A high-precision anti-interference semiconductor chip comprehensive testing method and device

Abstract

The application relates to the technical field of semiconductor chip testing, in particular to a high-precision anti-interference semiconductor chip comprehensive testing method and device. The method comprises the following steps: obtaining image data of a chip surface structure, obtaining position information through preprocessing, edge detection, contour matching and secondary verification; executing probe movement, guide rail sliding and gauge adjustment to obtain a probe position; obtaining a voltage signal and an output current through a micro-control module, digital-analog conversion, voltage following, voltage sampling, current amplification and load feedback processing; performing circuit anti-interference, sampling comparison, integration and digital logic processing to obtain test data for load feedback verification, motion control review and test program determination to obtain a test result. The application realizes high-precision positioning through visual guidance and a multi-stage mechanical calibration chain, and significantly improves testing accuracy and system adaptability by combining an anti-interference circuit and closed-loop data processing.

Description

Technical Field

[0001] This invention relates to the field of semiconductor chip testing technology, and in particular to a high-precision, anti-interference semiconductor chip comprehensive testing method and apparatus. Background Technology

[0002] In the field of semiconductor chip testing technology, existing solutions typically acquire image data of the surface structure of a semiconductor chip through an image acquisition module, generate position information through an image analysis module, and then control probe movement through a motion control module. This is combined with a microcontroller (MCU), a digital-to-analog converter (DAC), a voltage follower module, a voltage sampling module, a current amplification module, a load feedback module, a circuit anti-interference module, a sampler, a first comparator, a second comparator, an integrator, a reference voltage source, a clock unit, and a digital logic module to complete the test program. However, this approach suffers from limitations such as unstable alignment between position information and probe position, fragmented test data and test results, and insufficient coordination between the load feedback module and the motion control module. Existing methods often rely on directly driving the probe to contact the chip after generating position information from image data, or directly entering the test program after the voltage signal and output current are generated, resulting in disconnections in the test chain.

[0003] In semiconductor chip integrated testing scenarios, issues such as probe position misalignment with the semiconductor chip's surface structure and disconnection between test data and the current probe position can easily arise, making it difficult to meet the requirement of stable test results. In existing solutions, image data, position information, probe position, voltage signal, output current, test data, and test results are usually scattered across different processing stages. The processing status of the previous stage is difficult to completely transfer to the next stage, causing the test program to require repeated calibration during continuous operation, which can easily affect the test cycle time in production and the connection between testing during processing.

[0004] Regarding the joint processing of test data, load feedback module, motion control module, and test program, existing technologies generally suffer from shortcomings such as difficulty in unifying and organizing feedback records and motion control records, difficulty in directly driving the motion control module to call preceding processes, and discontinuous utilization of sampling result fields and digital logic output result fields by the test program. These shortcomings make it difficult to form a consistent process of acquisition, alignment, judgment, control, and recording in the comprehensive testing scenario of semiconductor chips, resulting in a lack of continuous correspondence between the test result generation process and the preceding position information processing process, voltage signal and output current processing process. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a high-precision anti-interference semiconductor chip comprehensive testing method and apparatus, comprising:

[0006] S100: Acquire image data of the surface structure of the semiconductor chip, perform preprocessing, edge detection and secondary verification to obtain position information;

[0007] The image data is the raw data of the semiconductor chip surface structure, which includes pads, cathode combs, and the contour boundaries of the semiconductor chip surface.

[0008] S200. Based on the position information, perform probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment to obtain the probe position;

[0009] S300. Based on the probe position, current amplification and load feedback closed-loop readback processing are performed to obtain voltage signal and output current;

[0010] S400. Based on the voltage signal and output current, the circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit and digital logic module are used to process the data to obtain test data.

[0011] S500. Based on the test data, the load feedback module, motion control module, and test program are processed to obtain the test results.

[0012] Further, the process of preprocessing, edge detection, and secondary verification to obtain location information includes:

[0013] The preprocessing includes denoising, grayscale conversion, and binarization.

[0014] The Canny algorithm is used for gradient change identification.

[0015] The find Contours algorithm is used to process continuous edges to obtain contour information, and background contours and interfering contours are filtered out.

[0016] The contour extraction results are compared with the contour template one by one to obtain contour matching results that include contour position, contour direction and contour integrity.

[0017] First, establish a reference position for laser positioning based on the center coordinates, and then establish a reference position for optical positioning based on the contour matching results. Compare the laser positioning results, optical positioning results, and initial position information. When there is a positional deviation between the two, call the contour matching results to correct the center coordinates and perform a second verification to obtain the position information. The position information includes the verified center coordinates, the verified contour matching results, and the corresponding surface structure position record.

[0018] Furthermore, the probe movement process includes:

[0019] The probe is coarsely moved according to the center coordinates, and the current position value of the distance adjustment gauge is continuously read until the current position value is consistent with the position corresponding to the center coordinates.

[0020] Furthermore, the process of fixing the guide rail sliding includes:

[0021] Based on the probe movement, the fixed guide rail is controlled to slide along the through hole according to the contour direction in the contour matching result. When the contour matching result shows boundary offset or abnormal contour integrity, the position information of S100 is called to recheck the center coordinates and then the fixed guide rail sliding process is executed.

[0022] Furthermore, the process of adjusting the distance gauge includes:

[0023] After the fixed guide rail slides, read the current distance between the probe and the target position, and perform forward or backward adjustment according to the center coordinates and surface structure position. After the adjustment is completed, read the current position value again and check it against the boundary position corresponding to the contour matching result. If the check is inconsistent, return to the fixed guide rail sliding process and re-execute.

[0024] The probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment are performed sequentially in the order of coarse movement, direction correction, and fine spacing adjustment to obtain the probe position; the probe position includes the current coordinates of the probe, the current position of the fixed guide rail, and the current position of the distance adjustment gauge.

[0025] Further, the process of performing current amplification and load feedback closed-loop readback processing to obtain the voltage signal and output current includes:

[0026] The process of current amplification and load feedback closed-loop readback includes: processing by the microcontroller module (MCU), DAC digital-to-analog converter module, voltage follower module, voltage sampling module, current amplification module, and load feedback module;

[0027] The microcontroller module (MCU) and the digital-to-analog converter (DAC) are connected via a serial peripheral interface (SPI). The MCU first reads the current coordinates of the probe, the current position of the fixed guide rail, and the current position of the distance adjustment gauge, and then generates control signals to write to the register of the DAC.

[0028] The voltage follower module receives the initial voltage data output by the DAC digital-to-analog converter module, performs voltage follower processing, and compares the followed voltage data with the probe position. When the comparison matches, it outputs a voltage signal.

[0029] After receiving the voltage signal, the voltage sampling module performs voltage sampling processing, and calibrates the current reading with the current position of the distance adjustment gauge. When the calibration is consistent, the sampled voltage data is output.

[0030] After receiving the sampled voltage data, the current amplification module performs current amplification processing and compares the current output current with the current position of the fixed guide rail and the current position of the distance adjustment gauge. If the comparison is consistent, the current output current is retained.

[0031] After receiving the output current, the load feedback module performs load feedback processing, including feedback reading, open circuit detection and feedback transmission. When an open circuit is detected, the microcontroller module (MCU) stops writing the current control signal and recalls the probe position output by S200.

[0032] The voltage follower module, voltage sampling module, and current amplification module form a voltage series negative feedback method.

[0033] Furthermore, the process of processing the circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit, and digital logic module includes:

[0034] The circuit anti-interference module first filters high-frequency interference signals and performs common-mode rectification on the voltage signal and output current, and then connects to the fully differential wide common-mode input operational amplifier to form the current sampling input state with the help of the reference voltage source;

[0035] The sampler samples and reads the current sampling input state within the current clock cycle and outputs the sampling result.

[0036] The first comparator receives the comparison input state after the sampling result is processed by a single-ended output amplifier, and compares it with the reference voltage to obtain the output result of the first comparator;

[0037] The integrator reads the output of the first comparator within a continuous cycle to form an integral result;

[0038] The second comparator receives the integration result and compares it with the reference voltage to obtain the output result of the second comparator;

[0039] Under the control of the clock unit, the digital logic module receives the sampling result, the output result of the first comparator, the integration result, and the output result of the second comparator according to the current clock cycle, and forms the digital logic output result in chronological order.

[0040] The test data includes a sampling result field, a first comparator output result field, an integration result field, a second comparator output result field, and a digital logic output result field.

[0041] Furthermore, the processing steps of the load feedback module include:

[0042] The sampling result field and digital logic output result field are read from the test data, and the status comparison and open circuit detection are performed to generate a load feedback record. When the sampling result field and the digital logic output result field are inconsistent, the inconsistent status is written into the load feedback record and sent to the motion control module.

[0043] Furthermore, the processing steps of the motion control module include:

[0044] The system receives load feedback records, the first comparator output result field, and the second comparator output result field. It compares the first comparator output result field and the second comparator output result field according to the same clock unit beat. Based on the comparison result, it determines whether the current probe position is maintained at the current test position. When continuous deviations occur, it calls the probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment processing in S200 to generate motion control results.

[0045] Furthermore, the processing steps of the test program include:

[0046] It receives load feedback records, motion control results, integral result fields, and digital logic output result fields. First, it reads the integral result field to obtain the cumulative state within the continuous cycle, then reads the digital logic output result field to obtain the current cycle logic state, and then forms the current cycle test judgment data. After performing electrical test processing, it outputs electrical test records, and then forms a defect judgment record corresponding to the current test position.

[0047] The test results include load feedback results, motion control results, electrical test records, and defect determination records.

[0048] The key innovations of this invention include:

[0049] (1) Based on the test data, the test result is generated by a joint processing chain of load feedback module, motion control module and test program, wherein the load feedback module records the test data, the motion control module performs position verification based on the test data, and the test program outputs the test result in combination with the test data.

[0050] (2) Based on the position information, the probe position is generated by a continuous processing chain of probe movement, fixed guide rail sliding and distance adjustment gauge adjustment, wherein the position information first enters the motion control module and then acts on the probe, fixed guide rail and distance adjustment gauge in sequence to form the probe position corresponding to the surface structure of the semiconductor chip.

[0051] (3) To comprehensively test the semiconductor chip, a sequential processing chain is constructed for image data, position information, probe position, voltage signal, output current, test data and test results, wherein the result of the previous processing is used as the input of the next processing. The circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit and digital logic module are continuously connected with the load feedback module, the motion control module and the test program in the same process.

[0052] The following are its main beneficial effects:

[0053] (1) To address the problem of scattered test data and test results links, the test data is processed jointly by the load feedback module, the motion control module and the test program, so that the test results are formed in the same processing chain and the test results and the test data maintain a continuous correspondence.

[0054] (2) To address the problem of unstable connection between position information and probe position, a sequential transmission relationship between the position information and the probe position is formed through continuous processing of probe movement, fixed guide rail sliding and distance adjustment gauge adjustment, and the probe position corresponds to the surface structure of the semiconductor chip.

[0055] (3) To address the problem that image data, location information, probe position, voltage signal, output current, test data and test results are scattered in different processing stages, the processing objects are organized by sequential processing chain so that the previous processing result enters the next processing stage, forming a continuous process from acquisition to test result.

[0056] (4) To address the problem of insufficient coordination between the load feedback module and the motion control module in the existing solution, the load feedback module records the test data, and the motion control module reads the relevant processing results, so that the position verification and feedback record during the test process form a corresponding relationship.

[0057] (5) In view of the problem that existing methods often directly enter the test program after the voltage signal and output current are formed, the test data is generated by the circuit anti-interference module, the sampler, the first comparator, the second comparator, the integrator, the reference voltage source, the clock unit and the digital logic module, and then the test program processes the test data, so that the test program and the previous processing process are connected in a continuous manner. Attached Figure Description

[0058] Figure 1 A flowchart illustrating a high-precision anti-interference semiconductor chip comprehensive testing method provided in this application embodiment;

[0059] Figure 2 This is a structural block diagram of a high-precision anti-interference semiconductor chip integrated testing device provided in an embodiment of this application. Detailed Implementation

[0060] Example 1: Refer to Figure 1 This is a flowchart illustrating a high-precision anti-interference semiconductor chip comprehensive testing method provided by an embodiment of the present invention. The process may include at least steps S100-S500:

[0061] S100: Acquire image data of the surface structure of the semiconductor chip, perform preprocessing, edge detection and secondary verification to obtain position information;

[0062] S200. Based on the position information, perform probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment to obtain the probe position;

[0063] S300. Based on the probe position, current amplification and load feedback closed-loop readback processing are performed to obtain voltage signal and output current;

[0064] S400. Based on the voltage signal and output current, the circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit and digital logic module are used to process the data to obtain test data.

[0065] S500. Based on the test data, the load feedback module, motion control module, and test program are processed to obtain the test results.

[0066] S100: Acquire image data of the surface structure of the semiconductor chip, perform preprocessing, edge detection and secondary verification to obtain position information;

[0067] Specifically, the input source for this step is the semiconductor chip to be tested and the corresponding image acquisition module. The image acquisition module is preferably a camera. The camera captures images of the surface structure of the semiconductor chip, forming image data. This image data is raw data characterizing the surface structure of the semiconductor chip, which includes pads, solder pads, cathode combs, and the contour boundaries of the semiconductor chip surface. After the image data enters the image analysis module, it does not directly output the position information but first undergoes preprocessing. Preprocessing is a chain of processing steps to unify and organize the image data. Preprocessing includes denoising, grayscale conversion, and binarization. Denoising removes scattered interference and edge artifacts from the image data. Grayscale conversion organizes the image data into a single-channel level. Binarization separates the surface structure from the background. After preprocessing, the image data forms edge detection input data and is recorded as preprocessed image data in the image analysis module.

[0068] Further, the edge detection is a process of performing an edge detection algorithm on the preprocessed image data. The edge detection algorithm preferably uses the Canny algorithm. The Canny algorithm performs gradient change recognition on the preprocessed image data and outputs the edge information of the surface structure. This edge information is not the final output, but rather the input source for contour extraction. The contour extraction is a process of performing a contour extraction algorithm based on the edge information. The contour extraction algorithm preferably uses the find Contours algorithm. The find Contours algorithm organizes continuous edges into closed or semi-closed contour information. This contour information corresponds to the actual position and shape of the Pad, solder pad, cathode comb, and the semiconductor chip surface boundary. The image analysis module filters the contour information, retaining the contour information corresponding to the semiconductor chip surface structure, filtering out background contours and interfering contours, and recording the retained result as the contour extraction result.

[0069] Further, the contour matching is a process of determining the correspondence between the contour extraction results and the contour template. The contour template is template data representing the ideal contour of the Pad, solder pad, or cathode comb. The image analysis module compares the contour extraction results with the contour template one by one to obtain the contour matching results. The contour matching results include contour position, contour direction, and contour integrity. The image analysis module determines the center coordinates based on the contour matching results. The center coordinates are the positioning results of the Pad, solder pad, or cathode comb in the image coordinates. The center coordinates and the contour matching results together constitute the initial position information. When the aforementioned contour extraction results overlap, break, or shift, the image analysis module returns to the preprocessed image data and performs the edge detection and contour extraction again until the initial position information that can be used for verification is output, and records this processing process in the processing record corresponding to the image data.

[0070] Furthermore, the high-precision positioning module's secondary verification is a process of verifying laser positioning and optical positioning based on the initial position information. The high-precision positioning module is a system that performs secondary positioning determination based on the center coordinates and the contour matching result. Laser positioning is used to verify the position of the corresponding area of ​​the Pad, solder pad, or cathode comb. Optical positioning is used to verify the position of the contour boundary and the center coordinates. After receiving the initial position information, the high-precision positioning module first establishes a reference position for laser positioning according to the center coordinates, and then establishes a reference position for optical positioning according to the contour matching result; subsequently, it compares the laser positioning result, the optical positioning result, and the initial position information, outputting the secondary verification result. When the laser positioning result and the optical positioning result correspond to the same position area, the image analysis module writes this result into the position information; when there is a positional deviation, the image analysis module calls the contour matching result to correct the center coordinates and performs the high-precision positioning module's secondary verification process again. The position information obtained after the secondary verification process includes the verified center coordinates, the verified contour matching result, and the corresponding surface structure position record.

[0071] Understandably, the position information is the output of this step and also the direct input position for subsequent steps. After being output from the image analysis module, the position information is used in S200 for probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment processing. In other words, S200 does not directly read the image data, but rather reads the position information output from this step; wherein, the center coordinates are used for the initial positioning of the probe movement, and the contour matching result is used to determine the fixed guide rail sliding direction and the distance adjustment gauge adjustment position. Thus, this step completes the processing chain from the image data to the position information and achieves cross-step connection with the input position of S200.

[0072] The technical effects of this step can be summarized as follows. This step processes the image data into location information, which is not a single coordinate, but a positioning input composed of the center coordinates and the contour matching result. The high-precision positioning module's secondary verification connects the processing results of the image analysis module with laser positioning and optical positioning, forming a preliminary positioning processing chain. The location information directly enters S200, thus this step constitutes the preliminary foundation for subsequent probe movement and fixed guide rail sliding.

[0073] S200. Based on the position information, perform probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment to obtain the probe position;

[0074] Specifically, the input source for this step is the position information output by S100. The position information includes center coordinates, contour matching results, and surface structure position records. The center coordinates represent the position of the pad, solder pad, or cathode comb in the image data. The contour matching results represent the contour direction and contour integrity. The surface structure position records represent the correspondence between the semiconductor chip surface structure and the center coordinates. After the position information enters the motion control module, the motion control module first reads the center coordinates, then reads the contour matching results, and subsequently generates the probe movement path, the fixed guide rail sliding path, and the distance adjustment gauge adjustment path. This step does not directly output test data; instead, it first completes the position alignment between the probe and the semiconductor chip and records the aligned result as the probe position.

[0075] Furthermore, the probe is mounted on a pin holder. The pin holder includes a PCB (Printed Circuit Board) and a ceramic substrate. The PCB supports the connection structure of the probe. The ceramic substrate supports the mounting structure of the probe. The through-hole is provided on both the PCB and the ceramic substrate. The fixed guide rail passes through the through-hole. The distance adjustment gauge is fixed to the fixed guide rail. The fixed guide rail is a mechanical structure that guides the probe to slide along a predetermined direction. The distance adjustment gauge is a gauge structure that characterizes the current sliding position of the fixed guide rail. After receiving the position information, the motion control module first determines the deviation between the center coordinates and the current position of the probe, then determines the directional deviation corresponding to the contour matching result, and writes both types of deviations into the current adjustment record.

[0076] Furthermore, the probe movement is an alignment action performed based on the center coordinates. The motion control module maps the center coordinates to the movement direction of the probe card and drives the probe to move towards the area corresponding to the center coordinates. During the probe movement, the motion control module continuously reads the current position value of the distance adjustment gauge. When the current position value matches the position corresponding to the center coordinates, the probe stops moving. When the current position value does not match the position corresponding to the center coordinates, the motion control module continues to issue sliding commands. Understandably, the probe movement is not a single action, but a continuous process including coarse and fine movements. The coarse movement corresponds to the main position of the center coordinates. The fine movement corresponds to the directional deviation and boundary deviation in the contour matching result. Through this process, the probe gradually approaches the target position of the semiconductor chip surface structure.

[0077] Furthermore, the sliding of the fixed guide rail is a direction correction process performed based on the movement of the probe. After the fixed guide rail passes through the through hole, the sliding direction of the fixed guide rail is consistent with the extension direction of the through hole. The motion control module controls the fixed guide rail to slide along the through hole according to the contour direction in the contour matching result. During the sliding of the fixed guide rail, the distance adjustment gauge synchronously outputs the sliding distance. The motion control module compares the sliding distance with the surface structure position record. When the comparison matches, the current fixed guide rail position is retained. When the comparison does not match, the sliding of the fixed guide rail continues. When the aforementioned contour matching result shows boundary offset or abnormal contour integrity, the motion control module calls the position information in S100 to recheck the center coordinates, then performs the fixed guide rail sliding process again, and writes this checking process into the processing record.

[0078] Furthermore, the distance adjustment gauge adjustment is a spacing correction process performed after the fixed guide rail has slid. The distance adjustment gauge is fixed on the fixed guide rail, and its output is the current distance between the probe and the target position. After reading the current distance, the motion control module adjusts the distance adjustment gauge according to the center coordinates and the surface structure position record. The distance adjustment gauge adjustment includes forward adjustment and backward adjustment. When the current distance is greater than a preset position range, forward adjustment is performed. When the current distance is less than the preset position range, backward adjustment is performed. After the forward adjustment and backward adjustment are completed, the motion control module reads the current position value of the distance adjustment gauge again and checks it against the boundary position corresponding to the contour matching result. If the check matches, the probe position is output. If the check does not match, the fixed guide rail sliding process is returned to be re-executed. Thus, the probe movement, the fixed guide rail sliding, and the distance adjustment gauge adjustment form a continuous processing chain.

[0079] As an operational engineering embodiment, in a wafer testing scenario, the camera first acquires image data of the semiconductor chip surface structure. After S100 outputs the position information, the motion control module reads the center coordinates and controls the probe mounted on the probe holder to move towards the Pad area. When the probe approaches the Pad area, the fixed guide rail slides along the through hole, and the distance adjustment gauge simultaneously outputs the current position value. Subsequently, the motion control module corrects the position of the fixed guide rail based on the contour matching result and performs forward or backward adjustment based on the current position value of the distance adjustment gauge. After the adjustment is completed, the current position of the probe relative to the Pad area is recorded as the probe position. This probe position includes three items: the current coordinates of the probe, the current position of the fixed guide rail, and the current position of the distance adjustment gauge, and is written into the subsequent call record by the motion control module.

[0080] Understandably, the probe position is the output of this step and also the direct input position of S300. S300 does not directly call the position information, but instead calls the probe position output by this step. The current probe coordinates in the probe position are used by the microcontroller module (MCU) to call the DAC (digital-to-analog converter) module for timing; the current position of the fixed guide rail is used as the processing starting point for the corresponding positions of the voltage follower module and the voltage sampling module; and the current position of the distance adjustment gauge is used for recording the input position of the load feedback module. Thus, this step completes the conversion from the position information to the probe position and achieves cross-step connection with S300.

[0081] The technical effects of this step can be summarized as follows. This step converts the position information into the probe position. The probe position is not a single position value, but rather a result that includes the current coordinates of the probe, the current position of the fixed guide rail, and the current position of the distance adjustment gauge. The probe movement, the sliding of the fixed guide rail, and the adjustment of the distance adjustment gauge are executed continuously within the same processing chain, thus providing the probe position with a direct input relationship to S300. The motion control module completes position reading, path generation, deviation correction, and output recording within this step, connecting the preceding image data processing chain with the subsequent voltage signal processing chain into a complete process.

[0082] S300. Based on the probe position, current amplification and load feedback closed-loop readback processing are performed to obtain voltage signal and output current;

[0083] Specifically, the input source for this step is the probe position output by S200. The probe position includes the current probe coordinates, the current position of the fixed guide rail, and the current position of the distance adjustment gauge. The current probe coordinates represent the current positioning state of the probe relative to the surface structure of the semiconductor chip. The current position of the fixed guide rail represents the current state of the mechanical sliding structure within the probe holder. The current position of the distance adjustment gauge represents the current distance between the probe and the pad or solder pad. After the probe position is entered into the microcontroller unit (MCU), the MCU first reads the current probe coordinates, then reads the current position of the fixed guide rail and the current position of the distance adjustment gauge, and generates a control signal based on the correspondence between the three. After the control signal enters the digital-to-analog converter (DAC), the DAC converts the control signal into initial voltage data and writes the initial voltage data into the current processing record.

[0084] Furthermore, the microcontroller module (MCU) and the DAC (digital-to-analog converter) are connected via a Serial Peripheral Interface (SPI). The SPI receives control signals from the MCU and sends these signals to the registers of the DAC. Upon receiving the control signals, the DAC outputs the initial voltage data according to the current state corresponding to the probe position. When there is a deviation between the current coordinates of the probe and the current position of the fixed guide rail, the MCU adjusts the writing order of the control signals and re-drives the DAC to output new initial voltage data. When the current position of the distance adjustment gauge is inconsistent with the preset distance state, the MCU pauses subsequent processing, rereads the probe position, and executes the control signal writing again. After this processing, the DAC outputs stable initial voltage data and sends this data to the voltage follower module.

[0085] Further, after receiving the initial voltage data, the voltage follower module performs voltage follower processing. The voltage follower module is responsible for maintaining and transmitting the initial voltage data. It includes an operational amplifier and corresponding connection structures. The voltage follower processing includes input reading, follower output, and current state verification. Input reading corresponds to the initial voltage data entering the voltage follower module. Follower output corresponds to the voltage follower module outputting the followed voltage data. Current state verification corresponds to the voltage follower module comparing the followed voltage data with the probe position. If the comparison matches, the voltage signal is output. If the comparison does not match, the voltage follower module returns the abnormal state to the microcontroller module (MCU) and triggers a re-read of the probe position. After this processing, the voltage follower module outputs the voltage signal. The voltage signal is voltage data characterizing the test drive state at the current probe position and is also the common input for the subsequent voltage sampling module and current amplification module.

[0086] Further, after receiving the voltage signal, the voltage sampling module performs voltage sampling processing. The voltage sampling module includes a sampling resistor and an operational amplifier. The sampling resistor corresponds to the voltage reading path. The operational amplifier corresponds to the sampling amplification path. The voltage sampling processing includes sampling reading, sampling calibration, and sampling output. Sampling reading corresponds to the current reading of the voltage signal on the sampling resistor. Sampling calibration corresponds to the voltage sampling module calibrating the current reading against the distance status in the probe position. Sampling output corresponds to the voltage sampling module outputting the sampled voltage data. When the current reading is inconsistent with the current position of the distance adjustment gauge, the voltage sampling module sends this inconsistency back to the microcontroller module (MCU), and the MCU regenerates the control signal. After this processing, the voltage sampling module outputs the sampled voltage data and sends it to the current amplification module.

[0087] Further, after receiving the sampled voltage data, the current amplification module performs current amplification processing. The current amplification module converts the sampled voltage data into a current output state. The current amplification processing includes voltage reading, current conversion, amplified output, and state verification. Voltage reading corresponds to the sampled voltage data entering the current amplification module. Current conversion corresponds to the current amplification module reading the corresponding current drive state according to the current probe position. Amplified output corresponds to the current amplification module outputting the current output current. State verification corresponds to the current amplification module comparing the current output current with the current position of the fixed guide rail and the current position of the distance adjustment gauge. If the comparison matches, the current output current is retained. If the comparison does not match, the current amplification module sends the current state back to the voltage sampling module, which then returns it to the microcontroller module (MCU). Thus, the voltage following module, the voltage sampling module, and the current amplification module form a voltage series negative feedback mechanism. This voltage series negative feedback mechanism is the core processing chain in this step. Both the voltage signal and the output current are generated within this processing chain.

[0088] Further, after receiving the output current, the load feedback module performs load feedback processing. The load feedback module is a module that reads back the current state of the output current. The load feedback processing includes feedback reading, open circuit detection, and feedback feedback transmission. Feedback reading corresponds to the load feedback module reading the current state of the output current. Open circuit detection corresponds to the load feedback module determining whether an open circuit has occurred in the current output path. Feedback feedback transmission corresponds to the load feedback module sending the feedback result back to the microcontroller module (MCU). When the aforementioned load feedback module detects an open circuit, the MCU stops writing the current control signal and re-calls the probe position output by S200. When the aforementioned load feedback module does not detect an open circuit, it retains the current voltage signal and the output current. Thus, the load feedback module forms a closed-loop readback path in this step and writes the closed-loop readback path into the current processing record.

[0089] As a feasible engineering implementation, in a wafer testing scenario, after the probe position is determined, the microcontroller module (MCU) first reads the current coordinates of the probe, and then writes a control signal to the DAC (digital-to-analog converter) module via the serial peripheral interface (SPI). The DAC module outputs initial voltage data, the voltage follower module outputs the voltage signal, the voltage sampling module reads the current state of the voltage signal, the current amplification module outputs the output current according to the sampled voltage data, and the load feedback module performs feedback reading and open-circuit detection on the output current. When the open-circuit detection result is normal, the current voltage signal and output current are retained; when the open-circuit detection result is abnormal, the MCU rereads the probe position and rewrites the control signal. Through this engineering implementation process, those skilled in the art can directly complete the process from the probe position to the voltage signal and the output current.

[0090] Understandably, the voltage signal and the output current are the output products of this step and also the direct input locations of S400. S400 receives not the probe position, but the voltage signal and the output current output from this step. The voltage signal enters the circuit's anti-interference module and the sampler. The output current enters the sampler and the subsequent decision chain. Thus, this step completes the processing chain from the probe position to the voltage signal and the output current, and achieves cross-step connection with S400.

[0091] The technical effects of this step can be summarized as follows: This step converts the probe position into a voltage signal and an output current, which are formed within the same processing chain. The microcontroller module (MCU), the DAC (digital-to-analog converter), the voltage follower module, the voltage sampling module, the current amplification module, and the load feedback module operate continuously, forming the current test drive chain. The load feedback module returns the open-circuit detection result to the preceding module, thus this step forms a seamless connection with S200 and S400.

[0092] S400. Based on the voltage signal and output current, the circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit and digital logic module are used to process the data to obtain test data.

[0093] Specifically, the input sources for this step are the voltage signal and the output current output from S300. The voltage signal is the current drive signal continuously processed by the microcontroller module (MCU), the DAC (digital-to-analog converter), the voltage follower module, and the voltage sampling module. The output current is the current current state generated by the current amplification module and the load feedback module corresponding to the current probe position. After the voltage signal and the output current enter S400, they do not directly form the test data, but first enter the circuit anti-interference module. The circuit anti-interference module is a processing unit that performs high-frequency interference signal filtering, common-mode rectification, and input state calibration on the voltage signal and the output current. After receiving the voltage signal and the output current, the circuit anti-interference module first filters the voltage signal, then filters the output current, and writes the filtered current state into the current processing record. When the aforementioned voltage signal and the output current correspond consistently under the same clock unit cycle, the current processing result is output; when the aforementioned correspondence deviates, the clock unit issues the next cycle, and the circuit anti-interference module rereads the voltage signal and the output current.

[0094] Furthermore, the front-end input of the circuit anti-interference module is connected to a fully differential wide common-mode input operational amplifier. The fully differential wide common-mode input operational amplifier is an operational amplifier structure that differentially adjusts the current input state. The reference voltage source is connected to the fully differential wide common-mode input operational amplifier, providing a reference voltage to it. After reading the filtered voltage signal and the output current, the fully differential wide common-mode input operational amplifier performs differential amplification and common-mode calibration to form the current sampling input state. The current sampling input state enters the sampler. The sampler is a processing unit that performs timing reading and state retention of the current sampling input state. The sampler completes one sampling reading within the current clock cycle and outputs the sampling result. When the aforementioned sampling result corresponds to the reference voltage output by the reference voltage source, the current sampling result is retained; when the aforementioned sampling result deviates from the reference voltage, the sampler re-executes the sampling reading in the next clock cycle and writes the resampling process into the sampling record.

[0095] Furthermore, the output terminal of the sampler is connected to the non-inverting input terminal of a single-ended output amplifier. The single-ended output amplifier is an amplification unit that organizes the sampling results into a single-ended output state. After receiving the sampling results, the single-ended output amplifier outputs the current comparison input state. The first comparator receives the current comparison input state and the reference voltage output by the reference voltage source. The first comparator performs a comparison process on the two to obtain a first comparator output result. The first comparator output result represents the difference between the current input state and the reference voltage. The first comparator output result enters the integrator. The integrator is a processing unit that continuously accumulates the first comparator output result. The integrator reads the first comparator output result within a continuous cycle, forms an integration result, and writes the integration result into the integration record. When the aforementioned first comparator output result continuously changes in the same direction, the integrator continues to accumulate; when the aforementioned first comparator output result changes in the opposite direction, the integrator restarts the current round of integration processing.

[0096] Further, the second comparator receives the integration result and the reference voltage output by the reference voltage source. The second comparator performs a second comparison process on the integration result and the reference voltage to obtain the second comparator output result. The second comparator output result is not the final output, but rather the input source for the digital logic module. The digital logic module is a processing unit that performs state organization, clock correlation, and logic combination on the sampling result, the first comparator output result, the integration result, and the second comparator output result. Under the control of the clock unit, the digital logic module receives the sampling result, the first comparator output result, the integration result, and the second comparator output result according to the current clock cycle, and forms a digital logic output result in chronological order. The digital logic output result, together with the sampling result, the first comparator output result, the integration result, and the second comparator output result, constitutes the test data. The test data is recorded in this step as a sampling result field, a first comparator output result field, an integration result field, a second comparator output result field, and a digital logic output result field.

[0097] As an engineering embodiment, in a wafer testing scenario, after the S300 outputs the voltage signal and the output current, the circuit anti-interference module first performs high-frequency interference signal filtering on the voltage signal and the output current. Then, the fully differential wide common-mode input operational amplifier reads the filtered current state and forms the current sampling input state in conjunction with the reference voltage source. Subsequently, the sampler performs a sampling read under the control of the clock unit, the single-ended output amplifier outputs the current comparison input state, the first comparator generates the first comparator output result, the integrator generates the integration result within consecutive cycles, the second comparator generates the second comparator output result, and the digital logic module writes the above results into the test data in cycle order. When the aforementioned second comparator output result changes abnormally within two consecutive cycles, the digital logic module retains the current record and calls the clock unit to continue the next round of sampling processing. After this processing, the test data forms a complete current round test record.

[0098] Understandably, the test data is the output of this step and also the direct input to S500. S500 does not call the voltage signal and the output current, but rather the test data output in this step. The sampling result field is used by the load feedback module, the first comparator output result field and the second comparator output result field are used by the motion control module, and the integration result field and the digital logic output result field are used by the test program. Thus, this step completes the processing chain from the voltage signal and the output current to the test data and achieves cross-step connection with S500.

[0099] The technical effects of this step can be summarized as follows: This step converts the voltage signal and the output current into test data. The test data is not a single reading, but rather a current round record containing a sampling result field, a first comparator output result field, an integration result field, a second comparator output result field, and a digital logic output result field. The circuit anti-interference module, the sampler, the first comparator, the second comparator, the integrator, the reference voltage source, the clock unit, and the digital logic module operate continuously within the same processing chain. The test data directly enters S500, thus connecting the preceding drive chain and the subsequent test program into a complete process.

[0100] S500. Based on the test data, the load feedback module, motion control module, and test program are processed to obtain the test results.

[0101] Specifically, the input source for this step is the test data output by S400. The test data includes a sampling result field, a first comparator output result field, an integration result field, a second comparator output result field, and a digital logic output result field. The sampling result field represents the current state of the voltage signal and the output current after processing by the sampler. The first comparator output result field represents the current comparison result. The integration result field represents the cumulative state within a continuous cycle. The second comparator output result field represents the current judgment state after the second comparison. The digital logic output result field represents the logical processing state of the digital logic module for the aforementioned fields. After the test data enters this step, the test result is not directly output; instead, it is first read and processed sequentially by the load feedback module, the motion control module, and the test program. After processing, the processing result is recorded as the test result and written into the current round of test records.

[0102] Further, the load feedback module first receives the sampling result field and the digital logic output result field. In this step, the load feedback module is used to read back and verify the output current state formed in the previous main step. The processing content of the load feedback module includes field reading, state comparison, open circuit detection, and feedback recording. Field reading involves reading the sampling result field and the digital logic output result field from the test data. State comparison involves determining the correspondence between the current state corresponding to the sampling result field and the current state corresponding to the digital logic output result field. Open circuit detection involves determining the open circuit state in the current state. Feedback recording involves writing the above determination results into the load feedback record. When the aforementioned sampling result field and the digital logic output result field are consistent, the load feedback module outputs the current feedback result; when the two are inconsistent, the load feedback module writes the inconsistent state into the load feedback record and sends the record to the motion control module. Thus, the load feedback module completes the first round of processing of the test data in this step and generates feedback input for subsequent action calls.

[0103] Further, the motion control module receives the load feedback record, the first comparator output result field, and the second comparator output result field. In this step, the motion control module verifies the position status during the current test process. The processing content of the motion control module includes field comparison, position determination, action invocation, and record write-back. Field comparison involves comparing the first comparator output result field and the second comparator output result field according to the same clock unit cycle. Position determination determines whether the current probe position remains at the current test position based on the comparison result. Action invocation invokes the probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment processes in S200 when the first comparator output result field and the second comparator output result field show continuous deviations. Record write-back writes the current determination content into the motion control record. Understandably, the motion control module does not regenerate the probe position in this step, but instead verifies the current probe position status based on the test data. If the verification passes, the current test program is maintained; if the verification fails, an action invocation signal is output, and this signal is recorded as the motion control result.

[0104] Further, the test program receives the load feedback record, the motion control result, the integral result field, and the digital logic output result field. In this step, the test program performs electrical tests on the semiconductor chip and detects whether the semiconductor chip has defects based on the test results. Specifically, the test program first reads the integral result field to obtain the cumulative state within consecutive cycles; then it reads the digital logic output result field to obtain the current round logic state; subsequently, it organizes the cumulative state, the current round logic state, the load feedback record, and the motion control result to form the current round test judgment data. After the current round test judgment data enters the electrical test processing, it outputs an electrical test record. The electrical test record is correlated with the current test position of the semiconductor chip to form a defect judgment record. When the aforementioned defect judgment record indicates that a defect exists at the current test position, the test program outputs a defect status; when the aforementioned defect judgment record indicates that no defect exists at the current test position, the test program outputs a normal status. The defect status and the normal status together constitute the final test judgment content.

[0105] As an engineering embodiment, in a wafer testing scenario, after S400 outputs the test data, the load feedback module first reads the sampling result field and the digital logic output result field, and completes the current round open circuit detection; subsequently, the motion control module reads the first comparator output result field and the second comparator output result field to perform a status check on the current probe position; if the check fails, the motion control module issues an action call signal, invoking the probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment processing in S200; if the check passes, the test program reads the integration result field and the digital logic output result field, generates an electrical test record, and performs defect detection on the semiconductor chip based on the electrical test record. Through this implementation process, those skilled in the art can directly complete the process from the test data to the test result.

[0106] Furthermore, the test results are the output of this step. The test results include load feedback results, motion control results, electrical test records, and defect determination records. The load feedback results characterize the current round of feedback status. The motion control results characterize whether the current probe position calls S200. The electrical test records characterize the current round of electrical test status. The defect determination records characterize whether the semiconductor chip has defects. Understandably, the test results are written into the test result record after the current round of testing and are used in the next round of image data acquisition, probe movement, and microcontroller (MCU) processing. Thus, this step completes the processing chain from the test data to the test results and establishes a closed-loop connection between the preceding main step and the next round of testing.

[0107] The technical effects of this step can be summarized as follows: This step transforms the test data into test results, which consist of load feedback results, motion control results, electrical test records, and defect judgment records. The load feedback module, the motion control module, and the test program process the same set of test data sequentially, and the preceding comparisons, sampling, and logical states form a unified judgment within this step. After the test results are written into the current round of test records, the current round of testing and the next round of testing form a closed loop relationship.

[0108] In one specific embodiment:

[0109] In S500, based on the test data, the load feedback module, motion control module, and test program are processed to obtain the test results.

[0110] The input source for this step is the test data output by S400. The test data includes a sampling result field. The first comparator output field Points result field The output field of the second comparator and digital logic output result fields The sampling result field Characterizes the current amplitude state of the voltage signal and output current after processing by the sampler, with the dimension of volts (V); the first comparator output result field The current comparison result is represented by a dimensionless Boolean value; the integral result field The cumulative state within a continuous beat is represented by the unit volt-second (V·s); the output field of the second comparator The current determination state after the second comparison is represented by a dimensionless Boolean value; the digital logic output result field The logical arrangement state of the aforementioned fields by the digital logic module is represented by dimensionless enumeration values. After the test data enters this step, the load feedback module first reads the sampling result fields. and the digital logic output result field Consistency verification is then performed. In this step, the load feedback module is used to read back and verify the output current state generated in the previous main step. Its processing includes field reading, state comparison, circuit breaker detection, and feedback recording. Field reading extracts data from the test data. and State comparison refers to... The current amplitude and the digital logic output result field The consistency of the represented logical expectation is determined. To quantify this consistency, a load feedback consistency metric is defined. As given by formula ①:

[0111] Formula①

[0112]

[0113] in, Load feedback consistency metrics;

[0114] : Sampling result field, derived from the sampler output of S400;

[0115] The preset current-to-voltage conversion threshold is derived from the configuration register of the load feedback module and has a value of 0.5V.

[0116] : The modified linear unit function is defined as , Let be any real number;

[0117] Indicator function, takes the value 1 when the condition in parentheses is true, otherwise takes the value 0;

[0118] The digital logic output result field originates from the S400 digital logic module and has a value range. ;

[0119] : The preset normal state enumeration value, which takes the value "normal".

[0120] Simple numerical example: Let , ,but , ; Therefore ,but This indicates that consistency has been achieved and there is a margin of safety. After this section of processing is completed, the load feedback module generates the load feedback record field. This field contains consistency metrics. And the open circuit detection status, which can be called by the motion control module.

[0121] Receive the aforementioned load feedback records and the output field of the first comparator Second comparator output field The motion control module then begins to verify the position status during the current test. The processing by the motion control module includes field comparison, position determination, action invocation, and record write-back. Field comparison refers to... and Alignment is performed according to the same clock unit beat. An offset accumulation function is defined. This is used to quantify the deviation of comparison results within consecutive beats. It takes the comparison result sequence of the most recent N beats, where the output sequence of the first comparator is denoted as... The output sequence of the second comparator is denoted as , For beat indexing. The cross-correlation deviation metric is used, given by formula ②:

[0122] Formula②

[0123]

[0124] in, Deviation rate, range of values ;

[0125] The length of the sliding window is derived from the configuration parameters of the motion control module and has a value of 5.

[0126] : Beat index, positive integer, value ;

[0127] : No. The first comparator output field for each beat comes from the first comparator of S400 and is a boolean value, taking the value 0 or 1.

[0128] : No. The output field of the second comparator for each beat comes from the second comparator of S400 and is a boolean value, taking the value 0 or 1.

[0129] The XOR operation returns 1 when two Boolean values ​​are different and 0 when they are the same.

[0130] Formula ② addresses the practical problem of quantifying the inconsistency of the comparator results through XOR averaging within a sliding window, thereby determining whether the probe contact is stable. If Exceeding the preset threshold If the deviation is 0.6 (e.g., 0.6), it is considered a continuous deviation. At this point, the motion control module generates an action call signal, invoking the probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment processes in S200. Position determination further utilizes load feedback recording. In and To perform the fusion, the callback triggering condition is defined by formula ③:

[0131] Formula③

[0132]

[0133] in, : Callback trigger flag, a boolean value, 1 indicates that S200 needs to be called back, 0 indicates that it does not;

[0134] Deviation rate, obtained from formula ②;

[0135] The deviation rate threshold is derived from the configuration parameters of the motion control module and has a value of 0.6.

[0136] Indicator function, defined as in formula ①;

[0137] The load feedback consistency index is obtained from formula ①.

[0138] : Logical OR operation.

[0139] Simple numerical example: Let , ,but ;set up ,but ; This triggers a callback. The motion control module then writes the current judgment information into the motion control record field. The record contains , And a flag indicating whether to execute the callback. When the review passes ( The current test program is maintained; if the review fails, an action call signal is output and recorded as the motion control result. After this section is completed, the motion control result field... Provided for use by the test program.

[0140] Furthermore, the test program receives the load feedback record. The motion control results The integration result field and the digital logic output result field The test procedure in this step is used to perform electrical tests on the semiconductor chip and detect the presence of defects based on the test results. Specifically, the test procedure first reads the integration result field. Obtain the cumulative state within the continuous cycle, and then read the digital logic output result field. Obtain the current round's logic state. Define a comprehensive defect risk score to generate the current round's test judgment data. This scoring system integrates the cumulative energy of the integral, the logical state, and the results of load and motion verification. First, the integral results are normalized, defining the normalized integral energy. Given by formula ④:

[0141] Formula④

[0142]

[0143] in, Normalized integral energy, range of values ;

[0144] The integration result field, derived from the S400 integrator, has a range of values. ;

[0145] The preset maximum integral value is derived from the test program configuration and is set to 10 V·s.

[0146] Load feedback records and motion control results It also participates in the integration. Defining defect risk scoring. Given by formula ⑤:

[0147] Formula⑤

[0148]

[0149] in, Defect risk score, value range ;

[0150] The preset weighting coefficients are derived from the parameter table of the test program and have values ​​of 0.4, 0.3, 0.2, and 0.1, respectively.

[0151] The normalized integral energy is obtained from formula ④.

[0152] : Logical score, output field of numerical logic The mapping yields: If but ,like but ;

[0153] : Natural constant, approximately equal to 2.71828;

[0154] The load feedback consistency index is obtained from formula ①.

[0155] : Callback trigger flag, obtained from formula ③, is a Boolean value (0 or 1).

[0156] Simple numerical example: Let , , , ,but Finally, the test program is based on Compared with the preset defect threshold (For example, 0.5) is compared to generate a defect determination record. As given by formula ⑥:

[0157] Formula⑥

[0158]

[0159] in, Defect determination record, a Boolean value, where 1 indicates the presence of a defect and 0 indicates the absence of a defect;

[0160] Defect risk score, obtained from formula ⑤;

[0161] The defect threshold is determined by the test program configuration and has a value of 0.5.

[0162] Simple numerical example: Continuing from the previous example ,but This indicates that there are no defects at the current test location. The test program maps the electrical test records to the current test location to form the final test result. The test result includes the load feedback result (i.e., In and circuit breaker detection status), motion control results (i.e. In and callback execution status), electrical test records (i.e. (and intermediate quantities during the testing process) and defect judgment records .

[0163] After this step is completed, the test results are written into the current round of test records and used for the next round of image data acquisition, probe movement, and microcontroller module (MCU) processing.

[0164] As a complete engineering implementation example, in a wafer testing scenario, after the S400 outputs test data, the load feedback module first reads the sampling result field. and digital logic output result fields Calculate according to formula ① The system generates a load feedback record and transmits it to the motion control module. The motion control module reads the output sequences of the first and second comparators for the last five beats and calculates... Below the threshold of 0.6, and Therefore, formula ③ outputs No callback is triggered, and the motion control result is recorded as "Position verification passed". The test program reads the integral result field. and digital logic output According to formula ④, we get Calculate according to formula ⑤ If the value is less than the threshold of 0.5, Formula ⑥ will output... The test was deemed defect-free. The final test results were recorded and used for the closed-loop connection of the next round of testing.

[0165] This section summarizes the technical effects. This step transforms test data into test results containing multi-dimensional verification information through load feedback consistency indicators, motion control deviation accumulation and callback triggering conditions, and defect risk scoring that integrates integrals and logic. Formulas ① to ⑥ form a closed-loop link from the original fields to the final judgment. Among them, the callback mechanism of formula ③ directly drives the mechanical calibration of S200, and the integrated scoring of formula ⑤ integrates the outputs of all preceding modules, giving the test results self-verification and self-driving capabilities, which is significantly different from the conventional single threshold judgment.

[0166] Example 2: Figure 2 A structural block diagram of a high-precision anti-interference semiconductor chip integrated testing device according to an embodiment of the present invention is shown. Figure 2 As shown, the structure may include:

[0167] Image acquisition module 01, connected to the image analysis module, is used to acquire image data of the surface structure of the semiconductor chip. Specifically, the image acquisition module receives the current test state of the semiconductor chip and, after the probe, pin clip, fixed guide rail, and distance adjustment gauge are initially installed, captures images of the surface structure of the semiconductor chip. The image acquisition module includes a camera and a corresponding acquisition interface. The camera continuously captures images of the semiconductor chip surface, generating image data. If the current frame of image data has occlusion, offset, or missing boundaries, the image acquisition module continues to acquire the next frame of image data and writes the current abnormal state into the acquisition record. When the current image data meets the calling conditions, the image acquisition module sends the image data to the image analysis module and writes the current acquisition time and current acquisition state into the corresponding image data record for the image analysis module to read.

[0168] Image analysis module 02 is used to preprocess, detect edges, extract contours, and match contours on the image data to obtain position information. Specifically, the image analysis module receives the image data from the image acquisition module and sequentially performs denoising, grayscale conversion, and binarization on the image data to form a preprocessing result. The image analysis module then performs edge detection on the preprocessing result to obtain edge information, and then performs contour extraction based on the edge information to obtain contour information. The image analysis module calls a contour template to perform contour matching on the contour information to generate a contour matching result. The contour matching result includes center coordinates and surface structure position records. When the current contour matching result overlaps, breaks, or deviates, the image analysis module returns to the preprocessing result to re-execute edge detection, contour extraction, and contour matching, and registers the current reprocessing state in the corresponding record of the image data. The image analysis module organizes the contour matching result into position information and sends the position information to the high-precision positioning module, while providing the motion control module with an entry point for calling the position information.

[0169] A high-precision positioning module 03, connected to the image analysis module, is used to perform secondary verification of the position information using laser positioning and optical positioning. Specifically, the high-precision positioning module receives the position information from the image analysis module and reads the center coordinates, contour matching results, and surface structure position records. The high-precision positioning module first performs laser positioning, then performs optical positioning, and performs secondary verification on the same surface structure position. The laser positioning generates a first verification result. The optical positioning generates a second verification result. When the first verification result matches the second verification result, the high-precision positioning module outputs the verified position information. When the first verification result does not match the second verification result, the high-precision positioning module returns the deviation status to the image analysis module and calls the image analysis module to re-execute contour matching. The verified position information is sent to the motion control module as input for probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment, and the current verification status is recorded in the high-precision positioning module record.

[0170] The motion control module 04 is used to move the probe, slide the fixed guide rail, and adjust the distance adjustment gauge based on the position information to obtain the probe position. Specifically, the motion control module receives the verified position information from the high-precision positioning module and reads the center coordinates and surface structure position records. The motion control module first controls the probe to perform a main movement along the corresponding path of the pin holder, then controls the fixed guide rail to slide along the through hole direction according to the contour matching result, and then reads the current position of the distance adjustment gauge and performs fine adjustment. The probe is fixed on the pin holder. The pin holder includes a PCB board and a ceramic substrate. The fixed guide rail passes through the through hole. The distance adjustment gauge is fixed on the fixed guide rail. After the current main movement of the probe is completed, the motion control module checks the current position of the fixed guide rail and the current position of the distance adjustment gauge. When both correspond to the position information, the motion control module outputs the probe position. When there is a deviation, the motion control module repeats the sliding of the fixed guide rail and the adjustment of the distance adjustment gauge, and writes the current deviation status into the motion control record. The probe position is sent to the voltage follower module, voltage sampling module, current amplification module, and load feedback module for subsequent drive chain calls.

[0171] A voltage follower module, a voltage sampling module, a current amplification module, and a load feedback module 05 are connected to the motion control module and are used to generate a voltage signal and an output current based on the probe position. Specifically, the voltage follower module, voltage sampling module, current amplification module, and load feedback module receive the probe position from the motion control module. The voltage follower module reads the current position information from the probe position and generates a current voltage signal after receiving a control signal from the microcontroller module. The voltage sampling module samples the current voltage signal and sends the sampling result back to the preceding processing chain. The current amplification module outputs the current output current based on the current voltage signal and the sampling result. The load feedback module reads the output current, performs open circuit detection and feedback recording, and returns the feedback result to the voltage sampling module and the motion control module. When the current load feedback module detects an open circuit, it issues a feedback interruption record, and the motion control module calls the previous probe position for re-verification. When no open circuit is detected, the voltage follower module, voltage sampling module, and current amplification module maintain the current link state. The voltage signal and the output current are sent to the circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit and digital logic module, and the current state is recorded in the load feedback module record.

[0172] The circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit, and digital logic module 06 are used to generate test data based on the voltage signal and the output current. Specifically, the circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit, and digital logic module receive the voltage signal and the output current from the voltage follower module, voltage sampling module, current amplification module, and load feedback module. The circuit anti-interference module first performs high-frequency interference signal filtering on the voltage signal and the output current to generate the current sampling input state. The sampler performs clock sampling on the current sampling input state. The first comparator performs a first comparison between the sampling result and the reference voltage provided by the reference voltage source. The integrator performs continuous integration on the first comparison result. The second comparator performs a second comparison between the integration result and the reference voltage. The clock unit provides clock signals corresponding to each processing stage. The digital logic module reads the sampling result, the first comparison result, the integration result, and the second comparison result in clock order and generates a digital logic output result. If a sampling anomaly, comparison anomaly, or integration anomaly occurs within any current cycle, the clock unit issues a call for the next cycle. The sampler and the comparison chain re-execute the current round of processing and write the anomaly status into the corresponding record of the test data. The digital logic module organizes the sampling result, the first comparison result, the integration result, the second comparison result, and the digital logic output result into test data, and sends the test data to the test program, while simultaneously providing read interfaces to the load feedback module and the motion control module.

[0173] Test program module 07 is used to generate test results based on the test data. Specifically, the test program receives test data from the digital logic module, feedback records from the load feedback module, and motion control records from the motion control module. The test program first reads the fields of the test data, then organizes the feedback records and motion control records with the test data to form the current round of test judgment data. The test program performs electrical test processing according to the current round of test judgment data and writes the electrical test processing status into the electrical test record. Subsequently, the test program maps the electrical test record to the current test position to generate a defect judgment record. When the current defect judgment record corresponds to an abnormal state, the test program writes the abnormal state into the test result; when the current defect judgment record corresponds to a normal state, the test program writes the normal state into the test result. The test result includes the feedback status, position status, electrical test record, and defect judgment record. The test result returns the call chain corresponding to the load feedback module, the motion control module, and the image acquisition module for the state connection of the next round of test process, and maintains the sequential correspondence of the current round of test results in the test program record.

Claims

1. A high-precision, anti-interference semiconductor chip comprehensive testing method, characterized in that, include: S100: Acquire image data of the surface structure of the semiconductor chip, perform preprocessing, edge detection and secondary verification to obtain position information; The image data is the raw data of the semiconductor chip surface structure, which includes pads, cathode combs, and the contour boundaries of the semiconductor chip surface. S200. Based on the position information, perform probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment to obtain the probe position; S300. Based on the probe position, current amplification and load feedback closed-loop readback processing are performed to obtain voltage signal and output current; S400. Based on the voltage signal and output current, the circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit and digital logic module are used to process the data to obtain test data. S500. Based on the test data, the load feedback module, motion control module, and test program are processed to obtain the test results.

2. The method according to claim 1, characterized in that, The process of obtaining location information through preprocessing, edge detection, and secondary verification includes: The preprocessing includes denoising, grayscale conversion, and binarization. The Canny algorithm is used for gradient change identification. The find Contours algorithm is used to process continuous edges to obtain contour information, and background contours and interfering contours are filtered out. The contour extraction results are compared with the contour template one by one to obtain contour matching results that include contour position, contour direction and contour integrity. First, establish a reference position for laser positioning based on the center coordinates, and then establish a reference position for optical positioning based on the contour matching results. Compare the laser positioning results, optical positioning results, and initial position information. When there is a positional deviation between the two, call the contour matching results to correct the center coordinates and perform a second verification to obtain the position information. The position information includes the verified center coordinates, the verified contour matching results, and the corresponding surface structure position record.

3. The method according to claim 2, characterized in that, The probe movement process includes: The probe is coarsely moved according to the center coordinates, and the current position value of the distance adjustment gauge is continuously read until the current position value is consistent with the position corresponding to the center coordinates.

4. The method according to claim 3, characterized in that, The process of fixed guide rail sliding includes: Based on the probe movement, the fixed guide rail is controlled to slide along the through hole according to the contour direction in the contour matching result. When the contour matching result shows boundary offset or abnormal contour integrity, the position information of S100 is called to recheck the center coordinates and then the fixed guide rail sliding process is executed.

5. The method according to claim 4, characterized in that, The process of adjusting the distance measuring instrument includes: After the fixed guide rail slides, read the current distance between the probe and the target position, and perform forward or backward adjustment according to the center coordinates and surface structure position. After the adjustment is completed, read the current position value again and check it against the boundary position corresponding to the contour matching result. If the check is inconsistent, return to the fixed guide rail sliding process and re-execute. The probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment are performed sequentially in the order of coarse movement, direction correction, and fine spacing adjustment to obtain the probe position; the probe position includes the current coordinates of the probe, the current position of the fixed guide rail, and the current position of the distance adjustment gauge.

6. The method according to claim 1, characterized in that, The process of performing current amplification and load feedback closed-loop readback processing to obtain the voltage signal and output current includes: The process of current amplification and load feedback closed-loop readback includes: processing by the microcontroller module (MCU), DAC digital-to-analog converter module, voltage follower module, voltage sampling module, current amplification module, and load feedback module; The microcontroller module (MCU) and the digital-to-analog converter (DAC) are connected via a serial peripheral interface (SPI). The MCU first reads the current coordinates of the probe, the current position of the fixed guide rail, and the current position of the distance adjustment gauge, and then generates control signals to write to the register of the DAC. The voltage follower module receives the initial voltage data output by the DAC digital-to-analog converter module, performs voltage follower processing, and compares the followed voltage data with the probe position. When the comparison matches, it outputs a voltage signal. After receiving the voltage signal, the voltage sampling module performs voltage sampling processing, and calibrates the current reading with the current position of the distance adjustment gauge. When the calibration is consistent, the sampled voltage data is output. After receiving the sampled voltage data, the current amplification module performs current amplification processing and compares the current output current with the current position of the fixed guide rail and the current position of the distance adjustment gauge. If the comparison is consistent, the current output current is retained. After receiving the output current, the load feedback module performs load feedback processing, including feedback reading, open circuit detection and feedback transmission. When an open circuit is detected, the microcontroller module (MCU) stops writing the current control signal and recalls the probe position output by S200. The voltage follower module, voltage sampling module, and current amplification module form a voltage series negative feedback method.

7. The method according to claim 6, characterized in that, The process of processing circuit anti-interference module, sampler, first comparator, second comparator, integrator, reference voltage source, clock unit and digital logic module includes: The circuit anti-interference module first filters high-frequency interference signals and performs common-mode rectification on the voltage signal and output current, and then connects to the fully differential wide common-mode input operational amplifier to form the current sampling input state with the help of the reference voltage source; The sampler samples and reads the current sampling input state within the current clock cycle and outputs the sampling result. The first comparator receives the comparison input state after the sampling result is processed by a single-ended output amplifier, and compares it with the reference voltage to obtain the output result of the first comparator; The integrator reads the output of the first comparator within a continuous cycle to form an integral result; The second comparator receives the integration result and compares it with the reference voltage to obtain the output result of the second comparator; Under the control of the clock unit, the digital logic module receives the sampling result, the output result of the first comparator, the integration result, and the output result of the second comparator according to the current clock cycle, and forms the digital logic output result in chronological order. The test data includes a sampling result field, a first comparator output result field, an integration result field, a second comparator output result field, and a digital logic output result field.

8. The method according to claim 7, characterized in that, The processing steps of the load feedback module include: The sampling result field and digital logic output result field are read from the test data, and the status comparison and open circuit detection are performed to generate a load feedback record. When the sampling result field and the digital logic output result field are inconsistent, the inconsistent status is written into the load feedback record and sent to the motion control module.

9. The method according to claim 8, characterized in that, The processing steps of the motion control module include: The system receives load feedback records, the first comparator output result field, and the second comparator output result field. It compares the first comparator output result field and the second comparator output result field according to the same clock unit beat. Based on the comparison result, it determines whether the current probe position is maintained at the current test position. When continuous deviations occur, it calls the probe movement, fixed guide rail sliding, and distance adjustment gauge adjustment processing in S200 to generate motion control results.

10. The method according to claim 1, characterized in that, The testing process includes: It receives load feedback records, motion control results, integral result fields, and digital logic output result fields. First, it reads the integral result field to obtain the cumulative state within the continuous cycle, then reads the digital logic output result field to obtain the current cycle logic state, and then forms the current cycle test judgment data. After performing electrical test processing, it outputs electrical test records, and then forms a defect judgment record corresponding to the current test position. The test results include load feedback results, motion control results, electrical test records, and defect determination records.

Images