A non-destructive testing method for computer memory modules

Abstract

The present invention discloses a non-destructive detection method for computer memory modules. The present invention relates to the technical field of memory module detection. First, a pre-built camera is used to respectively capture color images on both sides of the gold fingers of the memory module to be tested, and the two captured color images are fitted onto the same color image and then converted into a grayscale image; the obtained grayscale image is analyzed to screen out the abnormal memory modules corresponding to the defective gold fingers, and a preliminary elimination operation is performed on the abnormal memory modules, and the remaining memory modules to be tested enter the next step; an oscilloscope is used to monitor the memory modules to be tested to obtain a clock signal curve graph, and in combination with the analysis of the standard capacitor characteristic curve graph, the qualified and unqualified memory modules are calibrated, the unqualified memory modules are eliminated and the qualified memory modules 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 existing technology 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 abnormal 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 have an impact on 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 existing technology, 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 of the existing technology before the memory modules leave the factory.

[0005] The object of the present invention can be achieved by the following technical solutions:

[0006] A non-destructive detection method for computer memory modules includes the following steps:

[0007] 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;

[0008] Analyze the grayscale image, screen out abnormal memory modules, and perform elimination operations. The remaining memory modules to be tested enter the next step;

[0009] Step 2: Use an oscilloscope to monitor several memory modules of the same type to be tested, obtain the clock signals of the processing processes of several memory modules to be tested within a monitoring period and generate a standard clock signal change curve graph;

[0010] Step 3: Use an oscilloscope to monitor the memory module to be tested before leaving the factory, and obtain the processing process of the memory module to be tested The clock signal is used to generate a change curve graph of the clock signal to be measured, and the graph is analyzed in combination with the constructed standard capacitance characteristic change curve graph to calibrate qualified memory modules and unqualified memory modules;

[0011] Perform an elimination operation on the unqualified memory modules and output the qualified memory modules.

[0012] As a further solution of the present invention, the specific method for screening and eliminating abnormal memory modules in step one is as follows:

[0013] Obtain the color images of the gold fingers on both sides of any memory module to be measured, 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 ;

[0014] Extract the edge images of all the gold fingers in the grayscale image by using an edge detection algorithm, and record it as ;

[0015] Taking as the background, subtract the pixel values to remove from to obtain a grayscale image after removing the edge image of the gold fingers, which only contains the grayscale pixels of the gold fingers themselves in the memory module;

[0016] Eliminate the blank pixel points in the grayscale image after removing the edge image to make the adjacent gold finger bodies closely connected, and obtain the final body grayscale image ;

[0017] 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 measured associated with the body grayscale image . If there are defects, regard the memory module as an abnormal memory module and perform an elimination operation.

[0018] 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:

[0019] Determine the number of rows and the number of columns 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 ;

[0020] Combine the average gray value Calculate the gray value standard deviation of the body gray image ; ;

[0021] According to Define the gray value interval of the finger pixel points , where is the minimum value in the interval, is the maximum value in the interval, is a preset adjustment coefficient;

[0022] From the first pixel point in start traversing to the last pixel point , obtain any one pixel point among them, and obtain its pixel value . If then the pixel point is regarded as a normal pixel point, otherwise it is an abnormal pixel point;

[0023] If it is determined that is an abnormal pixel point, then immediately obtain all adjacent pixel points, and judge 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 judge whether there are other abnormal pixel points; and so on, until it is determined that there are no abnormal pixel points in the adjacent pixel points of all abnormal pixel points and stop;

[0024] For the several abnormal pixel points extracted, fit them into an abnormal area, and by calculating the area , perimeter and aspect ratio of the determined abnormal area, determine the nature of the abnormal area.

[0025] As a further solution of the present invention, the specific method for determining the nature of the abnormal area is:

[0026] Obtain the area threshold preset by the operator, compare the area of the abnormal area with the area threshold . If , then obtain the perimeter threshold preset by the operator, and compare the perimeter of the abnormal area with the perimeter threshold Perform a comparison. If , then compare the aspect ratio of the abnormal area with the aspect ratio threshold preset by the operator . If , then regard the abnormal area as noise;

[0027] If any condition in the above comparison process is not satisfied, then regard the abnormal area as a surface defect;

[0028] Repeat the above steps, and sum up the areas of all abnormal areas regarded as surface defects , and determine the proportion in the body grayscale image . If the proportion exceeds , then it is regarded that there is a defect in the gold finger associated with the body grayscale image , and the memory module under test where the gold finger is located is regarded as an abnormal memory module. Among them, is the preset value of the operator.

[0029] As a further solution of the present invention, the determination method of the monitoring period described in step two is:

[0030] Select several memory modules of the same type and process the same process . 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.

[0031] As a further solution of the present invention, the specific method for obtaining the clock signals of several memory modules under test in one monitoring period to generate a standard clock signal change curve is:

[0032] Use an oscilloscope to monitor the clock signals of several memory modules of the same type in the monitoring period, and extract the fundamental wave amplitude, frequency, and period from the clock signals;

[0033] Determine several different fundamental wave amplitudes, frequencies, and periods at different time points in 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 ;

[0034] 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 scales of the coordinate axes are determined by the operator.

[0035] As a further solution of the present invention, the standard clock signal change curve The variation law and variation period of the middle curve are consistent with the average frequency and average period.

[0036] As a further solution of the present invention, a change curve graph of the clock signal to be measured is generated, and combined with the constructed change curve graph of the standard capacitor characteristics for analysis. The specific method for calibrating qualified memory modules and unqualified memory modules is as follows:

[0037] According to the generation method, generate the change curve graph of the clock signal to be measured for any memory module to be measured ;

[0038] Place the obtained in the two-dimensional coordinate system where is located, and determine the time point corresponding to the first peak in the standard clock signal change curve graph , and the time point corresponding to the first peak in , and use to obtain the time difference between the first peaks of and and ; ;

[0039] According to the above method, calculate the time difference between each group of corresponding peaks in and during the monitoring period in chronological order, and a total of time differences are obtained, which are successively recorded as . Arrange them in ascending order according to the time difference values to obtain the time difference sequence ;

[0040] Taking the time line as the horizontal axis and the values of the time differences as the vertical axis, construct a two-dimensional coordinate system, mark the time difference sequence in the two-dimensional coordinate system to obtain a marked data points. Starting from the first data point, connect them to the subsequent data points with short lines to obtain a broken line 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 area of the closed region formed by the horizontal axis, the broken line Z, the straight line ;

[0041] Compare the area of the calculated closed region with the area threshold of the closed region preset by the operator. If , then mark the memory module to be measured corresponding to the broken line Z as an unqualified memory module, otherwise, mark the memory module to be measured corresponding to the broken line Z as a qualified memory module.

[0042] As a further solution of the present invention, the memory sticks that do not meet the standards are again eliminated, and are not output, and the operator is notified;

[0043] Output the memory sticks that meet the requirements.

[0044] Beneficial effects of the present invention:

[0045] (1) The present invention forms a comprehensive memory stick quality inspection process by combining appearance inspection with clock signal monitoring. In appearance inspection, image processing technology is used to accurately extract the gold finger body part and identify abnormal areas, objectively and accurately detect minor appearance defects, and reduce manual inspection errors. In clock signal monitoring, a standard curve is generated based on a large amount of memory stick data of the same type to provide an accurate comparison benchmark to determine whether the memory stick clock signal is normal. This method improves the comprehensiveness and accuracy of the inspection, and also improves the inspection efficiency through an automated process, reducing manual operation time. In addition, strict quality control reduces after-sales risks caused by product problems, improves user trust in the product, enhances market competitiveness, and ensures that the memory sticks shipped meet high quality standards in appearance and performance, providing users with more reliable products.

[0046] (2) The present invention analyzes the clock signal of the memory stick to be tested, generates a clock signal change curve of the memory stick to be tested, and compares it with the standard curve, so as to construct the concept of closed area area to provide a quantitative and intuitive indicator for the performance evaluation of the memory stick; so that the operator can clearly understand the stability of the memory stick performance. The larger the area, the greater the degree of change of the time difference sequence, and the more unstable the memory stick performance. This intuitive judgment method facilitates the operator to make decisions quickly, thereby improving the quality control efficiency in the production process;

[0047] (3) The present invention performs defect detection on the gold finger part of the memory stick through an automated process. By calculating the average gray value and standard deviation of the gray image of the main body, the gray features of the gold finger can be quantitatively analyzed, and the gray value range of the pixel points can be further determined, so as to more accurately identify abnormal pixels. Then, the connected domain analysis method is used for the abnormal pixels to accurately extract the abnormal area, and the nature of the abnormal area is further determined by analyzing the area, perimeter and aspect ratio of the abnormal area. This multi-dimensional analysis method effectively avoids misjudgment and missed judgment, and ensures the authenticity and reliability of the detection results. It provides an efficient, reliable and economical technical means for the production and quality control of enterprises, and has strong practical value and promotion significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The present invention will be further described below in conjunction with the accompanying drawings.

[0049] Figure 1 is a schematic flow chart of the method of the present invention;

[0050] Figure 2 is a schematic flow chart of the method described in Embodiment 2 of the present invention;

[0051] Figure 3 is a schematic flow chart of the method described in Embodiment 3 of the present invention. Detailed implementation manners

[0052] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with 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 the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0053] Embodiment 1

[0054] A non-destructive detection method for a computer memory module, as Figure 1 shown, specifically includes the following:

[0055] After the memory module to be tested to be shipped enters the quality inspection area, first, appearance inspection needs to be carried out. 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.

[0056] 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 photographing the images.

[0057] When the cameras photograph the color images of the gold fingers of the memory module, the light source should be able to evenly irradiate 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 images are clearly visible; and when photographing each memory module to be tested, the installed position is appropriate and constant.

[0058] 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.

[0059] At this time, the grayscale image only contains the images of the gold fingers. However, 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 from the original grayscale image, a grayscale image containing only 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 ;

[0060] 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 ;

[0061] 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 to calibrate the normal grayscale pixel value range, abnormal pixel points are screened out from the main body grayscale image . Analyze the pixel points in the neighborhood of the abnormal pixel points to identify whether there are other abnormal pixel points in the neighborhood of the abnormal pixel points. Finally, extract all the obtained abnormal pixel points and connect all the abnormal pixel points to form an abnormal area. By further analyzing the abnormal area, determine the nature of the abnormal area, and judge whether the current memory module to be tested is an abnormal memory module. If so, remove the abnormal memory module. If not, proceed to the next step

[0062] For a number of memory modules to be tested of the same type that have passed the appearance inspection, use an oscilloscope to monitor the clock signal in real time during the process of processing the same section of the process within a monitoring period ;

[0063] The determination of the monitoring period is as follows. Take the time point when any memory module to be tested starts to process the process as the initial moment. When the process of the memory module to be tested ends, 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

[0064] 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 section of the process . Obtain a number of groups of clock signals, and respectively extract the fundamental wave amplitude, frequency, and period in the number of groups of clock signals

[0065] Again, by taking the average method, the associated average fundamental wave amplitude, average frequency, and average period are obtained, and a corresponding standard clock signal change curve graph is generated. ; 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 will not be elaborated here too much;

[0066] It needs to be determined here that the change rule and change period of the curve in the standard clock signal change curve graph are consistent with the average frequency and average period.

[0067] Then, taking the time line as the horizontal axis and the fundamental wave amplitude as the vertical axis, a two-dimensional coordinate system is constructed, and the determined standard clock signal change curve graph is placed in the constructed two-dimensional coordinate system to enter the next step.

[0068] Then, any memory module to be shipped out for testing is obtained, and the memory module to be tested is detected by an oscilloscope according to the above method, and the clock signal during the processing process of the memory module to be tested is obtained, and a clock signal change curve graph of the memory module to be tested is generated ; Then, the clock signal change curve graph of the memory module to be tested is fitted into the two-dimensional coordinate system where the standard clock signal change curve graph is located for analysis, and it is determined whether the current memory module to be tested meets the standard. If it is a qualified memory module, it is output. If the memory module to be tested is an unqualified memory module, it is not output. 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, the average value and variance of its pixel points are calculated, the normal grayscale pixel value range is calibrated, abnormal pixel points are screened and analyzed, and it is judged whether the memory module is abnormal.

[0069] In terms of clock signal monitoring, an oscilloscope is used to monitor the clock signal during the processing process of the memory module, the monitoring period is determined, parameters such as fundamental wave amplitude, frequency, and period are extracted, and a standard clock signal change curve graph is generated; finally, the clock signal change curve graph of the memory module to be tested is compared with the standard curve graph to judge whether the memory module to be tested meets the standard and decide whether to ship it out.

[0070] In terms of clock signal monitoring, an oscilloscope is used to monitor the clock signal during the processing process of the memory module, the monitoring period is determined, parameters such as fundamental wave amplitude, frequency, and period are extracted, and a standard clock signal change curve graph is generated; finally, the clock signal change curve graph of the memory module to be tested is compared with the standard curve graph to judge whether the memory module to be tested meets the standard and decide whether to ship it out.

[0071] Embodiment 2

[0072] Based on Embodiment 1, this embodiment further discloses a method for visually inspecting the gold fingers of a memory module, as follows: Figure 2 Specifically, it includes the following steps:

[0073] Based on the method described in Embodiment 1, a grayscale image of the gold fingers of the memory module to be tested can be obtained. Using an edge detection algorithm in the prior art, analyze the grayscale image and extract the edge images of all the gold fingers therein, denoted as ;

[0074] Then, regard the original grayscale image as the background image, and regard the edge image as the part to be removed from the original background image. By the method of subtracting pixel values , remove the edge image from the grayscale image to obtain a grayscale image that only contains the grayscale pixels of each individual gold finger in the memory module;

[0075] At this time, there are blank pixel points between every two adjacent gold fingers after the removal operation. Eliminate the blank pixel points and closely connect the adjacent gold fingers to obtain a final grayscale image in which all the gold fingers are closely connected, which is called the body grayscale image and denoted as ;

[0076] If there are no scratches or stains on the original gold fingers, the pixel values of the body grayscale image can be regarded as within a certain range. If there are scratches or other obvious stains on the original gold fingers, abnormal pixel points will appear in the body grayscale image . According to this principle, perform the following operations:

[0077] 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 being , use the formula:

[0078] ;

[0079] Calculate the average grayscale value of the body grayscale image , where represents the pixel value of the pixel point at the th row and th column in the body grayscale image , ... respectively represent the number of rows and columns of the body grayscale image , and both are positive integer counting indices, starting from 1, and ; ;

[0080] obtain the average grayscale value of the body grayscale image , and then according to the formula: ;

[0081] ;

[0082] calculate the standard deviation of the grayscale values of the body grayscale image , the larger the standard deviation, the greater the difference in the pixel grayscale values in the body grayscale image, which also means that there may be abnormal pixel points in the gold fingers associated with the body grayscale image; , in the body grayscale image ;

[0083] 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 the method of is: 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, is an adjustment coefficient, and its value is determined by the operator in combination with actual requirements;

[0084] 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.

[0085] For the pixel points in the body grayscale image , starting from the first pixel point until the last pixel point is obtained. For any pixel point among the pixel points, extract its pixel value and determine whether it is within the grayscale value range of the gold finger pixel points . If then it is determined that the pixel value of the current pixel point is within the grayscale value range of the gold finger pixel points , and this pixel point is a normal pixel point, otherwise it is determined that this pixel point is an abnormal pixel point is an abnormal pixel point;

[0086] 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 determine whether the pixel values of these 8 pixel points are within the gray value range of the FPC finger pixel points If the pixel value of any pixel point is not within the gray value range of the FPC finger pixel points , then determine whether there are abnormal pixel points among the pixel points adjacent to the pixel point . At this time, extract the abnormal pixel points that are within the neighborhood of the determined pixel point and different from the pixel point , together with the pixel point ;

[0087] Then, take 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 determine 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;

[0088] In the subsequent process of traversing from the first pixel point to the last pixel point , 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;

[0089] 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 , perimeter and aspect ratio (aspect ratio = maximum row span of the abnormal region / maximum column span of the abnormal region) of any one of the abnormal regions.

[0090] After determining the abnormal region, it cannot be directly determined that there are defects 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 by combining the following points:

[0091] Obtain the area threshold , perimeter threshold , aspect ratio threshold ;

[0092] Extract the area of any abnormal region , perimeter and aspect ratio , and compare the area of the abnormal region with the area threshold . If the area of the abnormal region is less than the area threshold , then continue to compare the perimeter of the abnormal region with the perimeter threshold . If the perimeter of the abnormal region is less than the perimeter threshold , then continue to compare the aspect ratio of the abnormal region with the aspect ratio threshold . If the aspect ratio 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;

[0093] For several abnormal regions, use the above method to make judgments 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 body grayscale image . If the proportion exceeds , then consider that there is a defect in the gold finger associated with the body grayscale image , 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 by the operator and is determined according to actual requirements;

[0094] In this embodiment, through edge detection and cropping operations on the grayscale image, a tightly connected grayscale image of the gold finger body is obtained. Then calculate the average grayscale value and standard deviation of this 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.

[0095] Embodiment 3

[0096] 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, it includes the following:

[0097] Based on the determined curve graph of the clock signal to be measured of the memory module to be tested , place the curve graph of the clock signal to be measured in the two-dimensional coordinate system where the standard curve graph of the clock signal is located. Due to the characteristics of the fundamental wave amplitude in the clock signal, it is determined that the curve graph of the fitted clock signal is a regular waveform graph similar to a sine function and a cosine function. Place the standard curve graph of the clock signal and the curve graph of the clock signal to be measured in the same two-dimensional coordinate system, and then obtain the time point on the horizontal axis corresponding to the first peak in the standard curve graph of the clock signal . 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 ;

[0098] ;

[0098] By adopting the calculation method, obtain the time difference between the first peak of the standard curve graph of the clock signal and the first peak of the curve graph of the clock signal to be measured ;

[0099] According to the above method, obtain the time point on the horizontal axis corresponding to the second peak in the standard curve graph of the clock signal 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 curve graph of the clock signal and the curve graph of the clock signal to be measured within the entire monitoring period. For the corresponding peaks of group a, a time differences are obtained in total. Denote them in the order of obtaining 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 . Then sort the obtained a time differences in ascending order to obtain the time difference sequence ;

[0100] 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 tested during the processing process Q and the standard clock signal, indicating that the working state of the current memory module to be tested during the processing process Q is more unstable. 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 tested 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 tested;

[0101] Using the sorting order of the time difference sequence (timeline order) as the horizontal axis and the value of the time difference as the vertical axis, construct a new two-dimensional coordinate system, and use each time difference in the time difference sequence to be marked on this two-dimensional coordinate system in the form of data points. A total of a marked data points are obtained. Starting from the first data point, connect it to the second data point with a short line, and then connect to the subsequent data points with short lines in sequence, finally passing through a marked data points to obtain a broken line representing the degree of fluctuation of the data points, denoted as Z;

[0102] Construct a straight line that passes through the first data point and is perpendicular to the horizontal axis and parallel to the vertical axis, denoted as ; then construct a straight line that passes through the last data point and is perpendicular to the horizontal axis and parallel to the vertical axis, denoted as ;

[0103] At this time, the broken line Z, the straight line , the straight line and the horizontal axis will form a closed area. Calculate the area of this closed area, denoted as ;

[0104] 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 test 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;

[0105] 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 mark the test memory module associated with the test clock signal change curve corresponding to this closed area as a non-compliant memory module. If the closed area is less than or equal to the closed area threshold , then mark the test memory module associated with the test clock signal change curve corresponding to this closed area as a compliant memory module;

[0106] Based on the determined non-compliant memory modules, eliminate them from the test memory modules, indicating that they have not passed quality inspection and are not output. Output the compliant memory modules, indicating that they have passed all quality inspections and can be shipped.

[0107] Embodiment 4

[0108] The technical solution of this embodiment lies in combining and implementing the solutions of the above-mentioned Embodiment 2 and Embodiment 3.

[0109] For some of the data in the formulas described above, the dimension is removed for numerical calculation. At the same time, the content not described in detail in this specification belongs to the prior art well-known to those skilled in the art.

[0110] The above content is only an example and explanation 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 for substitution. As long as they do not deviate from the invention or exceed the scope defined by this claim book, they shall fall within the protection scope of the present invention.

[0111] 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 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 non-destructive testing method for computer memory modules, characterized in that, It 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 onto the same color image, and then convert it into a grayscale image; Analyze the grayscale image, screen out the abnormal memory modules and perform the rejection operation, and the remaining memory modules to be tested enter the next step; Specific method for screening and removing abnormal memory modules: Obtain color images of the gold fingers on both sides of any memory module to be tested, fit the two color images into one color image in a 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 the edge detection algorithm, denoted as ; ; Taking as the background, the is removed by the pixel value subtraction method from to obtain 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 module. The grayscale image after removing the edge image is processed to eliminate the blank pixel points, 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, the memory module is regarded as an abnormal memory module and removed; Step 2: Use an oscilloscope to monitor several memory modules of the same type to be tested, and obtain the clock signals of the processing processes of several memory modules to be tested within a monitoring period, and generate a standard clock signal change curve graph; ​ The determination method of the monitoring period is as follows: Select several memory modules of the same type and process the same process , for each memory module, use 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 value of all time intervals as a monitoring period; Step 3: Use an oscilloscope to monitor the memory module to be tested before leaving the factory, and obtain the processing process of the memory module to be tested of the clock signal, generate a curve graph of the change of the clock signal to be tested, analyze it in combination with the constructed standard capacitance characteristic change curve graph, and calibrate the qualified memory modules and unqualified memory modules; Perform the rejection operation on the non-compliant memory modules and output the compliant memory modules.

2. The non-destructive testing method for a computer memory module according to claim 1, characterized in that, For The specific way to perform defect analysis is as follows: Determine the grayscale image of the body The number of rows of the pixel points And the number of columns , 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 , obtain The average grayscale value of ; Combined average gray value Calculate the gray value standard deviation of the body gray image ; ; According to Define the grayscale value range of the finger contact pad pixels , where is the minimum value within the range, is the maximum value within the range, is a preset adjustment coefficient; From the first pixel start traversing to the last pixel and obtain any one of the pixels value If then consider the pixel as a normal pixel, otherwise it is an abnormal pixel; If it is determined that is an abnormal pixel point, immediately obtain all adjacent pixel points, and determine whether there are any abnormal pixel points among the adjacent pixel points. If there are abnormal pixel points among the adjacent pixel points, then the abnormal pixel points among the determined adjacent pixel points, together with are extracted, and the adjacent pixel points of the abnormal pixel points different from the pixel point except for the pixel point are extracted, and it is determined whether there are other abnormal pixel points; and so on, until it is determined that there are no abnormal pixel points in the adjacent pixel points of all abnormal pixel points, then stop; For several extracted abnormal pixel points, they are fitted into an abnormal area, and by analyzing the area , perimeter and aspect ratio of the determined abnormal area, the nature of the abnormal area is determined.

3. A non-destructive testing method for a computer memory module according to claim 2, characterized in that, The specific method for determining the nature of the abnormal area is: Obtain the area threshold preset by the operator , and compare the area of this abnormal area with the area threshold . If , then obtain the perimeter threshold preset by the operator , and compare the perimeter of this abnormal area with the perimeter threshold . If , then compare the aspect ratio of the abnormal area with the aspect ratio threshold preset by the operator . If , then regard the abnormal area as noise; If any condition in the above comparison process is not met, the abnormal area is regarded as a surface defect; Repeat the above steps to calculate the sum of the areas of all abnormal regions that are surface defects , and determine the proportion in the body grayscale image . If the proportion exceeds , it is regarded that the gold finger associated with the body grayscale image has a defect. Then, the memory module under test where the gold finger is located is regarded as an abnormal memory module. Among them, is the preset value set by the operator 4. A non-destructive testing method for a computer memory module according to claim 1, characterized in that, In the second step, several processing processes of the memory modules to be tested within a monitoring period are obtained The specific method for generating the standard clock signal change curve of the clock signal 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 a number of 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 mean value to generate a standard clock signal variation curve graph ; Based on the determined standard clock signal variation curve graph , 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.

5. A non-destructive testing method for a computer memory module according to claim 4, characterized in that, The standard clock signal variation curve diagram The variation law and variation period of the curve in it are both consistent with the average frequency and average period.

6. The non-destructive testing method for a computer memory module according to claim 5, characterized in that, Generate a change curve graph of the clock signal to be tested, and analyze it in combination with the constructed standard capacitance characteristic change curve graph. The specific method for calibrating the compliant memory modules and non-compliant memory modules is: Generate according to a change curve graph of the clock signal to be measured for any memory module to be measured ; Place the obtained in the two-dimensional coordinate system where is located, and determine the time point corresponding to the first peak in the standard clock signal variation curve graph ; Then, for the time point corresponding to the first peak in and the time point corresponding to the first peak in , use to obtain the time difference between the first peaks ; According to the above method, calculate in chronological order within the monitoring period and the time differences of each corresponding wave peak in, and a total of a time differences are obtained, which are successively recorded as , and are arranged in ascending order according to the numerical values of the time differences to obtain the time difference sequence ; Taking the timeline as the horizontal axis and the numerical values of the time differences as the vertical axis, a two-dimensional coordinate system is constructed. The time difference sequence is marked in the two-dimensional coordinate system to obtain a 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 is drawn and a straight line parallel to the vertical axis and passing through the last data point . Calculate the area of the closed region formed by the horizontal axis, the broken line Z, the straight line and the straight line , and denote it as ; Calculate the area of the calculated closed region Compare it with the area threshold of the closed region preset by the operator If , then calibrate the memory module to be tested corresponding to the polyline Z as a non-compliant memory module, otherwise, calibrate the memory module to be tested corresponding to the polyline Z as a compliant memory module.

7. A non-destructive testing method for a computer memory module according to claim 6, characterized in that, Perform the rejection operation on the non-compliant memory modules again, do not output them, and notify the operator; Output the compliant memory modules.

Images