neural network) are used to build models that predict the performance of the memory under specific radiation environments. Characteristic engineering: features that help to improve model predictive power are selected or constructed, which may include composite indicators derived from raw sensor data. 6. Sensitivity analysis. Parameter adjustment: it is determined which physical parameters have the greatest impact on memory performance and test conditions or designs may be adjusted to improve the radiation resistance of the memory. Threshold value setting: setting critical values of physical parameters, and giving early warning when the data reach or exceed the critical values.
Matrix mapping: the association of sensor data and memory performance is mapped in a matrix so that potentially error prone or vulnerable locations can be quickly searched and identified. Matrix mapping is a data representation method that is commonly used to reduce complex data relationships or multidimensional datasets to a two-dimensional tabular form that is easy to understand. In the monitoring and analysis of the memory chip, the matrix mapping may help identify and visualize the relationships between parameters, the location of the fault points, and the pattern of radiation effects. The following are the steps and key points of matrix mapping: 1. a matrix dimension is defined. Parameter selection: it is determined which physical parameters or test conditions will constitute the rows and columns of the matrix. For example, the rows may represent different voltage levels, while the columns may represent different temperature settings. Range and resolution: for each parameter, its operating range and resolution, i.e. the specific value or interval represented by each element in the matrix, are defined. 2. And (5) data integration. Data aggregation: the data of the plurality of measurements is aggregated in accordance with the dimensions of the matrix so as to represent the measurement results under the same conditions within each cell in the matrix. And (3) standardization treatment: in order for the data to be comparable, it may be necessary to normalize it, for example by Z-score normalization or scaling the data to a fixed interval. 3. And (5) filling a matrix. Data mapping: the test results are mapped to corresponding locations of the matrix, with each cell representing a test output at a particular parameter combination. Missing data processing: possible methods to deal with the problem of missing data that may occur in the matrix include interpolation, using default values, or omitting missing data. 4. And (5) pattern recognition. And (3) fault identification: through matrix mapping, failure modes can be visually identified, e.g., failure points may form a line or a region in the matrix. Correlation analysis: the row-column relationships in the matrix can help identify potential correlations between parameters, revealing correlations between faults and specific conditions. 5. And (5) visual display. Heat map: color coding is used to represent the values in the matrix cells to facilitate distinguishing between different performance levels or levels of failure. Contour plot: adding contours on the basis of matrix mapping can help identify progressive trends in the data. 6. Analysis and interpretation. Trend analysis: the overall trend of the data in the matrix is observed, such as a pattern in which performance increases or decreases with an increase in a certain parameter. Abnormal point positioning: matrix mapping can help to locate outlier data points quickly, which may be due to test errors or real failures. 7. Application and optimization. Feedback application: the memory design or test conditions are optimized based on the analysis results of the matrix mapping. Iterative improvement: the matrix mapping results are used to guide further testing, and parameter selection and testing strategies are continually optimized through an iterative process. Through matrix mapping, complex data sets can be converted into intuitive formats, thereby speeding up problem diagnosis, providing valuable insight for design and testing of memory chips.
By the method, the fragile part of the memory under the radiation environment can be predicted and identified more accurately, and powerful data support is provided for the design and application of the memory. In addition, the method is also beneficial to simplifying the test flow and reducing the cost and difficulty when testing the memory in the radiation environment.
Claims (9)
1. a method of testing an irradiation detection memory chip, comprising:
Step one, arranging a plurality of high-sensitivity sensors on a chip to be tested to obtain parameter quantity and parameter variation quantity of the high-sensitivity sensors at the same time in the test process, wherein the parameter quantity comprises a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value, and converting analog signals collected by the plurality of high-sensitivity sensors into digital signals and transmitting the digital signals to an external processing system;
Writing input data into a plurality of identical chips to be tested by the external processing system, acquiring read output data of the chips to be tested, and recording parameter amounts and parameter variation amounts of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value and a magnetic field intensity value at the moment when the data inversion occurs in the chips to be tested is identified by the external processing system, acquiring data inversion with the occurrence frequency higher than a preset frequency value, and correspondingly acquiring the parameter amounts and parameter variation amounts of the high-sensitivity sensors at the moment, and taking the parameter amounts and the parameter variation amounts as judgment standards for judging whether the data inversion occurs in the chips to be tested;
And thirdly, marking the chip to be tested as data turnover when other identical chips to be tested have identical changes of current value, voltage value, resistance value, capacitance value, inductance value, temperature value and magnetic field intensity value.
2. The method for testing an irradiation detection memory chip according to claim 1, wherein a multi-dimensional matrix map is established between the parameter quantity collected by the high-sensitivity sensor and the chip performance to be tested, and when the same parameter variation occurs in the multi-dimensional matrix as when data inversion occurs, corresponding fault points are recorded.
3. The method according to claim 2, wherein failure modes corresponding to different types of failure points are preset, and values corresponding to cells in the multi-dimensional matrix are preset for color coding so as to distinguish failure degrees corresponding to the values of the cells in the multi-dimensional matrix with different colors.
4. The method for testing the radiation detecting memory chip according to claim 1, wherein the step two further comprises: and (3) cleaning data, and eliminating parameter variation values and noise which are irrelevant and abnormal in the parameter quantity collected by the high-sensitivity sensor.
5. the method of testing an irradiance detection memory chip of claim 1, further comprising:
And (3) signal processing, namely collecting time series data of the parameter by the high-sensitivity sensor, analyzing the time series data, and establishing a Fourier transform curve image by the frequency characteristic and the time dependence of the parameter so as to predict whether data inversion occurs at a certain moment.
6. The method for testing the irradiation detection memory chip according to claim 1, wherein the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the same time when the data inversion occurs are used as input data, fault modes corresponding to different types of fault points are used as reference data, a trained artificial neural network model is obtained, the parameter quantity and the parameter variation quantity collected by the high-sensitivity sensor are input into the artificial neural network model, and the fault modes to be generated are estimated as output data.
7. the method according to claim 1, wherein the irradiation condition is adjusted to reduce the degree of variation of the corresponding parameter number by acquiring one of a current value, a voltage value, a resistance value, a capacitance value, an inductance value, a temperature value, and a magnetic field intensity value, which have the greatest degree of variation when data inversion occurs, based on the parameter number and the parameter variation amount at the same time of the high-sensitivity sensor during the test.
8. The method for testing the irradiation detection memory chip according to claim 1, wherein the threshold interval of the parameter is preset based on the parameter quantity and the parameter variation quantity of the high-sensitivity sensor at the same time in the testing process, and the external processing system sends out the early warning when the parameter quantity collected by the high-sensitivity sensor exceeds the threshold interval.
9. The method for testing the memory chip by irradiation according to claim 1, wherein the high-sensitivity sensor is calibrated based on the parameter quantity and the parameter variation quantity at the same time in the testing process before starting the test, so as to improve the accuracy of collecting the parameter quantity by the high-sensitivity sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410190601.7A CN117766014B (en) | 2024-02-21 | 2024-02-21 | Method for testing irradiation detection memory chip |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410190601.7A CN117766014B (en) | 2024-02-21 | 2024-02-21 | Method for testing irradiation detection memory chip |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117766014A true CN117766014A (en) | 2024-03-26 |
| CN117766014B CN117766014B (en) | 2024-04-19 |
Family
ID=90326066
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410190601.7A Active CN117766014B (en) | 2024-02-21 | 2024-02-21 | Method for testing irradiation detection memory chip |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117766014B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060154642A1 (en) * | 2004-02-20 | 2006-07-13 | Scannell Robert F Jr | Medication & health, environmental, and security monitoring, alert, intervention, information and network system with associated and supporting apparatuses |
| US20140156215A1 (en) * | 2012-12-04 | 2014-06-05 | Mapmyfitness, Inc. | Gait analysis system and method |
| CN114544557A (en) * | 2022-03-03 | 2022-05-27 | 南京邮电大学 | Wide-spectrum high-sensitivity high-flux biochemical sensor and sensing method thereof |
| DE102022124361A1 (en) * | 2021-09-29 | 2023-03-30 | Nvidia Corporation | VISIBILITY ASSESSMENT USING DEEP LEARNING IN AUTONOMOUS MACHINE APPLICATIONS |
| CN117367589A (en) * | 2023-10-13 | 2024-01-09 | 国网河南省电力公司信息通信分公司 | All-weather auxiliary temperature measurement system for substation main equipment based on AI technology |
-
2024
- 2024-02-21 CN CN202410190601.7A patent/CN117766014B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060154642A1 (en) * | 2004-02-20 | 2006-07-13 | Scannell Robert F Jr | Medication & health, environmental, and security monitoring, alert, intervention, information and network system with associated and supporting apparatuses |
| US20140156215A1 (en) * | 2012-12-04 | 2014-06-05 | Mapmyfitness, Inc. | Gait analysis system and method |
| DE102022124361A1 (en) * | 2021-09-29 | 2023-03-30 | Nvidia Corporation | VISIBILITY ASSESSMENT USING DEEP LEARNING IN AUTONOMOUS MACHINE APPLICATIONS |
| CN114544557A (en) * | 2022-03-03 | 2022-05-27 | 南京邮电大学 | Wide-spectrum high-sensitivity high-flux biochemical sensor and sensing method thereof |
| CN117367589A (en) * | 2023-10-13 | 2024-01-09 | 国网河南省电力公司信息通信分公司 | All-weather auxiliary temperature measurement system for substation main equipment based on AI technology |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117766014B (en) | 2024-04-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108647136B (en) | Hard disk damage prediction method and device based on SMART information and deep learning | |
| JP5306902B2 (en) | System and method for high performance condition monitoring of asset systems | |
| CA2508008C (en) | A method for developing a unified quality assessment and providing an automated fault diagnostic tool for turbine machine systems and the like | |
| US7421351B2 (en) | Monitoring and fault detection in dynamic systems | |
| US8433472B2 (en) | Event-driven data mining method for improving fault code settings and isolating faults | |
| US20070294591A1 (en) | Method and apparatus for identifying a failure mechanism for a component in a computer system | |
| US7248939B1 (en) | Method and apparatus for multivariate fault detection and classification | |
| CN105044759B (en) | A kind of state estimation of digital nuclear detector is with ensureing maintaining method and system | |
| US20100292959A1 (en) | Telemetry data analysis using multivariate sequential probability ratio test | |
| CN112284440B (en) | Sensor data deviation self-adaptive correction method | |
| CN113255783B (en) | Sensor fault detection method and device based on unsupervised learning | |
| CN117251812A (en) | High-voltage power line operation fault detection method based on big data analysis | |
| CN117766014B (en) | Method for testing irradiation detection memory chip | |
| CN111314110B (en) | Fault early warning method for distributed system | |
| Bae et al. | Detecting abnormal behavior of automatic test equipment using autoencoder with event log data | |
| Bhargava et al. | Anomaly detection in wireless sensor networks using S-Transform in combination with SVM | |
| CN115460061A (en) | Health degree evaluation method and device based on intelligent operation and maintenance scene | |
| CN111309584A (en) | Data processing method and device, electronic equipment and storage medium | |
| CN114112390A (en) | Early fault diagnosis method for nonlinear complex system | |
| CN114371677A (en) | Industrial process state monitoring method based on spectral radius-interval principal component analysis | |
| CN117572159B (en) | Power failure detection method and system based on big data analysis | |
| Gross et al. | Machine Learning Innovation for High Accuracy Remaining Useful Life (RUL) Estimation for Critical Assets in IoT Infrastructures | |
| CN117093821B (en) | Energy efficiency and water efficiency measuring system and method for washing machine | |
| Dekate | Prognostics and engine health management of vehicle using automotive sensor systems | |
| CN116660761B (en) | Lithium ion battery detection method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |