Nondestructive testing method for computer memory bank

Abstract

The invention discloses a nondestructive testing method for a computer memory bank, and relates to the technical field of memory bank detection.The nondestructive testing method comprises the steps that firstly, a pre-built camera is used for shooting color images on the two sides of a golden finger of a to-be-tested memory bank respectively, the two shot color images are fitted on the same color image, and then the color images are converted into gray images; analyzing the obtained grayscale image, screening out an abnormal memory bank corresponding to the defective golden finger, performing preliminary elimination operation on the abnormal memory bank, and entering the next step for the rest of memory banks to be detected; an oscilloscope is adopted to monitor the memory bank to be detected, a clock signal curve graph is obtained, standard and substandard memory banks are analyzed and calibrated in combination with the standard capacitance characteristic curve graph, the substandard memory banks are removed, and the standard memory banks are output.

Description

Technical Field

[0001] The present invention belongs to the technical field of memory module detection. Specifically, it relates to a non-destructive detection method for computer memory modules. Background Art

[0002] As one of the crucial basic components in a computer system, the memory module is a bridge for communication with the CPU. Whether it works properly and its performance level have a great impact on the computer. During the operation of the computer, the execution of all programs needs to be carried out in the memory. Therefore, the stability, reliability, and compatibility of the memory module directly determine the overall performance of the computer system.

[0003] When detecting memory modules before leaving the factory, the prior art mostly uses manual methods to detect the appearance of memory modules, with low efficiency and poor detection quality. Secondly, it only focuses on testing basic functions such as read and write tests of memory modules. Such a detection method will cause a large number of memory modules with performance different from normal memory modules but passing the read and write tests to flow into the market. The inflow of such memory modules into the market will not only affect consumers in terms of use but also further affect the reliability and popularity of the memory module brand. Therefore, in order to solve the above problems, the present invention proposes a non-destructive detection method for computer memory modules. Summary of the Invention

[0004] Aiming at the deficiencies of the prior art, the present invention provides a non-destructive detection method for computer memory modules, which solves the problems of difficult quality inspection and low quality inspection accuracy before the memory modules leave the factory in the prior art.

[0005] The object of the present invention can be achieved by the following technical solutions: A non-destructive detection method for computer memory modules includes the following steps: Step 1: Obtain the color images of the gold fingers on both sides of the memory module to be tested, fit the two color images on the same color image, and then convert it into a grayscale image; Analyze the grayscale image, screen out abnormal memory modules, and perform rejection operations. The remaining memory modules to be tested enter the next step; Step 2: Use an oscilloscope to monitor several memory modules of the same type to be tested, obtain the clock signals of the processing process Q of several memory modules to be tested within a monitoring period, and generate a standard clock signal change curve graph; Step 3: Use an oscilloscope to monitor the memory modules to be tested before leaving the factory, obtain the clock signals of the processing process Q of the memory modules to be tested, generate a to-be-tested clock signal change curve graph, and analyze it in combination with the constructed standard capacitance characteristic change curve graph to calibrate qualified memory modules and unqualified memory modules; Perform rejection operations on unqualified memory modules and output qualified memory modules.

[0006] As a further solution of the present invention, the specific method for screening and removing abnormal memory modules in step one is as follows: Obtain the color images of the gold fingers on both sides of any memory module to be tested, and fit the two color images into one color image in the way that the gold fingers are connected end to end, and convert it into a grayscale image ; Extract the edge images of all the gold fingers in the grayscale image through an edge detection algorithm, denoted as ; ; Using as the background, remove from through the pixel value subtraction method to obtain the grayscale image after removing the edge image of the gold finger, which only contains the grayscale pixels of the gold finger itself in the memory module; Eliminate the blank pixel points in the grayscale image after removing the edge image so that the adjacent gold finger bodies are closely connected to obtain the final body grayscale image ; Perform defect analysis on the pixel values of the body grayscale image to determine whether there are defects in the gold fingers of the memory module to be tested associated with the body grayscale image . If there are defects, regard this memory module as an abnormal memory module and perform the removal operation.

[0007] As a further solution of the present invention, the specific method for performing defect analysis on the pixel values of the body grayscale image is as follows: Determine the number of rows M and the number of columns N of the pixel points in the body grayscale image , traverse all the pixel points in , extract the pixel values and accumulate them, and then divide the accumulated result by the number of pixel points in to obtain the average grayscale value of ; ; Combine the average grayscale value to calculate the standard deviation of the grayscale values of the body grayscale image ; ; According to , define the grayscale value interval of the gold finger pixel points , where is the minimum value in the interval, is the maximum value in the interval, and K is a preset adjustment coefficient; Starting from the first pixel point in Start traversing to the last pixel point , and obtain any one of the pixel points 's pixel value . If , then consider the pixel point as a normal pixel point, otherwise it is an abnormal pixel point; If it is determined that is an abnormal pixel point, then immediately obtain all adjacent pixel points, and determine whether there are abnormal pixel points among the adjacent pixel points. If there are abnormal pixel points among the adjacent pixel points, then extract the abnormal pixel points in the determined adjacent pixel points together with , and extract the adjacent pixel points of the abnormal pixel points different from the pixel point except the pixel point , and determine whether there are other abnormal pixel points; and so on, until there are no abnormal pixel points in the adjacent pixel points of all the determined abnormal pixel points to stop; For the extracted several abnormal pixel points, fit them into an abnormal area, and determine the nature of the abnormal area by analyzing the area S, perimeter L, and aspect ratio B of the determined abnormal area.

[0008] As a further solution of the present invention, the specific way to determine the nature of the abnormal area is: Obtain the area threshold preset by the operator , compare the area S of this abnormal area with the area threshold . If , then obtain the perimeter threshold preset by the operator , compare the perimeter L of this abnormal area with the perimeter threshold . If , then compare the aspect ratio B of the abnormal area with the aspect ratio threshold preset by the operator . If , then regard the abnormal area as noise; If any of the above comparison processes does not meet the conditions, then regard this abnormal area as a surface defect; Repeat the above steps, and count the sum of the areas of all the abnormal areas that are surface defects , and determine the proportion in the body grayscale image . If the proportion exceeds , then regard it as a defect in the gold finger associated with the body grayscale image , and regard the memory module under test where the gold finger is located as an abnormal memory module, where is the preset value of the operator.

[0009] As a further solution of the present invention, the determination method of the monitoring period in step two is as follows: Select several memory modules of the same type and process the same process Q. For each memory module, take the start time of the process as the initial moment and the end time of the process as the end moment, record the time interval between the two moments, and take the average of all time intervals as a monitoring period.

[0010] As a further solution of the present invention, the specific method for obtaining the clock signals of several memory modules to be tested processing process Q within a monitoring period and generating a standard clock signal change curve is as follows: Use an oscilloscope to monitor the clock signals of several memory modules of the same type within the monitoring period, and extract the fundamental wave amplitude, frequency, and period from the clock signals; Determine several different fundamental wave amplitudes, frequencies, and periods at different time points within the same monitoring period, and determine the associated average fundamental wave amplitude, average frequency, and average period by taking the average method to generate a standard clock signal change curve ; Based on the determined standard clock signal change curve , place it in a two-dimensional coordinate system with the time line as the horizontal axis and the fundamental wave amplitude as the vertical axis, and the scale of the coordinate axes is determined by the operator.

[0011] As a further solution of the present invention, the change law and change period of the curve in the standard clock signal change curve are consistent with the average frequency and average period.

[0012] As a further solution of the present invention, the specific method for generating a change curve of the clock signal to be tested, analyzing it in combination with the constructed standard capacitance characteristic change curve, and calibrating qualified and unqualified memory modules is as follows: Generate a change curve of the clock signal to be tested for any memory module to be tested according to the method of generating ; Place the obtained in the two-dimensional coordinate system where is located, determine the time point corresponding to the first peak in the standard clock signal change curve , and the time point corresponding to the first peak in , and use to obtain the time difference between the first peaks of and ; and ; ; According to the above method, calculate the and The time difference for each group corresponding to the wave crest is obtained, and a total of time differences are obtained, which are successively denoted as . Arranged in ascending order according to the time difference values, a time difference sequence is obtained; Taking the time line as the horizontal axis and the value of the time difference as the vertical axis, a two-dimensional coordinate system is constructed, and the time difference sequence is marked in the two-dimensional coordinate system to obtain marked data points. Starting from the first data point, the subsequent data points are successively connected by short lines to obtain a broken line Z. Then, a straight line parallel to the vertical axis and passing through the first data point and a straight line parallel to the vertical axis and passing through the last data point are made. Calculate the area of the closed region formed by the horizontal axis, the broken line Z, the straight line and the straight line , which is denoted as ; Compare the calculated area of the closed region with the area threshold of the closed region preset by the operator . If , then the memory module to be tested corresponding to the broken line Z is calibrated as a non-compliant memory module. Otherwise, the memory module to be tested corresponding to the broken line Z is calibrated as a compliant memory module.

[0013] As a further solution of the present invention, the non-compliant memory modules are subjected to the rejection operation again, not output, and the operator is notified; The compliant memory modules are output.

[0014] Advantages of the present invention: (1) By combining appearance detection and clock signal monitoring, the present invention forms a comprehensive memory module quality inspection process; in appearance detection, image processing technology is used to accurately extract the main body part of the gold finger and identify abnormal areas, objectively and accurately detect minor appearance defects, and reduce manual detection errors; in clock signal monitoring, a standard curve is generated based on a large amount of data of the same type of memory modules, providing an accurate comparison benchmark to judge whether the clock signal of the memory module is normal. This method improves the comprehensiveness and accuracy of detection, and also improves the detection efficiency through an automated process, reducing the manual operation time; in addition, strict quality control reduces the after-sales risk caused by product problems, enhances the user's trust in the product, enhances market competitiveness, and ensures that the memory modules leaving the factory meet high-quality standards in terms of appearance and performance, providing more reliable products for users; (2) By analyzing the clock signal of the memory module to be tested, the present invention generates a curve graph of the clock signal change of the memory module to be tested, and compares it with the standard curve graph. By constructing the concept of the enclosed area, a quantitative and intuitive index is provided for the performance evaluation of the memory module; enabling the operator to clearly understand the stability of the memory module performance. The larger the area, the greater the degree of change in the time difference sequence, and the more unstable the memory module performance. This intuitive determination method facilitates the operator to make decisions quickly and improves the quality control efficiency in the production process; (3) The present invention performs defect detection on the gold finger part of the memory module through an automated process. By calculating the average gray value and standard deviation of the grayscale image of the main body, the grayscale characteristics of the gold finger can be quantitatively analyzed, and the grayscale value range of the pixel points can be further determined, so as to more accurately identify abnormal pixel points. Then, by using the method of connected domain analysis for the abnormal pixel points, the abnormal area can be accurately extracted, and by analyzing the characteristics such as the area, perimeter, and aspect ratio of the abnormal area, its nature can be further determined. This multi-dimensional analysis method effectively avoids misjudgment and missed judgment, ensuring the authenticity and reliability of the detection results; providing an efficient, reliable, and economical technical means for the production and quality control of enterprises, and having strong practical value and promotion significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 is a schematic flow chart of the method of the present invention; Figure 2 is a schematic flow chart of the method described in Embodiment 2 of the present invention; Figure 3 is a schematic flow chart of the method described in Embodiment 3 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0018] Embodiment 1

[0019] A non-destructive testing method for a computer memory module, as Figure 1 shown, specifically includes the following: After the memory module to be shipped enters the quality inspection area, it is first necessary to conduct an appearance inspection. When the present invention processes the appearance inspection of the memory module, it mainly targets the gold finger area on the memory module, that is, the interface part between the memory module and the computer motherboard. Through the cameras installed on both sides of the memory module, the gold fingers on both sides of the memory module to be tested are photographed to obtain two color images of the current memory module to be tested.

[0020] The environment of the quality inspection area should be stable to ensure that the images captured by the cameras will not be deviated due to environmental factors when taking pictures.

[0021] When the camera takes a color image of the gold fingers of the memory module, it should be ensured that the light source can evenly illuminate the surface of the gold fingers, and the light sources on both sides do not affect each other, ensuring that the detailed features of the gold fingers in the captured color image are clearly visible; and when taking pictures of each memory module to be tested, the installation position is appropriate and constant.

[0022] Through the image fitting technology, the memory modules in the two color images are connected end to end and fitted into a color image, and then using grayscale conversion, the color image is converted into a grayscale image, denoted as ; while reducing the data volume, the main feature information of the color image is retained, which is convenient for subsequent image analysis and processing.

[0023] At this time, the grayscale image only contains the images of the gold fingers, but the adjacent gold fingers are not tightly connected, but are combined through the PCB board. Therefore, it is necessary to remove the part including the PCB board from the original grayscale image. The edge detection algorithm can extract the edge part from the grayscale image. By removing the edge part in the original grayscale image, a grayscale image only containing the main body part of the gold fingers is obtained. Then, the blank pixel points after removing the edge part are removed, so that the adjacent gold fingers are tightly connected, and the final grayscale image is obtained, denoted as the main body grayscale image ; Based on the determined main body grayscale image , calculate the average value and variance of the grayscale pixels of all pixel points in the main body grayscale image ; With the characteristic that the pixels of the main body of the gold finger are in a relatively stable pixel value range, by combining the average value and variance, the normal grayscale pixel value range is calibrated, from the main body grayscale image Screen out abnormal pixel points, analyze the pixel points within the neighborhood of the abnormal pixel points to identify whether there are other abnormal pixel points within the neighborhood of the abnormal pixel point. Finally, perform an extraction operation on all the obtained abnormal pixel points, connect all the abnormal pixel points, combine them into an abnormal area, and through further analysis of the abnormal area, determine the nature of the abnormal area, and judge whether the currently tested memory module is an abnormal memory module. If so, remove the abnormal memory module; if not, proceed to the next step.

[0024] For a number of memory modules of the same type to be tested that have passed the appearance inspection, use an oscilloscope to monitor the clock signal of the memory module to be tested in real time during the process of processing the same process Q within a monitoring period. The determination of the monitoring period is as follows: Take the time point when any memory module to be tested starts to process process Q as the initial moment. When the memory module to be tested finishes processing process Q, record the current time point as the end moment, calculate the time interval between the start moment and the end moment, record the time intervals between the start moments and the end moments of a number of memory modules to be tested, and take the average value. Take the average value of the number of time intervals as a monitoring period. Use an oscilloscope to monitor the clock signals of a number of memory modules of the same type to be tested during the process of the same process Q, obtain several groups of clock signals, and respectively extract the fundamental wave amplitude, frequency, and period in several groups of clock signals. And again, through the method of taking the average value, obtain the associated average fundamental wave amplitude, average frequency, and average period, and generate a corresponding standard clock signal change curve graph. ; The method of generating the standard clock signal change curve graph can be obtained by fitting the obtained average fundamental wave amplitude, average frequency, and average period through the prior art, and no more details will be elaborated here. It should be determined here that the change law and change period of the curve in the standard clock signal change curve graph are consistent with the average frequency and average period.

[0025] Then, using the time line as the horizontal axis and the fundamental wave amplitude as the vertical axis, construct a two-dimensional coordinate system, and place the determined standard clock signal change curve graph in the constructed two-dimensional coordinate system, and proceed to the next step.

[0026] Then obtain any memory module to be tested that is to be shipped out, use an oscilloscope to detect the memory module to be tested according to the above method, obtain the clock signal of the memory module to be tested during the process of processing process Q, and generate a change curve graph of the clock signal to be tested. ; Then fit the change curve graph of the clock signal to be tested to the standard clock signal change curve graph Analyze in the two-dimensional coordinate system where it is located, calibrate whether the current memory module to be tested meets the standard. If it is a qualified memory module, output it; if the memory module to be tested is unqualified, do not output it.

[0027] In this embodiment, the method of the present invention is divided into two parts: appearance detection and clock signal monitoring. In terms of appearance detection, after the memory module enters the quality inspection area, color images of the gold finger area are taken by cameras on both sides, and a color image is generated through image fitting and then converted into a grayscale image; the PCB board part is removed by using an edge detection algorithm to obtain a grayscale image containing only the gold finger body, calculate the average value and variance of its pixel points, calibrate the normal grayscale pixel value range, screen and analyze abnormal pixel points, and judge whether the memory module is abnormal.

[0028] In terms of clock signal monitoring, use an oscilloscope to monitor the clock signal during the processing process of the memory module, determine the monitoring period, extract parameters such as the fundamental wave amplitude, frequency, and period, and generate a standard clock signal change curve graph; finally, compare the clock signal change curve graph of the memory module to be tested with the standard curve graph to judge whether the memory module to be tested meets the standard and decide whether to leave the factory.

[0029] Embodiment 2 On the basis of Embodiment 1, this embodiment further discloses a method for appearance detection of the gold finger part of a memory module, as Figure 2 shown, which specifically includes the following: Based on the method described in Embodiment 1, a grayscale image of the gold finger of the memory module to be tested can be obtained , and use the edge detection algorithm in the prior art to analyze the grayscale image , and extract the edge images of all gold fingers therein, denoted as ; Then regard the original grayscale image as the background image, regard the edge image as the part to be removed from the original background image, and through the method of subtracting pixel values , remove the edge image from the grayscale image , and obtain a grayscale image containing only the grayscale pixels of each individual gold finger itself in the memory module; At this time, there are blank pixel points after the removal operation between every two adjacent gold fingers. Eliminate the blank pixel points and closely connect the adjacent gold fingers to obtain a final grayscale image in which all gold fingers are closely connected, which is called the body grayscale image and denoted as ; If there are no scratches or stains on the original gold finger, the body grayscale image The pixel values are considered to be within a certain range. If there are scratches or other obvious stains on the original gold finger, abnormal pixel points will appear in the body grayscale image and the following operations are performed according to this principle: Perform abnormal analysis on the pixel points in the body grayscale image Based on the determined body grayscale image with the number of pixel rows as M and the number of pixel columns as N , use the formula: Calculate the average grayscale value of the body grayscale image , where represents the pixel value of the pixel point at the i-th row and j-th column in the body grayscale image , M and N respectively represent the number of rows and columns of the body grayscale image , i and j are both positive integer counting indices, starting from 1, and ; After obtaining the average grayscale value of the body grayscale image , then according to the formula: Calculate the standard deviation of the grayscale values of the body grayscale image . The larger the standard deviation, the greater the difference in pixel grayscale values in the body grayscale image , which also means that there may be abnormal pixel points in the gold finger associated with the body grayscale image ; Define the grayscale value range of the gold finger pixel points applicable to the current type of memory module , where the method of defining is: 、Define 's method is: , where is the minimum value of the grayscale value range of the gold finger pixel points, is the maximum value of the grayscale value range of the gold finger pixel points, and K is an adjustment coefficient, and its value is determined by the operator in combination with actual requirements; Pixel points with pixel values within this grayscale value range of the gold finger pixel points are all regarded as normal pixel points, otherwise the pixel point is regarded as an abnormal pixel point.

[0030] For the pixel points in the body grayscale image , starting from the first pixel point until the last pixel point , for any pixel point among the pixel points, extract its pixel value And determine whether it is within the gray value range of the FPC finger pixel points If then determine that the current pixel point pixel value is within the gray value range of the FPC finger pixel points within, this pixel point is a normal pixel point, otherwise determine that this pixel point is an abnormal pixel point; If it has been determined that the pixel point is an abnormal pixel point, immediately obtain the 8 pixel points in 8 directions within the neighborhood of this abnormal pixel point, and respectively judge whether the pixel values of these 8 pixel points are located within the gray value range of the FPC finger pixel points within. If the pixel value of any pixel point is not located within the gray value range of the FPC finger pixel points within, then determine whether there are abnormal pixel points among the pixel points adjacent to the pixel point At this time, the abnormal pixel points that are determined to be within the neighborhood of the pixel point and different from the pixel point together with the pixel point are subjected to an extraction operation; Then use the abnormal pixel point that is within the neighborhood of the pixel point and different from the pixel point as the central pixel point, continue to obtain the 8 pixel points in 8 directions within the neighborhood of this central pixel point, and remove the pixel points that have been determined to be abnormal pixel points from the 8 pixel points , to obtain the remaining 7 pixel points, repeat the above steps to judge whether there are abnormal pixel points among the neighborhood pixel points of this central pixel point until there are no abnormal pixel points in the adjacent pixel points of all the locked abnormal pixel points; During the subsequent traversal from the first pixel point to the last pixel point process, for the pixel points that have been determined to be abnormal pixel points, skip this pixel point and continue to judge the next pixel point; After the traversal operation is completed, extract all the determined abnormal pixel points, connect the adjacent pixel points in sequence, fit them into 1 to several abnormal regions, and obtain the area S, perimeter L and aspect ratio B of any one of the abnormal regions (aspect ratio = maximum row span of the abnormal region / maximum column span of the abnormal region).

[0031] After determining the abnormal region, it cannot be directly determined that there is a defect in the FPC finger where this abnormal region is located, because the abnormal region may be caused by noise. In order to more accurately judge the nature of the abnormal region, it is determined in combination with the following points: Obtain the area threshold associated with the abnormal region formulated by the operator based on experience , perimeter threshold , aspect ratio threshold ; Extract the area S, perimeter L, and aspect ratio B of any abnormal region, and compare the area S of the abnormal region with the area threshold . If the area S of the abnormal region is less than the area threshold , then continue to compare the perimeter L of the abnormal region with the perimeter threshold . If the perimeter L of the abnormal region is less than the perimeter threshold , then continue to compare the aspect ratio B of the abnormal region with the aspect ratio threshold . If the aspect ratio B of the abnormal region is less than the aspect ratio threshold , consider the determined abnormal region as noise. If any parameter does not meet the threshold set by the operator during this process, consider the abnormal region as a surface defect; For several abnormal regions, use the above method to make determinations in sequence, obtain the abnormal regions determined as surface defects, and calculate the sum of the areas of all abnormal regions. Calculate the proportion of the sum of the areas of the abnormal regions in the entire grayscale image of the main body . If the proportion exceeds , then consider that there is a defect in the gold finger associated with the grayscale image of the main body , and consider the memory module under test corresponding to the gold finger as an abnormal memory module and remove it from the memory modules under test. Among them, is a preset value set by the operator and is determined according to actual requirements; In this embodiment, through edge detection and cropping operations on the grayscale image, a grayscale image of the tightly connected gold finger main body is obtained. Then, calculate the average grayscale value and standard deviation of the image, determine the grayscale value range of the gold finger pixel points, screen out the abnormal pixel points, and form abnormal regions by connecting the abnormal pixel points through neighborhood analysis. Finally, combine features such as area, perimeter, and aspect ratio with the thresholds set by the operator for verification to determine whether the abnormal region is a surface defect. If the proportion of the defect area exceeds the preset value, determine that the corresponding memory module is an abnormal memory module and remove it from the memory modules under test.

[0032] Embodiment 3 On the basis of Embodiment 1, this embodiment further discloses a method for analyzing the clock signal of a memory module under test and calibrating qualified and unqualified memory modules, as shown in Figure 3 , specifically including the following: Based on the determined curve graph of the clock signal change of the memory module under test , place the curve graph of the clock signal change of the memory module under test on the standard curve graph of the clock signal change In the two-dimensional coordinate system where it is located, due to the characteristics of the fundamental wave amplitude in the clock signal, the curve graph of the fitted clock signal change is determined to be a regular waveform graph similar to the sine function and cosine function. The standard clock signal change curve graph and the curve graph of the clock signal to be measured After being placed in the same two-dimensional coordinate system, obtain the time point on the horizontal axis corresponding to the first peak in the standard clock signal change curve graph , and similarly obtain the time point on the horizontal axis corresponding to the first peak in the curve graph of the clock signal to be measured , ; ; By adopting The calculation method to obtain the time difference between the first peaks of the standard clock signal change curve graph and the curve graph of the clock signal to be measured ; According to the above method, obtain the time point on the horizontal axis corresponding to the second peak of the standard clock signal change curve graph and the time point on the horizontal axis corresponding to the second peak in the curve graph of the clock signal to be measured , calculate the time difference again, and so on, until obtaining the time difference between each group of corresponding peaks in the standard clock signal change curve graph and the curve graph of the clock signal to be measured and the curve graph of the clock signal to be measured For the a groups of corresponding peaks, a time differences are obtained in total. Record them in the order of the time differences between the peaks (that is, the order of the time line on the horizontal axis of the two-dimensional coordinate system) as , and then sort the obtained a time differences in ascending order to obtain the time difference sequence ; The greater the degree of fluctuation of the time difference sequence, the worse the synchronization between the clock signal of the current memory module to be measured during the processing of process Q and the standard clock signal, which means the more unstable the working state of the current memory module to be measured during the processing of process Q. On the contrary, if the degree of fluctuation of the time difference sequence is relatively stable, it further indicates that the working state of the current memory module to be measured is relatively stable. Therefore, calculating the degree of fluctuation of the time difference of the peaks can quantify the working stability of the memory module to be measured; Using the sorting order (time line order) of the time difference sequence as the horizontal axis and the value of the time difference as the vertical axis, construct a new two-dimensional coordinate system, and use the time difference sequence Each time difference is marked on this two-dimensional coordinate system in the form of data points, and a total of a marked data points are obtained. Starting from the first data point, short lines are connected to the second data point, and then short lines are successively connected to the subsequent data points, finally passing through a marked data points to obtain a broken line representing the degree of fluctuation of the data points, denoted as Z; Construct a straight line passing through the first data point and perpendicular to the horizontal axis and parallel to the vertical axis, denoted as ; then construct a straight line passing through the last data point and perpendicular to the horizontal axis and parallel to the vertical axis, denoted as ; At this time, the broken line Z and the straight line , the straight line and the horizontal axis will form a closed area. Calculate the area of this closed area, denoted as ; The larger the area of the closed area, the greater the degree of change of the time difference sequence , and it further shows that there is an unstable situation in the change of the fundamental wave amplitude of the measured clock signal change curve compared with the standard clock signal change curve , which means that the performance of the corresponding memory module has a certain gap compared with the normal memory module; Then obtain the closed area threshold formulated by the operator according to actual experience. Compare the calculated closed area with the closed area threshold . If the closed area exceeds the closed area threshold , then calibrate the measured memory module associated with the measured clock signal change curve of this closed area as a non-compliant memory module. If the closed area is less than or equal to the closed area threshold , then calibrate the measured memory module associated with the measured clock signal change curve of this closed area as a compliant memory module; Based on the determined non-compliant memory modules, eliminate them from the measured memory modules, indicating that they have not passed the quality inspection and are not output. Output the compliant memory modules, indicating that they have passed all quality inspections and can be shipped.

[0033] Embodiment 4 The technical solution of this embodiment is to combine and implement the solutions of the above Embodiment 2 and Embodiment 3.

[0034] Some of the data in the formulas described above are numerically calculated after removing their dimensions, and the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0035] The above content is only an example and illustration of the present invention. Those skilled in the art of this technology can make various modifications or supplements to the specific embodiments described, or use similar methods to replace them. As long as they do not deviate from the invention or exceed the scope defined by this claim book, they should all fall within the protection scope of the present invention.

[0036] It should be stated that all user data collected in this application is collected with the consent and authorization of the users. Moreover, the uses of the user data are all legal and compliant, and the use and processing of the user data comply with the relevant laws, regulations and standards of the relevant regions.

Claims

1. A nondestructive testing method for computer memory sticks, characterized in that: The steps include: Step 1: Obtain color images of the gold fingers on both sides of the memory stick to be tested, fit the two color images on the same color image, and then convert them into grayscale images; Analyze the grayscale image, filter out abnormal memory bars, and remove them. The remaining memory bars to be tested proceed to the next step. Step 2: using an oscilloscope to monitor a number of memory bars of the same type to be tested, obtaining clock signals of the processing process Q of the memory bars to be tested within a monitoring cycle, and generating a standard clock signal change curve diagram; Step 3: Use an oscilloscope to monitor the memory stick to be tested that is ready to leave the factory, obtain the clock signal of the memory stick processing process Q, generate a curve diagram of the clock signal change to be tested, analyze it in combination with the constructed standard capacitance characteristic change curve diagram, and calibrate the memory sticks that meet the standards and the memory sticks that do not meet the standards; Memory bars that do not meet the standards are removed, and memory bars that meet the standards are output.

2. A nondestructive testing method for computer memory sticks according to claim 1, characterized in that: The specific method of filtering abnormal memory bars and removing them in step 1 is as follows: Get the color images of the gold fingers on both sides of any memory stick to be tested, fit the two color images into one color image by connecting the gold fingers end to end, and convert them into grayscale images ; Extract grayscale image through edge detection algorithm The edge images of all gold fingers in ; by As the background, the pixel value is subtracted from Middle Cut , obtaining a grayscale image after removing the edge image of the gold finger, which only contains the grayscale pixels of the gold finger itself in the memory bar; Grayscale image Image with edge removed The blank pixels after the process are eliminated to make the adjacent gold finger bodies closely connected and obtain the final body grayscale image. ; To the grayscale image The pixel value is used for defect analysis to determine the grayscale image of the entity. Check whether the gold finger of the associated memory stick to be tested has defects. If so, the memory stick is regarded as an abnormal memory stick and is removed.

3. A nondestructive testing method for computer memory sticks according to claim 2, characterized in that: To the grayscale image The specific method of defect analysis based on the pixel value is as follows: Determine the grayscale image of the entity The number of rows M and columns N of pixels in the middle, traverse All the pixels in the image are extracted and the pixel values ​​are accumulated, and then the accumulated result is divided by The number of pixels in ,get The average gray value ; Combined with the average gray value Calculate the grayscale image of the entity The standard deviation of gray value ; according to Define the gray value range of the gold finger pixel ,in, is the minimum value in the interval, is the maximum value in the interval, and K is the preset adjustment coefficient; since The first pixel in Start traversing to the last pixel , get any pixel point Pixel value ,like The pixel point It is a normal pixel, otherwise it is an abnormal pixel; If confirmed If it is an abnormal pixel, it will be obtained immediately All adjacent pixels are determined, and it is determined whether there are abnormal pixels among the adjacent pixels. If there are abnormal pixels among the adjacent pixels, the abnormal pixels among the determined adjacent pixels are combined with Extract and extract the pixel points Different abnormal pixels except pixels The adjacent pixels other than the abnormal pixels are determined, and whether there are other abnormal pixels; and so on, until it is determined that there are no abnormal pixels in the adjacent pixels of all abnormal pixels; The extracted abnormal pixel points are fitted into abnormal regions, and the properties of the abnormal regions are determined by analyzing the area S, perimeter L and aspect ratio B of the determined abnormal regions.

4. A nondestructive testing method for computer memory sticks according to claim 3, characterized in that: The specific method for determining the nature of the abnormal area is: Get the area threshold preset by the operator , the area S of the abnormal region and the area threshold For comparison, if , then obtain the perimeter threshold preset by the operator , the perimeter L of the abnormal area and the perimeter threshold For comparison, if , then the aspect ratio B of the abnormal area is compared with the aspect ratio threshold preset by the operator For comparison, if , then the abnormal area is regarded as noise; If any of the conditions in the above comparison process are not met, the abnormal area will be regarded as a surface defect; Repeat the above steps to calculate the sum of the areas of all abnormal areas that are surface defects. , and determine In the grayscale image If the proportion exceeds , it is regarded as the grayscale image of the body If the associated gold finger has a defect, the memory module to be tested where the gold finger is located is regarded as an abnormal memory module. Preset values ​​for operators.

5. A nondestructive testing method for computer memory sticks according to claim 4, characterized in that: The monitoring period described in step 2 is determined as follows: Select several memory sticks of the same type to process the same process Q. For each memory stick, take the start time of the process as the initial time and the end time of the process as the end time. Record the time interval between the two times and take the average of all time intervals as a monitoring period.

6. A nondestructive testing method for computer memory sticks according to claim 5, characterized in that: The specific method of obtaining the clock signals of the processing processes Q of the memory bars to be tested within a monitoring cycle and generating the standard clock signal change curve diagram in step 2 is: Use an oscilloscope to monitor the clock signals of several memory modules of the same type within the monitoring period, and extract the fundamental wave amplitude, frequency, and period from the clock signals; Determine several different fundamental amplitudes, frequencies, and periods at different time points in the same monitoring cycle, determine the associated average fundamental amplitude, average frequency, and average period by taking the average value, and generate a standard clock signal change curve ; Based on the determined standard clock signal change curve , place it in a two-dimensional coordinate system with the time line as the horizontal axis and the fundamental wave amplitude as the vertical axis. The scale of the coordinate axis is set by the operator.

7. A nondestructive testing method for computer memory sticks according to claim 6, characterized in that: The standard clock signal variation curve diagram The changing pattern and period of the middle curve are consistent with the average frequency and average period.

8. A nondestructive testing method for computer memory sticks according to claim 7, characterized in that: Generate a curve chart of the clock signal change to be tested, and analyze it in combination with the constructed standard capacitance characteristic change curve chart. The specific method of calibrating the memory sticks that meet the standard and the memory sticks that do not meet the standard is as follows: According to the generated The method generates a clock signal change curve diagram of any memory module to be tested. ; The obtained Place with In the two-dimensional coordinate system, determine the standard clock signal change curve The time point corresponding to the first peak in ,and The time point corresponding to the first peak in ,use get and Time difference of the first peak ; According to the above method, calculate the monitoring period in chronological order. and The time difference of each group of corresponding peaks is The time difference is recorded as , sort in ascending order according to the time difference value, and get the time difference sequence ; With the timeline as the horizontal axis and the time difference value as the vertical axis, a two-dimensional coordinate system is constructed, and the time difference sequence is marked in the two-dimensional coordinate system to obtain Marked data points, starting from the first data point to the subsequent data points with short lines to obtain the polyline Z, and then draw a straight line parallel to the vertical axis passing through the first data point and a straight line parallel to the vertical axis passing through the last data point , calculate the horizontal axis, the broken line Z, and the straight line and straight lines The area of ​​the closed region formed is denoted by ; The area of ​​the enclosed region will be calculated The closed area threshold preset by the operator For comparison, if , the memory bar to be tested corresponding to the broken line Z is marked as a substandard memory bar, otherwise, the memory bar to be tested corresponding to the broken line Z is marked as a standard memory bar.

9. A nondestructive testing method for computer memory sticks according to claim 8, characterized in that: Memory sticks that do not meet the standards are removed again, not output, and the operator is notified; Output the memory sticks that meet the requirements.

Images