CN115774185B - Vehicle-mounted chip DPAT detection method and device - Google Patents

Vehicle-mounted chip DPAT detection method and device Download PDF

Info

Publication number
CN115774185B
CN115774185B CN202310101574.7A CN202310101574A CN115774185B CN 115774185 B CN115774185 B CN 115774185B CN 202310101574 A CN202310101574 A CN 202310101574A CN 115774185 B CN115774185 B CN 115774185B
Authority
CN
China
Prior art keywords
production
data
test
dpat
robust
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.)
Active
Application number
CN202310101574.7A
Other languages
Chinese (zh)
Other versions
CN115774185A (en
Inventor
徐祖峰
赵伟
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Jiangsu Taizhi Technology Co ltd
Original Assignee
Jiangsu Taizhi Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Jiangsu Taizhi Technology Co ltd filed Critical Jiangsu Taizhi Technology Co ltd
Priority to CN202310101574.7A priority Critical patent/CN115774185B/en
Publication of CN115774185A publication Critical patent/CN115774185A/en
Application granted granted Critical
Publication of CN115774185B publication Critical patent/CN115774185B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Landscapes

  • Testing Or Measuring Of Semiconductors Or The Like (AREA)

Abstract

The invention discloses a DPAT (differential pulse automatic transmission) detection method and device for a vehicle-gauge chip. The method comprises the following steps: after a batch of test is completed, test data of the batch is obtained, and the test data is ensured to accord with set DPAT calculation conditions; determining production influence factors according to the test items, wherein each production influence factor X has a plurality of production objects in the test process, each production object corresponds to a group of test data, and determining the data distribution of the test data of each production object; judging whether the data distribution accords with the drift removal adjustment condition, if so, carrying out drift removal adjustment on the data set of each production object under the production influence factor X; and calculating the specification threshold value of the DPAT based on the data set after drift removal adjustment, loading the test item values of all chips in the batch, comparing the test item values with the specification threshold value, and identifying the abnormal chips. According to the invention, the influence of process parameter drift on the test item value in the chip production process is considered, and the accuracy of DPAT detection is improved through drift removal processing.

Description

Vehicle-mounted chip DPAT detection method and device
Technical Field
The invention relates to the field of chip testing, in particular to a vehicle-mounted chip DPAT detection method and device.
Background
The automobile gauge chip is a chip applied to an automobile, and compared with consumer-grade and industrial-grade chips, the automobile gauge chip faces the challenges of large cold and hot temperature range, large humidity change, more dust, harmful gas erosion, jolt, impact and the like in the use process, and has higher requirements on reliability. Moreover, the automobile is closely related to personal safety, so that no careless mistakes can be made, and the safety requirement is extremely high. Therefore, there is a more stringent quality requirement for the on-board chip, and DPPM (Defect part per million, the number of defective products per million defective opportunities) is usually controlled to be within 10 or even 0, which requires that defects and abnormal products be found out as much as possible during the chip test process, and prevented from flowing into the downstream automobile application.
For the test item of reliability of the automobile-scale chip, the U.S. automobile electronics Committee AEC-Q001 specification recommends a dynamic part average test (Dynamic part average testing, DPAT) method, the basic idea of which is as follows: and carrying out sampling test and data statistics on the products according to batches, determining a specification threshold applicable to the products in the current batch, and screening abnormal products on the products in the current batch by taking the specification threshold as a criterion. DPAT defines a batch of test item values that deviate significantly from the overall distribution as outliers. However, in an actual service scenario, the test result of the chip often generates a difference due to the drift of related process parameters, so that the data distribution is changed, and the conventional DPAT method used in a simple and general way inevitably generates missed judgment or misjudgment, so that abnormal products flow out, and uncontrollable risks are brought.
Disclosure of Invention
The invention aims to: aiming at the defects of the prior art, the invention provides a vehicle-mounted chip DPAT detection method, which improves the detection accuracy and further avoids abnormal product outflow caused by missed judgment or misjudgment.
The invention also provides a vehicle-mounted chip DPAT detection device, computer equipment and a computer storage medium.
The technical scheme is as follows: according to a first aspect of the invention, a vehicle-mounted chip DPAT detection method comprises the following steps:
after a batch of test is completed, test data of the batch is obtained, whether the test data accords with set DPAT calculation conditions is judged, and the test data accords with the set DPAT calculation conditions is ensured;
determining production influence factors according to test items, wherein each production influence factor X has a plurality of production objects in the test process, each production object corresponds to a group of test data, and determining the data distribution condition of the test data of each production object under the production influence factors;
judging whether the data distribution condition accords with a drift removal adjustment condition, if so, carrying out drift removal adjustment on the data set of each production object under the production influence factor X;
calculating a specification threshold value of the DPAT based on the data set after drift removal adjustment, loading test item values of all chips in the batch, comparing the test item values with the specification threshold value, giving new classification to the test item values exceeding the specification threshold value, and identifying the corresponding chips as abnormal products.
As a preferred embodiment, the set DPAT calculation conditions include: the number of chips with good products in the batch reaches a preset threshold, and the data of all wafers are complete.
As a preferred embodiment, the determining the data distribution of the test data of each production object under the production influence factor includes: taking 50% fractional value Q2 of data set corresponding to each production object under production influence factor X as steady average value
Figure SMS_1
The medium represents the median;
acquiring a 25% quantile value Q1 and a 75% quantile value Q3 according to the data set corresponding to each production object under the production influence factor X, and calculating a robust standard deviation value
Figure SMS_2
；
Based on robust mean RobustMean and robust standard deviationRobustSigmaThe data distribution coefficient coficient is calculated according to the following formula:
Figure SMS_3
，
the larger the Cofficient is, the larger the drift among all production objects under the production influence factor X is, and the more obvious the data is in a plurality of distributions; wherein, the liquid crystal display device comprises a liquid crystal display device,AllRobustMeansis a robust average value set, which contains robust average values calculated by data sets corresponding to various production objects, and is expressed as follows:
Figure SMS_4
wherein X is 1 RM represents a robust average value calculated from the data set corresponding to the first production object under the production influence factor X, X n RM represents a robust average value obtained by calculation of a data set corresponding to the nth production object;
AllRobustSigmasis a robust standard deviation set, which includes the robust standard deviation calculated by the data set corresponding to each production object, and is expressed as follows:
Figure SMS_5
wherein X is 1 RS represents a robust standard deviation value calculated by a data set corresponding to a first production object under the production influence factor X, X n RS represents the robust standard deviation calculated for the data set corresponding to the nth production object.
As a preferred embodiment, the determining whether the data distribution situation meets the drift removal adjustment condition includes: and judging whether the data distribution coefficient Cofficient is larger than a specified discrete threshold M, and if so, conforming to a drift removal adjustment condition.
As a preferred embodiment, the performing drift-removal adjustment on the data set includes: obtaining the average value Mean (AllRobustMeans) of each robust average value in the robust average value set and the robust average value X of the ith production object i The difference delta between the RMs,
Figure SMS_6
；
compensating the original test value OriginalValue by using the difference delta to obtain a value after deshifting:
Figure SMS_7
；
and putting test data of the chip in the batch after drift removal together for subsequent specification threshold calculation.
As a preferred embodiment, calculating the specification threshold of the DPAT based on the de-drift adjusted data set includes: according to
Figure SMS_8
Calculating the upper limit of the specification threshold according to
Figure SMS_9
A specification threshold lower limit is calculated.
As a preferred embodiment, the production influencing factors comprise test sites Site where the chip is located and a mask plate used in etching exposure.
According to a second aspect of the present invention, a vehicle-mounted chip DPAT detection apparatus includes:
the data acquisition module is used for acquiring test data of a batch after the batch is tested, judging whether the test data accords with the set DPAT calculation conditions or not, and ensuring that the test data accords with the set DPAT calculation conditions;
the data distribution determining module is used for determining production influence factors according to the test items, each production influence factor X is provided with a plurality of production objects in the test process, each production object corresponds to one group of test data, and the data distribution condition of the test data of each production object under the production influence factors is determined;
the drift removing processing module is used for judging whether the data distribution situation accords with the drift removing adjustment condition, and if so, the drift removing adjustment is carried out on the data set of each production object under the production influence factor X;
the detection and identification module is used for calculating the specification threshold value of the DPAT based on the data set after drift removal adjustment, loading the test item values of all chips in the batch, comparing the test item values with the specification threshold value, endowing new classification with the test item values exceeding the specification threshold value, and identifying the corresponding chips as abnormal products.
According to a third aspect of the invention, a computer device comprises:
one or more processors;
a memory; and
one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, which when executed by the processors implement the steps of the on-board chip DPAT detection method as described above.
According to a fourth aspect of the present invention, a computer storage medium has stored thereon a computer program which, when executed by a processor, implements the steps of the vehicle-mounted chip DPAT detection method as described above.
The beneficial effects are that: aiming at the high reliability and high safety requirements of the vehicle-mounted chip test, the invention provides a vehicle-mounted chip DPAT detection method and device, which are used for identifying factors influencing the test result as production influence factors by combing the influence of the change of the data distribution of the test result caused by the difference generated by the drift of related process parameters in the test flow, wherein each production influence factor X is provided with a plurality of production objects in the test process, each production object corresponds to a group of test data, whether the test result drifts or not is identified by determining the data distribution condition of the test data of each production object under the production influence factors, drift removal processing is carried out, DPAT detection is carried out based on the data set after drift removal adjustment, the accuracy of the chip test is improved, abnormal product outflow caused by missing judgment or misjudgment is avoided, and the product delivery quality of a chip manufacturer is improved.
Drawings
FIG. 1 is a flow chart of a DPAT detection method for a vehicle-mounted chip of the present invention;
FIG. 2 is an overall data distribution diagram of test item data in an embodiment of the present invention;
FIG. 3 is a diagram showing data distribution corresponding to two sites in an embodiment of the present invention;
FIG. 4 is a graph showing the overall data distribution after de-drifting in accordance with an embodiment of the present invention;
FIG. 5 is a Wafer diagram showing the abnormal products detected by the method in the example of the present invention.
Detailed Description
The technical scheme of the invention is further described below with reference to the accompanying drawings.
Referring to fig. 1, the vehicle-mounted chip DPAT detection method provided by the invention comprises the following steps:
step 1, obtaining test data of a batch, judging whether the test data accords with the set DPAT calculation conditions, and ensuring that the test data accords with the set DPAT calculation conditions.
After a batch test is completed, all test data files generated by the batch test are collected, typically STDF (Standard Test Data Format ) files, and all test data can be obtained after analysis and processing. The test data herein is a test item for chip functions, including current, voltage, etc., such as the common IDDQ leakage current test.
Judging whether the data accords with the set DPAT calculation condition, and if so, entering step 2. According to an embodiment of the present invention, the DPAT calculation conditions include: the number of chips with good products in the batch reaches a preset threshold, and the data of all wafers are complete. For example, good product identification in the test results is Good Die, and the DPAT calculation condition may be that the batch test contains at least 100/200/300 or other number of Good Die, while all Wafer data is complete. All Wafer data are complete, and is a link of data verification, for example, 10 wafers are tested in actual production, and then the DPAT calculation needs to be performed until the data of the 10 wafers are successfully resolved. To ensure that DPAT calculations are performed on a reasonable, efficient data set, avoiding false positives, both conditions need to be met at the same time. If the calculation conditions are not met, DPAT calculation is not carried out, and special records are carried out on the batch for subsequent manual judgment.
And 2, determining production influence factors according to the test items, wherein each production influence factor has a plurality of production objects in the test process, each production object corresponds to one group of test data, and determining the data distribution condition of the batch of test data.
Chip testing is largely divided into three categories: chip function test, performance test, reliability test. To achieve these tests, various means are required to test and detect several items on the chip, and the main means are CP test and FT test. Wafer CP (Chip Probing) is often used in functional testing and performance testing to see if the chip functions properly and to screen out faulty chips in the wafer. The CP test is to contact the chip on Wafer with a Probe (Probe), input various signals into the chip, grasp the output response of the chip and perform comparison and calculation. The equipment to be used mainly includes Automatic Test Equipment (ATE), a Probe station (Probe), and some instruments and meters, and a hardware Probe Card (Probe Card) needs to be manufactured. To save the cost of the tester, the probe card is often subjected to multi-Site testing. The post-package finished product FT test is often applied to functional tests, performance tests and reliability tests, checks whether the chip functions normally and whether defects occur in the packaging process, and helps to detect whether the chip is still functional after being subjected to a severe environment in the reliability tests. The equipment to be used mainly includes Automatic Test Equipment (ATE), a mechanical arm (Handler), and corresponding instruments, and the hardware to be manufactured is a test board (loader board or Loadboard), a Socket (Socket or base), and the like. The packaged chip is arranged on a Socket, the Loadboard is placed on a test machine, the test machine adds required voltage and current to an interface on the Loadboard, the voltage and current are loaded on the Socket through a circuit on the Loadboard, and the Socket adds the voltage and current on the chip. In order to accelerate the test efficiency, a Loadboard is often provided with a plurality of sites (test bits), so that the test of a plurality of chips can be performed simultaneously.
The chip test projects are numerous, and in an actual service scene, the test term value distribution caused by drift of some process parameters is offset/drifted, i.e. a certain deviation is generated compared with a normal value. For example, if a probe corresponding to Site is dirty, the contact is unstable, and the test result is drifted. For example, non-uniformity of exposure of reticles (also called masks, reticles, masks during exposure) in semiconductor front-end-of-line processes causes shifts in test results. In the present invention, these drifts of process parameters affecting the test results are referred to as production impact factors, i.e. production impact factors represent what factors the test results are affected by. In actual business, the production influence factors can be determined according to test items, and according to the current working practice of the applicant, the production influence factors with actual production significance mainly include Site (test bit) and Reticle (mask), and are applicable to CP test and FT test, thus being recommended factors. During the test, each production influence factor corresponds to a plurality of production objects, and each production object corresponds to a set of test data. It is of course not excluded that there is only one production object under one production influence factor. For example, in chip testing, each Tester corresponds to a Loadboard, and a plurality of sites are located on a Loadboard, which is similar to a patch board with a plurality of jacks, and each Site is placed with a chip for testing. Thus, when Site is selected as an influencing factor, the object of production is all sites (Site 1/Site 2.). If Reticle is selected as the influencing factor, the object of production is all reticles (Reticle 1/reticle2.). Each Site or each Reticle corresponds to a set of test data.
The present invention uses factor level to represent the number of objects produced under a production impact factor. For each set production influence factor X, a plurality of production objects are included
Figure SMS_10
For example, if there are 8 sites on the tester during a test, site1, site2, site8 is a production object, the factor level is 8, and each production object corresponds to a set of test data. The invention determines the data distribution of the batch test based on the test data.
The measurement of the data distribution is mainly to look at the degree of data difference, i.e. the situation that the data shift exceeds a certain threshold. According to the invention, the drift degree among all production objects under the production influence factors can be identified by acquiring the data distribution condition. Eliminating these drifts will help to improve the accuracy of the test detection. According to an embodiment of the invention, the measurement of the data distribution is represented by a data distribution coefficient coficient. The larger the cofcient, the greater the drift between individual production objects at the production impact factor X, and the more obvious the data will appear as multiple distributions. In one embodiment, the Coffinity is calculated as follows:
Figure SMS_11
，
wherein Max () represents maximum value, min () represents minimum value, robustMean () represents robust average value; allRobustMeans is a set of robust averages comprising robust averages calculated for data sets corresponding to individual production objects (robustmeans), expressed as:
Figure SMS_12
wherein X is 1 RM represents the robust average value calculated from the data set corresponding to the first production object under the production influence factor X, RM is an abbreviation of RobustMean, X n RM represents a robust average value obtained by calculation of a data set corresponding to the nth production object;
AllRobustSigmas is a set of robust standard deviations, which contains the robust standard deviations (RobustSigma) calculated for the data sets corresponding to each production object, expressed as:
Figure SMS_13
wherein X is 1 RS represents the robust standard deviation value calculated from the data set corresponding to the first production object under the production influence factor X, RS is an abbreviation of RobustSigma, X n RS represents the robust standard deviation calculated for the data set corresponding to the nth production object.
According to an embodiment of the present invention, the robust mean RobustMean and robust standard deviation RobustSigma are calculated as follows:
Figure SMS_14
，
Figure SMS_15
，
wherein Q1, Q2 and Q3 respectively represent 25%, 50% and 75% quantiles.
Figure SMS_16
Representing the 50% quantile of the corresponding test item data, the mean represents the median, e.g., for current and voltage test data, the 50% quantile of these current and voltage values are obtained as Q2, and similarly, Q1 and Q3 represent the 25% and 75% quantile of the test item data, respectively.
Taking the production influencing factors Site as an example, the tester for testing a certain batch has 2 sites, each Site corresponds to a group of test data, i.e. X 1 ，X 2 . RM and RS are calculated by taking Q1, Q2, Q3 quantile values:
Figure SMS_17
，
Figure SMS_18
，
Figure SMS_19
，
Figure SMS_20
，
the data distribution coefficients are as follows:
Figure SMS_21
。
and step 3, judging whether the data distribution situation accords with the drift removal adjustment condition, and if so, carrying out drift removal adjustment on the data set of each production object under the production influence factor X.
And (3) judging whether the data distribution coefficient Cofficient obtained by calculation in the step (2) exceeds a set discrete threshold value M, and if so, performing drift elimination adjustment on the data set of each production object under the production influence factor X. The discrete threshold may be a set of distribution coefficients coficients calculated from historical lot data over a period of time, selected based on the distribution of the set of data, such as selecting 5000 lots of data over the past half year, calculating 5000 coficients, and taking the median as the threshold. In batch testing, chips corresponding to different production objects are often consistent, but the center drift is caused by testing production process parameters, so that the influence caused by factors is eliminated by carrying out drift removal processing. After the drift removal, the test data of the batch of chips are taken together to be considered as a perfect data set.
In one embodiment, for each production object's data set under the production impact factor X, the deshifted data values are calculated according to the following method:
Figure SMS_22
，
Figure SMS_23
，
OriginalValue is the original test value, delta is the mean Mean (AllRobustMeans) of each robust mean in the robust mean set and the robust mean X of the ith production object i Difference between RMs.
Taking the production influencing factors Site as an example, the tester for testing a certain batch has 2 sites, each Site corresponds to a group of test data, i.e. X 1 、X 2 . The way to calculate the data value after the decentration drift is as follows:
Figure SMS_24
，
Figure SMS_25
。
and 4, calculating upper and lower limits by using a traditional DPAT method aiming at the data set in the step 3, loading corresponding test item values of all chips in the batch, judging whether the upper and lower limits are exceeded, if so, giving a new BIN classification to the chips, modifying test result data, and treating the chips as abnormal products in a subsequent process.
The formula for calculating the upper and lower limits by the traditional DPAT method is as follows:
Figure SMS_26
，
the calculation methods of RobustMean and RobustSigma are referred to in the previous description of step 2, and are not described here again.
In step 2, if a plurality of production influencing factors are selected, steps 2, 3 and 4 are applied one by one, and the abnormal product is identified as an abnormal product as long as one abnormality is found.
In one embodiment, after a lot of a certain vehicle-level chip is tested, 25 STDF files generated during a wafer test process are collected, and after statistics, 25 wafers of a selected lot have complete data, and a total of 10350 Good Die meet a set condition (the lot at least contains 100 Good Die), and data of a certain IDDQ leakage current test item is selected to apply DPAT detection. For comparison, the detection was performed according to the original DPAT method and the DPAT method of the present invention, respectively.
According to the original DPAT method, for convenience of viewing, the data of the test item is plotted into a histogram shown in fig. 2, and the abscissa in fig. 2 represents the test value, in which the IDDQ leakage current test item is indicated in the example, the unit mv, and the vertical axis represents the frequency, and the number of occurrences of the corresponding test value is indicated. The data for this test item was calculated to obtain the relevant parameters shown in table 1.
TABLE 1 relevant parameters for raw DPAT method detection
Data set Q1 Q2 Q3 Robust Mean Robust Sigma
Integral body 0.546 2.6203 4.1224 2.6203 2.6492
The DPAT upper and lower limits are obtained by calculation based on the relevant parameters in table 1:
Figure SMS_27
，
Figure SMS_28
。
based on the upper and lower limits of the DPAT, loading all chip data of the test item in the batch, judging whether the upper and lower limits are exceeded, and if so, obtaining an abnormal product. After analysis and calculation, the original DPAT method is adopted for detection, and no abnormal product is found.
According to the method of the invention, site is selected as a production influence factor, 2 Sites are arranged on a tester in the batch during the test process, and test item data corresponding to the 2 Sites are drawn into a histogram shown in fig. 3 for convenience of viewing, and the meaning of the abscissa is the same as that of fig. 2. And respectively calculating the test item data corresponding to the 2 sites to obtain the related parameters shown in the table 2.
Table 2 Site related parameters
Data set Q1 Q2 Q3 Robust Mean Robust Sigma
Site1 3.3693 4.1015 4.8105 4.1015 1.0676
Site2 -0.4762 0.5461 1.5327 0.5461 1.4881
Based on the relevant parameters of Site, the data distribution coefficient 2.7825 of the production influence factor Site is calculated.
Figure SMS_29
。
And judging whether the data distribution coefficient exceeds a set threshold value. The threshold value is set at 0.6 here, i.e. the center-drift distance brought about by the production influencing factor Site should not exceed 0.6 average dispersion. Since the data distribution coefficient 2.7825 exceeds the threshold value of 0.6, the center drift caused by the production influence factor Site needs to be subjected to drift removal adjustment, and the test data of the batch of chips are taken into consideration together after drift removal. The data sets for Site1 and Site2 are each adjusted as follows.
Figure SMS_30
，
Figure SMS_31
。
For ease of viewing, the data after the deghost is plotted as a histogram as shown in fig. 4. The abscissa in fig. 4 has the same meaning as fig. 2. The data for this test item was calculated to obtain the relevant parameters shown in table 3.
Table 3 relevant parameters detected by the DPAT method of the present invention
Data set Q1 Q2 Q3 Robust Mean Robust Sigma
Integral body 1.4563 2.3238 3.1472 2.3238 1.2525
Based on which new DPAT upper and lower limits are obtained by calculation:
Figure SMS_32
，
Figure SMS_33
。
and loading all chip data of the test items of the batch, and judging whether the upper limit and the lower limit are exceeded, wherein the excess is an abnormal product. After analysis and calculation, three abnormal products were found by detection using the DPAT method of the present invention, as shown in Table 4 below.
TABLE 4 detection of abnormal products found by the DPAT method of the present invention
Sequence number Test value after drift removal Raw data value Wafer Num Coordinate position
1 -6.9401 -8.7176 7 （32,17）
2 -7.0243 -8.8020 12 （28,14）
3 -7.4105 -9.1882 18 （16,15）
FIG. 5 is a Wafer plot presentation of where one of the anomalies is located, the location indicated by the arrow being the anomaly detected by the DPAT method after modification of the present invention, which was assigned to the BIN97 type in the test.
By comparing the detection results of the two methods, the improved DPAT detection method can be used for effectively finding abnormal products which cannot be found by the original DPAT method. In the actual production scene, further test and verification are carried out on the abnormal products found in the same way for a plurality of times, and the problem of the abnormal products is confirmed. If such an abnormal chip flows into the downstream automobile market, the automobile is broken down due to light weight, and accidents are caused due to heavy weight, so that immeasurable results are brought about. According to the invention, the influence of the process parameter drift on the test item value in the chip production process is considered, the accuracy of DPAT detection is improved through drift removal processing, abnormal product outflow caused by missed judgment or misjudgment is avoided, and the practical production application meaning of DPAT is enhanced.
Another embodiment of the present invention further provides a device for detecting a vehicle-mounted chip DPAT, including:
the data acquisition module is used for acquiring test data of a batch after the batch is tested, judging whether the test data accords with the set DPAT calculation conditions or not, and ensuring that the test data accords with the set DPAT calculation conditions;
the data distribution determining module is used for determining production influence factors according to the test items, each production influence factor X is provided with a plurality of production objects in the test process, each production object corresponds to one group of test data, and the data distribution condition of the test data of each production object under the production influence factors is determined;
the drift removing processing module is used for judging whether the data distribution situation accords with the drift removing adjustment condition, and if so, the drift removing adjustment is carried out on the data set of each production object under the production influence factor X;
the detection and identification module is used for calculating the specification threshold value of the DPAT based on the data set after drift removal adjustment, loading the test item values of all chips in the batch, comparing the test item values with the specification threshold value, endowing new classification with the test item values exceeding the specification threshold value, and identifying the corresponding chips as abnormal products.
It will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process of the apparatus and modules described above may refer to the corresponding process in the foregoing method embodiment, which is not repeated herein.
Embodiments of the present invention also provide a computing device comprising: the device comprises a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface complete communication with each other through the communication bus; the memory is used for storing at least one executable instruction, and the executable instruction enables the processor to execute the operation corresponding to the vehicle-mounted chip DPAT detection method.
It should be noted that, the computing device is a computing device corresponding to the above method, and all implementation manners in the above method embodiments are applicable to the embodiments of the computing device, so that the same technical effects can be achieved.
Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the vehicle-mounted chip DPAT detection method as described above.
Furthermore, it should be noted that in the apparatus and method of the present invention, it is apparent that the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent aspects of the present invention. Also, the steps of performing the series of processes described above may naturally be performed in chronological order in the order of description, but are not necessarily performed in chronological order, and some steps may be performed in parallel or independently of each other. It will be appreciated by those of ordinary skill in the art that all or any of the steps or components of the methods and apparatus of the present invention may be implemented in hardware, firmware, software, or a combination thereof in any computing device (including processors, storage media, etc.) or network of computing devices, as would be apparent to one of ordinary skill in the art after reading this description of the invention.
The object of the invention can thus also be achieved by running a program or a set of programs on any computing device. The computing device may be a well-known general purpose device. The object of the invention can thus also be achieved by merely providing a program product containing program code for implementing said method or apparatus. That is, such a program product also constitutes the present invention, and a storage medium storing such a program product also constitutes the present invention. It is apparent that the storage medium may be any known storage medium or any storage medium developed in the future.

Claims (8)

1. The vehicle-mounted chip DPAT detection method is characterized by comprising the following steps of:
after a batch of test is completed, test data of the batch is obtained, whether the test data accords with set DPAT calculation conditions is judged, and the test data accords with the set DPAT calculation conditions is ensured;
determining production influence factors according to test items, wherein each production influence factor X has a plurality of production objects in the test process, each production object corresponds to a group of test data, and determining the data distribution condition of the test data of each production object under the production influence factors;
judging whether the data distribution condition accords with a drift removal adjustment condition, if so, carrying out drift removal adjustment on the data set of each production object under the production influence factor X;
calculating a specification threshold value of the DPAT based on the data set after drift removal adjustment, loading test item values of all chips in the batch, comparing the test item values with the specification threshold value, giving new classification to the test item values exceeding the specification threshold value, and identifying the corresponding chips as abnormal products;
the determining the data distribution condition of the test data of each production object under the production influence factor comprises the following steps: taking 50% fractional value Q2 of data set corresponding to each production object under production influence factor X as steady average value
Figure QLYQS_1
The medium represents the median;
acquiring a 25% quantile value Q1 and a 75% quantile value Q3 according to the data set corresponding to each production object under the production influence factor X, and calculating a robust standard deviation value
Figure QLYQS_2
；
Based on RobustMean and RThe data distribution coefficient Cofficient was calculated from the following equation:
Figure QLYQS_3
，
the larger the Cofficient is, the larger the drift among all production objects under the production influence factor X is, and the more obvious the data is in a plurality of distributions; the AllRobustMeans is a robust average set, which includes robust averages calculated from data sets corresponding to respective production objects, and is expressed as:
Figure QLYQS_4
wherein X is 1 RM represents a robust average value calculated from the data set corresponding to the first production object under the production influence factor X, X n RM represents a robust average value obtained by calculation of a data set corresponding to the nth production object;
AllRobustSigmas is a set of robust standard deviations, which contains the robust standard deviations calculated for the data sets corresponding to the individual production objects, expressed as:
Figure QLYQS_5
wherein X is 1 RS represents a robust standard deviation value calculated by a data set corresponding to a first production object under the production influence factor X, X n RS represents the robust standard deviation value calculated by the data set corresponding to the nth production object;
performing drift-removal adjustment on the data set includes: obtaining the average value Mean (AllRobustMeans) of each robust average value in the robust average value set and the robust average value X of the ith production object i The difference delta between the RMs,
Figure QLYQS_6
；
compensating the original test value OriginalValue by using the difference delta to obtain a value after deshifting:
Figure QLYQS_7
；
and putting test data of the chip in the batch after drift removal together for subsequent specification threshold calculation.
2. The vehicle-mounted chip DPAT detection method of claim 1, wherein said set DPAT calculation conditions comprise: the number of chips with good products in the batch reaches a preset threshold, and the data of all wafers are complete.
3. The method for detecting the DPAT of the vehicle-mounted chip according to claim 1, wherein the determining whether the data distribution condition meets the drift-removal adjustment condition comprises: and judging whether the data distribution coefficient Cofficient is larger than a specified discrete threshold M, and if so, conforming to a drift removal adjustment condition.
4. The vehicle-mounted chip DPAT detection method of claim 1, wherein calculating a DPAT specification threshold based on the de-drifting adjusted data set comprises: according to
Figure QLYQS_8
Calculating the upper limit of the specification threshold according to
Figure QLYQS_9
A specification threshold lower limit is calculated.
5. The method for detecting the DPAT of the on-board chip as set forth in claim 1, wherein the production influencing factors include a test Site where the chip is located and a mask used in etching exposure.
6. The utility model provides a car rule level chip DPAT detection device which characterized in that includes:
the data acquisition module is used for acquiring test data of a batch after the batch is tested, judging whether the test data accords with the set DPAT calculation conditions or not, and ensuring that the test data accords with the set DPAT calculation conditions;
the data distribution determining module is used for determining production influence factors according to the test items, each production influence factor X is provided with a plurality of production objects in the test process, each production object corresponds to one group of test data, and the data distribution condition of the test data of each production object under the production influence factors is determined;
the drift removing processing module is used for judging whether the data distribution situation accords with the drift removing adjustment condition, and if so, the drift removing adjustment is carried out on the data set of each production object under the production influence factor X;
the detection and identification module is used for calculating the specification threshold value of the DPAT based on the data set after drift removal adjustment, loading the test item values of all chips in the batch, comparing the test item values with the specification threshold value, endowing new classification with the test item values exceeding the specification threshold value, and identifying the corresponding chips as abnormal products;
the determining the data distribution condition of the test data of each production object under the production influence factor comprises the following steps: taking 50% fractional value Q2 of data set corresponding to each production object under production influence factor X as steady average value
Figure QLYQS_10
The medium represents the median;
acquiring a 25% quantile value Q1 and a 75% quantile value Q3 according to the data set corresponding to each production object under the production influence factor X, and calculating a robust standard deviation value
Figure QLYQS_11
；
Based on RobustMean and RobustSigma, the data distribution coefficient coficient is calculated according to the following formula:
Figure QLYQS_12
，
the larger the Cofficient is, the larger the drift among all production objects under the production influence factor X is, and the more obvious the data is in a plurality of distributions; the AllRobustMeans is a robust average set, which includes robust averages calculated from data sets corresponding to respective production objects, and is expressed as:
Figure QLYQS_13
wherein X is 1 RM represents a robust average value calculated from the data set corresponding to the first production object under the production influence factor X, X n RM represents a robust average value obtained by calculation of a data set corresponding to the nth production object;
AllRobustSigmas is a set of robust standard deviations, which contains the robust standard deviations calculated for the data sets corresponding to the individual production objects, expressed as:
Figure QLYQS_14
wherein X is 1 RS represents a robust standard deviation value calculated by a data set corresponding to a first production object under the production influence factor X, X n RS represents the robust standard deviation value calculated by the data set corresponding to the nth production object; />
Performing drift-removal adjustment on the data set includes: obtaining the average value Mean (AllRobustMeans) of each robust average value in the robust average value set and the robust average value X of the ith production object i The difference delta between the RMs,
Figure QLYQS_15
；
compensating the original test value OriginalValue by using the difference delta to obtain a value after deshifting:
Figure QLYQS_16
；
and putting test data of the chip in the batch after drift removal together for subsequent specification threshold calculation.
7. A computer device, comprising:
one or more processors;
a memory; and
one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, which when executed by the processor implement the steps of the vehicle-mounted chip DPAT detection method of any one of claims 1-5.
8. A computer storage medium having stored thereon a computer program, which when executed by a processor performs the steps of the vehicle-mounted chip DPAT detection method according to any one of claims 1-5.
CN202310101574.7A 2023-02-13 2023-02-13 Vehicle-mounted chip DPAT detection method and device Active CN115774185B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310101574.7A CN115774185B (en) 2023-02-13 2023-02-13 Vehicle-mounted chip DPAT detection method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310101574.7A CN115774185B (en) 2023-02-13 2023-02-13 Vehicle-mounted chip DPAT detection method and device

Publications (2)

Publication Number Publication Date
CN115774185A CN115774185A (en) 2023-03-10
CN115774185B true CN115774185B (en) 2023-05-05

Family

ID=85393568

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310101574.7A Active CN115774185B (en) 2023-02-13 2023-02-13 Vehicle-mounted chip DPAT detection method and device

Country Status (1)

Country Link
CN (1) CN115774185B (en)

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102017200826A1 (en) * 2017-01-19 2018-07-19 Conti Temic Microelectronic Gmbh Method for operating a monitoring device of a data network of a motor vehicle and monitoring device, control device and motor vehicle
US10393802B2 (en) * 2017-06-14 2019-08-27 Nuvoton Technology Corporation System and method for adaptive testing of semiconductor product
CN110596566B (en) * 2018-06-12 2022-03-04 北京华峰测控技术股份有限公司 DPAT (dual port automatic test) method for ATE (automatic test equipment) system
CN110458188A (en) * 2019-06-27 2019-11-15 精锐视觉智能科技（深圳）有限公司 Industrial vision detection data processing method, device, storage medium and terminal device
CN111220889B (en) * 2020-01-02 2022-08-12 长江存储科技有限责任公司 Wafer test data processing method and equipment
CN111626351B (en) * 2020-05-26 2024-03-22 清华大学 Method and system for acquiring concept drift amount of data distribution
CN114254261A (en) * 2020-09-23 2022-03-29 爱德万测试股份有限公司 Method and system for detecting product test data, electronic device and storage medium
CN112397409A (en) * 2020-11-24 2021-02-23 安测半导体技术(江苏)有限公司 Chip wafer test data analysis method and system
CN113407219A (en) * 2021-07-07 2021-09-17 安测半导体技术(江苏)有限公司 Method and system for updating threshold of semiconductor test program
CN114020971A (en) * 2021-11-05 2022-02-08 光大科技有限公司 Abnormal data detection method and device
CN114201350A (en) * 2021-12-29 2022-03-18 上海赛美特软件科技有限公司 Wafer chip testing method and device, electronic equipment and storage medium
CN114755552A (en) * 2022-05-11 2022-07-15 通富微电子股份有限公司 Semiconductor device testing method and device, electronic device, and storage medium

Also Published As

Publication number Publication date
CN115774185A (en) 2023-03-10

Similar Documents

Publication Publication Date Title
US7386418B2 (en) Yield analysis method
US7117057B1 (en) Yield patrolling system
US7096140B2 (en) Test system, test method and test program for an integrated circuit by IDDQ testing
US20060236184A1 (en) Fault detecting method and layout method for semiconductor integrated circuit
US8054097B2 (en) Method and system for automatically managing probe mark shifts
CN110596566B (en) DPAT (dual port automatic test) method for ATE (automatic test equipment) system
JP2010197385A (en) Method and apparatus for data analysis
CN111653500A (en) Method for judging wafer yield loss
CN101290901A (en) Wafer quality analysis method and device
CN115774185B (en) Vehicle-mounted chip DPAT detection method and device
JP2004047542A (en) Chip quality determining method, chip quality determining program, marking mechanism using the program, and fault generation analyzing method of wafer
CN111044878A (en) Integrated circuit testing and monitoring method based on ATE system
US6681361B1 (en) Semiconductor device inspection apparatus and semiconductor device inspection method
WO1995035544A1 (en) System and method for inspection of products with warranties
CN111562503B (en) Method for analyzing and processing failure of lithium ion battery charging and discharging equipment
US7230442B2 (en) Semi-conductor component testing process and system for testing semi-conductor components
US7265568B2 (en) Semi-conductor component test process and a system for testing semi-conductor components
CN112346920A (en) Integrated circuit test data analysis method and system
CN114461457A (en) Detection method and detection device for wafer tester
TWI389245B (en) Chip sorter with prompt chip pre-position and optical examining process thereof
CN112462233B (en) Site control method and system in integrated circuit test
CN116167313B (en) Training data generation method and system for integrated circuit design
CN113393422B (en) Method and device for determining probe card abnormity, terminal equipment and storage medium
CN114839514B (en) Dynamic optimization method and system for chip test engineering
CN118280881A (en) Method for monitoring and calculating equipment particles

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