Signal testing method, system, laser probe, device, apparatus, and storage medium

Abstract

The embodiment of the application relates to the technical field of hardware testing, in particular to a signal testing method, system, laser probe, device, equipment and storage medium, and aims to improve the accuracy and efficiency of signal testing. The method comprises the following steps: placing a to-be-tested board card on a board card testing table, taking a sample of the to-be-tested board card, and obtaining a board card photo of the to-be-tested board card; performing to-be-tested point analysis on a PCB diagram of the to-be-tested board card, and marking theoretical to-be-tested points and a theoretical grounding point on the PCB diagram; inputting the board card photo and the marked PCB diagram into a to-be-tested point coordinate reading network, and obtaining actual to-be-tested point coordinates and actual grounding point coordinates of the to-be-tested board card; sending the actual to-be-tested point coordinates and the actual grounding point coordinates to a probe controller; controlling, by the probe controller, a laser probe to perform a test on the to-be-tested board card according to the actual to-be-tested point coordinates and the actual grounding point coordinates, and obtaining a test signal waveform diagram at the actual to-be-tested point.

Description

Technical Field

[0001] This application relates to the field of hardware testing technology, and more specifically, to a signal testing method, system, laser probe, device, equipment, and storage medium. Background Technology

[0002] In the hardware development process, the signal quality of each component directly determines the reliability of the product. Therefore, signal integrity testing is crucial in hardware development. Typically, during hardware development testing, test engineers mark the test points and ground points on the PCB layout of the board under test, based on the type and path of the signal under test. They then label and add jumper wires to these points on the board, and finally use probes to penetrate the test points to obtain the waveform of the signal under test for further analysis.

[0003] In practice, it is time-consuming and prone to errors for test engineers to mark and fly wire the test points and grounding points on the circuit board. When manually pressing the test point with a probe, the probe often slips and fails to capture the signal because the test point is often smaller than the probe diameter and the circuit board is relatively smooth. Furthermore, when there are more than two test points, it is impossible for a single person to complete the operation, which greatly limits the accuracy and efficiency of signal testing. Summary of the Invention

[0004] This application provides a signal testing method, system, laser probe, device, equipment, and storage medium, aiming to improve the accuracy and efficiency of signal testing.

[0005] A first aspect of this application provides a signal testing method, the method comprising:

[0006] The board under test is placed on the board test bench, and the board under test is photographed to obtain a board photo.

[0007] The test points are analyzed on the PCB diagram of the board under test, and the theoretical test points and theoretical ground points are marked on the PCB diagram to obtain the marked PCB diagram.

[0008] The circuit board photo and the labeled PCB diagram are input into a pre-trained test point coordinate reading network to obtain the actual test point coordinates and actual ground point coordinates of the circuit board under test.

[0009] The actual coordinates of the test point and the actual grounding point are sent to the probe controller on the board test bench.

[0010] The laser probe is controlled by the probe controller to perform a test on the board under test based on the actual coordinates of the test point and the actual grounding point, thereby obtaining the test signal waveform at the actual test point.

[0011] Optionally, the training steps of the network for reading the coordinates of the point to be measured include:

[0012] Collect PCB diagrams with multiple theoretical test points marked and photos of test boards with multiple unmarked theoretical test points corresponding to the PCB diagrams;

[0013] The multiple PCB diagrams and the photos of the multiple boards under test corresponding to the PCB diagrams are superimposed to obtain multiple training images;

[0014] For each of the multiple training images, the actual test point coordinates of the test board corresponding to the training image are used as the label of the training image to obtain multiple sets of training data.

[0015] The pre-constructed convolutional neural network is trained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

[0016] Optionally, training the pre-constructed convolutional neural network using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured includes:

[0017] The multiple sets of training data are divided into N subsets on average. A subset of the N subsets is used as the training set, and another subset is used as the validation set.

[0018] The convolutional neural network is trained sequentially using the training set and the test set to determine the optimal structure and optimal hyperparameters of the convolutional neural network, thereby obtaining a convolutional neural network with determined structure and parameters.

[0019] The convolutional neural network with determined structural parameters is retrained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

[0020] Optionally, the step of placing the board under test on the board test bench and taking pictures of the board under test to obtain a photograph of the board under test includes:

[0021] The board to be tested is placed on the board test platform according to the preset limit coordinates;

[0022] The test board is photographed by a camera on the test board to obtain a photograph of the test board.

[0023] Optionally, the step of taking pictures of the board under test by the camera on the board test bench to obtain a photograph of the board includes:

[0024] Using the camera, the test board is partially sampled according to preset sampling coordinates to obtain multiple partial photos of the test board.

[0025] The multiple partial photos of the board under test are stitched together to obtain the photo of the board under test.

[0026] Optionally, the step of performing test point analysis on the PCB layout of the board under test, and marking the theoretical test points on the PCB layout to obtain the marked PCB layout, includes:

[0027] Based on the signal type and link of the signal under test in the PCB diagram, determine the theoretical coordinates of the test point and the ground point in the PCB diagram;

[0028] Based on the theoretical test point coordinates and the grounding point coordinates, the theoretical test point and the grounding point are marked on the PCB diagram to obtain the marked PCB diagram.

[0029] Optionally, the step of controlling the laser probe through the probe controller to perform a piercing test on the board under test based on the actual coordinates of the test point and the actual grounding point, and obtaining a test signal waveform at the actual test point, includes:

[0030] The probe controller controls the laser probe to move above the actual coordinates of the point to be measured;

[0031] The probe controller controls the laser probe to move downwards;

[0032] When the probe inside the laser probe contacts the board under test, the test signal at the actual test point is sent to the oscilloscope.

[0033] The oscilloscope generates a waveform diagram of the test signal based on the test signal.

[0034] A second aspect of this application provides a signal testing apparatus, the apparatus comprising:

[0035] The image sampling module is used to place the board under test on the board test bench, take images of the board under test, and obtain a picture of the board under test.

[0036] The theoretical test point annotation module is used to perform test point analysis on the PCB diagram of the board under test, and to annotate the theoretical test points and theoretical ground points on the PCB diagram to obtain the annotated PCB diagram;

[0037] The actual test point acquisition module is used to input the board photo and the labeled PCB diagram into a pre-trained test point coordinate reading network to obtain the actual test point coordinates and actual ground point coordinates of the board under test.

[0038] The coordinate sending module is used to send the actual coordinates of the test point and the actual grounding point to the probe controller on the board test bench.

[0039] The test signal waveform acquisition module is used to control the laser probe through the probe controller to perform a test on the board under test according to the actual test point coordinates and the actual grounding point coordinates, and obtain the test signal waveform at the actual test point.

[0040] Optionally, the training steps of the network for reading the coordinates of the point to be measured include:

[0041] Collect PCB diagrams with multiple theoretical test points marked and photos of test boards with multiple unmarked theoretical test points corresponding to the PCB diagrams;

[0042] The multiple PCB diagrams and the photos of the multiple boards under test corresponding to the PCB diagrams are superimposed to obtain multiple training images;

[0043] For each of the multiple training images, the actual test point coordinates of the test board corresponding to the training image are used as the label of the training image to obtain multiple sets of training data.

[0044] The pre-constructed convolutional neural network is trained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

[0045] Optionally, training the pre-constructed convolutional neural network using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured includes:

[0046] The multiple sets of training data are divided into N subsets on average. A subset of the N subsets is used as the training set, and another subset is used as the validation set.

[0047] The convolutional neural network is trained sequentially using the training set and the test set to determine the optimal structure and optimal hyperparameters of the convolutional neural network, thereby obtaining a convolutional neural network with determined structure and parameters.

[0048] The convolutional neural network with determined structural parameters is retrained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

[0049] Optionally, the imaging sampling module includes:

[0050] The board under test placement submodule is used to place the board under test on the board test platform according to the preset limit coordinates;

[0051] The board photo acquisition submodule is used to capture and sample the board under test using a camera on the board test bench to obtain the board photo.

[0052] Optionally, the board photo acquisition submodule includes:

[0053] The local board under test acquisition submodule is used to take local images of the board under test by means of the camera according to the preset sampling coordinates, and obtain multiple local images of the board under test.

[0054] The photo stitching submodule is used to stitch together the multiple partial photos of the board under test to obtain the photo of the board under test.

[0055] Optionally, the theoretical test point annotation module includes:

[0056] The theoretical test point coordinate determination submodule is used to determine the theoretical test point coordinates and ground point coordinates in the PCB diagram based on the signal type and link of the signal under test in the PCB diagram.

[0057] The theoretical test point coordinate annotation submodule is used to annotate the theoretical test point and the grounding point in the PCB diagram according to the theoretical test point coordinates and the grounding point coordinates, so as to obtain the annotated PCB diagram.

[0058] Optionally, the test signal waveform acquisition module includes:

[0059] The first probe control submodule is used by the probe controller to control the laser probe to move above the actual coordinates of the point to be measured;

[0060] The second probe control submodule is used by the probe controller to control the laser probe to move downward;

[0061] The signal transmission submodule is used to send the test signal at the actual test point to the oscilloscope when the probe inside the laser probe contacts the board under test.

[0062] The test signal waveform generation submodule is used by the oscilloscope to generate the test signal waveform based on the test signal.

[0063] A third aspect of this application provides a signal testing system, comprising:

[0064] Test bench, main unit, oscilloscope;

[0065] The test bench is used to place the board under test and acquire the test signals of the board under test, and includes: N three-dimensional displacement platforms, an adjustable horizontal optical platform, and an electrical discharge controller;

[0066] The N three-dimensional displacement platforms are used to perform tack testing on the test board placed on the adjustable horizontal optical platform;

[0067] The N three-dimensional displacement platforms are installed side by side in the slide rail on one side of the adjustable horizontal optical platform;

[0068] The N three-dimensional displacement platforms are connected to the EDM controller via serial cables;

[0069] Each of the N three-dimensional displacement platforms is equipped with a laser probe, and a camera is installed on the first of the N three-dimensional displacement platforms.

[0070] The laser probe is used to receive the test signal from the board under test and send the test signal to the host and the oscilloscope;

[0071] The host computer is used to control the test bench and the oscilloscope. The host computer is connected to the test bench and the oscilloscope via a data cable.

[0072] The oscilloscope is used to generate a test signal waveform diagram corresponding to the received test signal.

[0073] The fourth aspect of this embodiment provides a laser probe, which is applied to a signal testing system and includes:

[0074] Pump source, used to emit near-infrared laser with a wavelength of the first wavelength;

[0075] A reflector for reflecting the near-infrared laser emitted by the pump source;

[0076] A beam splitter is used to split a reflected near-infrared laser into a first beam and a second beam.

[0077] An optical path compensation system is used to increase the optical path of the first beam;

[0078] A condenser lens is used to receive the first beam and the second beam, and to focus the first beam and the second beam onto the BBO crystal;

[0079] BBO crystal is used to generate a second harmonic based on the first beam and the second beam, and to send the second harmonic to the laser sensor.

[0080] A laser sensor is used to send a signal to the host computer after detecting the second harmonic.

[0081] A fifth aspect of this application provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps described in the first aspect of this application.

[0082] A sixth aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the method described in the first aspect of this application.

[0083] Using the signal testing method provided in this application, the board under test (BUT) is placed on a board test bench, and the BUT is photographed to obtain a photograph of the BUT. The PCB diagram of the BUT is analyzed for test points, and theoretical test points are marked on the PCB diagram to obtain a marked PCB diagram. The BUT photograph and the marked PCB diagram are input into a pre-trained test point coordinate reading network to obtain the actual test point coordinates and actual ground point coordinates of the BUT. The actual test point coordinates and actual ground point coordinates are sent to a probe controller on the board test bench. The probe controller controls a laser probe to perform a penetration test on the BUT based on the actual test point coordinates and actual ground point coordinates, obtaining a test signal waveform at the actual test point.

[0084] In this application, the board under test is tested using a board test bench. The testing process is fully automated. When selecting test points on the board under test, a photograph of the board is first acquired, and then the theoretical test points are marked on the PCB diagram of the board under test. Further, a pre-trained test point coordinate reading network analyzes the board photograph and PCB diagram to determine the actual test point coordinates. Because there may be slight differences between the board under test and the PCB diagram during the printing process, and slight differences may also occur in the position of the board under test on the board test bench, affecting the reliability of the test point marking, the test point coordinate reading network analyzes the board photograph and PCB diagram to eliminate possible slight differences and quickly and accurately obtain the actual test point coordinates. The use of test point coordinates improves the accuracy of board signal testing and saves testing time. When performing penetration testing on the board under test, a probe controller is used to control the laser probe, avoiding potential errors from human operation. The laser probe is equipped with a laser sensor, which has high sensitivity and immediately transmits the signal back to the host and stops the penetration test upon detection, preventing damage to the board under test. Furthermore, the laser does not interact with the electronic components on the board under test, thus avoiding signal quality interference and affecting test results. This ensures accurate signal readings, facilitating subsequent analysis. This application uses network analysis to obtain the actual test point coordinates based on the test point coordinates and controls the laser probe through the probe controller, effectively improving the accuracy and efficiency of hardware signal testing. Attached Figure Description

[0085] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0086] Figure 1 This is a flowchart of a signal testing method proposed in an embodiment of this application;

[0087] Figure 2 This is a schematic diagram of the signal testing system structure proposed in one embodiment of this application;

[0088] Figure 3 This is a schematic diagram of the structure of a laser probe according to an embodiment of this application;

[0089] Figure 4 This is a schematic diagram of the second harmonic generation process according to an embodiment of this application;

[0090] Figure 5 This is a schematic diagram of the optimal neural network architecture proposed in one embodiment of this application;

[0091] Figure 6 This is a schematic diagram of a signal testing device according to an embodiment of this application. Detailed Implementation

[0092] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0093] refer to Figure 1 , Figure 1 This is a flowchart of a signal testing method proposed in an embodiment of this application. Figure 1 As shown, the method includes the following steps:

[0094] S11: Place the board under test on the board test platform, take a picture of the board under test, and obtain a picture of the board under test.

[0095] In this embodiment, reference Figure 2 , Figure 2 This is a schematic diagram of a signal testing system structure proposed in an embodiment of this application. Figure 2 As shown, the system includes:

[0096] Test bench, main unit, oscilloscope;

[0097] The test bench is used to place the board under test and acquire the test signals of the board under test, and includes: N three-dimensional displacement platforms, an adjustable horizontal optical platform, and an electrical discharge controller;

[0098] The N three-dimensional displacement platforms are used to perform tack testing on the test board placed on the adjustable horizontal optical platform;

[0099] The N three-dimensional displacement platforms are installed side by side in the slide rail on one side of the adjustable horizontal optical platform;

[0100] The N three-dimensional displacement platforms are connected to the EDM controller via serial cables;

[0101] Each of the N three-dimensional displacement platforms is equipped with a laser probe, and a camera is installed on the No. 1 three-dimensional displacement platform among the four three-dimensional displacement platforms.

[0102] The laser probe is used to receive the test signal from the board under test and send the test signal to the host and the oscilloscope;

[0103] The host computer is used to control the test bench and the oscilloscope. The host computer is connected to the test bench and the oscilloscope via a data cable.

[0104] The oscilloscope is used to generate a test signal waveform corresponding to the received test signal. The system is characterized by comprising:

[0105] Test bench, main unit, oscilloscope;

[0106] The test bench is used to place the board under test and acquire the test signals of the board under test, and includes: N three-dimensional displacement platforms, an adjustable horizontal optical platform, and an electrical discharge controller;

[0107] The N three-dimensional displacement platforms are used to perform tack testing on the test board placed on the adjustable horizontal optical platform;

[0108] The N three-dimensional displacement platforms are installed side by side in the slide rail on one side of the adjustable horizontal optical platform;

[0109] The N three-dimensional displacement platforms are connected to the EDM controller via serial cables;

[0110] Each of the N three-dimensional displacement platforms is equipped with a laser probe, and a camera is installed on the first of the N three-dimensional displacement platforms.

[0111] The laser probe is used to receive the test signal from the board under test and send the test signal to the host and the oscilloscope;

[0112] The host computer is used to control the test bench and the oscilloscope. The host computer is connected to the test bench and the oscilloscope via a data cable.

[0113] The oscilloscope is used to generate a test signal waveform diagram corresponding to the received test signal.

[0114] In this embodiment, the system includes three parts: a board test platform, a host (13), and an oscilloscope (14). In actual operation, the test platform is an adjustable horizontal optical stage (5) of 1.2×1.2 meters. Four three-dimensional displacement platforms (1,2,3,4) are placed side by side on its left side, which can realize that four probes can be inserted at the same time. In actual application, the number of three-dimensional displacement platforms can be flexibly adjusted as needed. The x-axis and y-axis guide rails of the four three-dimensional displacement platforms have the same length, and the height of the z-axis guide rail decreases from No. 1 to No. 4. This can prevent the probes from interfering with each other when they are displaced. A special laser probe (6,7,8,9) is installed on the platform of the x-axis guide rail of each three-dimensional displacement platform. An additional platform is placed on the x-axis guide rail of the No. 1 three-dimensional displacement platform. A high-definition camera (10) is placed on the platform. The four three-dimensional displacement platforms transmit data to the EDM controller (12) through a serial port. The host controls the EDM controller through a data line. In actual testing, the board under test (11) is placed on the test bench for testing. Testing the board under test involves using a probe to probe the position of the test point on the board under test. When the probe receives the signal of the test point, the signal waveform of the test point is displayed on the oscilloscope. Then, the signal waveform of the test point is analyzed to complete the signal test of the board under test.

[0115] In this embodiment, the board under test is the integrated circuit board that needs to be tested for signals. The board test bench is the platform in the signal testing system on which the board under test is placed. The board photo is a photograph of the board under test taken from the board under test.

[0116] In this embodiment, the board under test is placed on the board test platform, and the board under test is photographed and sampled by a high-definition camera on the No. 1 three-dimensional displacement platform to obtain a board photo of the board under test.

[0117] In this embodiment, step S11 specifically includes the following steps:

[0118] S11-1: Place the board under test on the board test platform according to the preset limit coordinates.

[0119] In this embodiment, the preset limit coordinates are pre-set coordinates used to specify the placement position of the circuit board on the test bench.

[0120] In this embodiment, when performing signal testing on the board under test, the board under test is placed on the board test platform according to the preset limit coordinates.

[0121] For example, the preset limit coordinates are (0,0). During testing, the board under test is placed at the top corner according to the dotted line at the coordinate (0,0) end on the test bench to ensure that the host coordinates are read normally.

[0122] S11-2: The test board is photographed and sampled using a camera on the test board platform to obtain a photograph of the test board.

[0123] In this embodiment, after the board to be tested is placed, the camera on the test bench is used to take pictures of the board to obtain a picture of the board.

[0124] In this implementation, step S11-2 specifically includes the following steps:

[0125] S11-2-1: Using the camera, the test board is partially sampled according to the preset sampling coordinates to obtain multiple partial test board photos.

[0126] In this embodiment, the preset sampling coordinates are the coordinates of the camera's image capture position, and the camera captures a picture of the board under test at the preset sampling coordinates.

[0127] In this embodiment, because the detail requirements of the board photo of the board under test are high, and because the shooting height is limited, it is not possible to take a photo of the board under test in one go. Therefore, it is necessary to use a camera to take multiple partial photos of the board under test according to the preset sampling coordinates to obtain multiple partial photos of the board under test.

[0128] In this embodiment, when taking pictures of the board under test, the high-definition camera on the three-dimensional displacement platform is opened on the host side, and multiple local pictures of the board under test are taken according to the preset sampling coordinates to obtain multiple local pictures of the board under test.

[0129] For example, the preset sampling coordinates are (0, 0), (0, 1), (0, 2), ..., (10, 10). Local photos of the board under test are taken sequentially using these sampling coordinates.

[0130] S11-2-2: The multiple partial photos of the board under test are stitched together to obtain the photo of the board under test.

[0131] In this embodiment, after obtaining multiple partial photos of the board under test, the multiple partial photos of the board under test are stitched together according to the shooting order to obtain the photos of the board under test.

[0132] In this embodiment, when performing signal testing on the board under test, a camera mounted on a three-dimensional displacement platform takes a picture of the board under test. The actual coordinates of the test points are then analyzed based on the picture of the board under test. The host computer controls the camera to automatically take pictures, which quickly and accurately obtains the picture of the board under test, thus improving the efficiency of signal testing.

[0133] S12: Perform test point analysis on the PCB diagram of the board under test, mark the theoretical test points and theoretical ground points on the PCB diagram, and obtain the marked PCB diagram.

[0134] In this embodiment, the PCB (Printed Circuit Board) diagram is a diagram of the actual component placement and wiring drawn according to the circuit schematic. The actual circuit board can be manufactured based on the PCB diagram. The theoretical test point is the signal test point of the board under test that needs to be tested, determined according to the PCB diagram of the board under test. The grounding point is the point where the test circuit is grounded during signal testing.

[0135] In this embodiment, the PCB diagram of the board under test is analyzed for test points. The theoretical test points on the PCB diagram are determined according to the type of signal under test and the link. The theoretical test points are marked to obtain the marked PCB diagram.

[0136] In this embodiment, step S12 specifically includes:

[0137] S12-1: Based on the signal type and link of the signal under test in the PCB diagram, determine the theoretical coordinates of the test point and the grounding point in the PCB diagram.

[0138] In this embodiment, the signals to be tested are the signals of various components on the board under test (BUT) to be tested. The theoretical test point coordinates are the coordinates of the points where the probe should theoretically be inserted during signal testing. Because the actual printed board may differ from the positions of various components and circuits on the PCB layout, the actual test point coordinates on the board may also have slight differences from the theoretical test point coordinates. Depending on the circuit, there may be one or more theoretical test points. The ground point coordinates are the coordinates of the ground point on the BUT during testing; there is only one ground point.

[0139] In this embodiment, the PCB diagram of the board under test is opened on the host side. Based on the signal type of the signal under test and the links on the board under test, the coordinates of the theoretical test point in the PCB diagram are determined. The signal type of the signal under test and the links on the board under test can be divided into many types according to different circuits. The test point can be determined according to the PCB diagram.

[0140] For example, when the signal under test is a clock signal, based on the characteristics of the clock circuit, the test needs to be performed at the circuit's terminal; that is, the coordinates of the clock circuit's terminal are used as the theoretical test point coordinates. Similarly, when testing the voltage signal of a transistor on the board under test, the coordinates of the transistor's collector are used as the theoretical test point coordinates.

[0141] S12-2: Based on the theoretical test point coordinates and the grounding point coordinates, mark the theoretical test point and the grounding point on the PCB diagram to obtain the marked PCB diagram.

[0142] In this embodiment, after determining the coordinates of the theoretical test point and the grounding point, the theoretical test point is marked at the coordinates of the theoretical test point and the grounding point is marked at the coordinates of the grounding point using PCB layout software on the host side, thus obtaining the marked PCB layout.

[0143] For example, for a certain PCB diagram, after determining the coordinates of the theoretical test point and the ground point on the diagram, these two points are marked with red circles on the PCB diagram.

[0144] S13: Input the circuit board photo and the labeled PCB diagram into the pre-trained test point coordinate reading network to obtain the actual test point coordinates and actual ground point coordinates of the circuit board under test.

[0145] In this embodiment, the network for reading the coordinates of the test point is a convolutional neural network designed to determine the signal test points on the circuit board. The actual coordinates of the test point are the actual coordinates of the test point on the circuit board under test during signal testing. The actual ground point coordinates are the actual coordinates of the ground point on the circuit board under test during signal testing.

[0146] In this embodiment, the circuit board photo and the annotated PCB diagram are input into a pre-trained test point coordinate reading network. The coordinate reading network can map the theoretical test point coordinates in the PCB diagram to the corresponding positions in the circuit board photo, and eliminate the positional error between the PCB diagram and the actual test board based on calculations, thereby obtaining the actual test point coordinates and the actual grounding point coordinates.

[0147] In this embodiment, a convolutional neural network is constructed to read the actual coordinates. The advantage of the convolutional neural network is that it can extract key features from the image data through a series of operations such as convolution, rather than mechanically selecting coordinates, so that it has a certain error resistance when selecting the coordinates of the actual test points.

[0148] S14: Send the actual coordinates of the test point and the actual grounding point to the probe controller on the board test bench.

[0149] In this embodiment, the probe controller, also known as the electrical spark controller, is connected to the four three-dimensional displacement platforms via a serial cable to control the movement of the four three-dimensional displacement platforms. It is also connected to the host computer via a data cable and is controlled by the host computer.

[0150] In this embodiment, after determining the actual coordinates of the test point and the actual grounding point, the actual coordinates of the test point and the actual grounding point are sent to the probe controller on the board test bench.

[0151] S15: The laser probe is controlled by the probe controller to perform a test on the board under test according to the actual test point coordinates and the actual grounding point coordinates, and the test signal waveform at the actual test point is obtained.

[0152] In this embodiment, reference Figure 3 , Figure 3 This is a schematic diagram of a laser probe according to an embodiment of this application. The laser probe is applied to the signal testing system described in step S11, and includes:

[0153] Pump source, used to emit near-infrared laser with a wavelength of the first wavelength;

[0154] A reflector for reflecting the near-infrared laser emitted by the pump source;

[0155] A beam splitter is used to split a reflected near-infrared laser into a first beam and a second beam.

[0156] An optical path compensation system is used to increase the optical path of the first beam;

[0157] A condenser lens is used to receive the first beam and the second beam, and to focus the first beam and the second beam onto the BBO crystal;

[0158] BBO crystal is used to generate a second harmonic based on the first beam and the second beam, and to send the second harmonic to the laser sensor.

[0159] A laser sensor is used to send a signal to the host computer after detecting the second harmonic.

[0160] In this embodiment, the laser probe is a probe specifically designed for signal testing of the board under test. Figure 3 The system includes (1) a pump source (800nm ​​optical pulse generator), (2) a reflector, (3) a beam splitter, (4) an optical path compensation system, (5) a condenser lens, (6) a BBO (barium borate) crystal, (7) a 400nm laser sensor, (8) a 400nm second harmonic wave, (9) the point to be measured, and (10) the host computer. Figure 3The probe has a built-in pump source emitter that emits near-infrared laser light at a wavelength of 800nm. After reflection, the light is split into two beams of equal intensity by a beam splitter. One beam, the first beam, is incident on the optical path compensation system to increase the optical path before entering the condenser lens. The other beam, the second beam, is emitted from the probe tip, reflected by the test point, passes through the beam splitter, and enters the condenser lens. The two beams are focused by the condenser lens onto the nonlinear crystal BBO. When the probe tip just touches the test point, the two beams have equal optical path lengths. Due to the frequency doubling effect, the BBO crystal emits a second harmonic at a wavelength of 400nm to the 400nm laser sensor. After detecting the second harmonic, the sensor sends a signal to the host computer. The host computer then controls the EDM controller to terminate the movement of the three-dimensional displacement platform along the z-axis, completing the point test.

[0161] In this embodiment, the host sends the actual coordinates of the test point and the actual grounding point to the probe controller. The probe controller controls the probe to move to the actual test point coordinates and the actual grounding point coordinates to perform a test on the test board. The obtained signal is sent to the oscilloscope to generate a signal waveform diagram at the actual test point.

[0162] In this embodiment, step S15 specifically includes the following steps:

[0163] S15-1: The probe controller controls the laser probe to move above the actual coordinates of the point to be measured.

[0164] In this embodiment, after receiving the actual coordinates of the point to be measured and the actual grounding point sent by the host, the probe controller controls the corresponding three-dimensional displacement platform to move, moving the laser probe directly above the point to be measured.

[0165] For example, when the actual coordinates of the measured point are (1,2) and the actual grounding point is (3,3), control the No.1 three-dimensional displacement platform to move and move the No.6 laser probe to the coordinates (1,2), and control the No.2 three-dimensional displacement platform to move and move the No.7 laser probe to the coordinates (3,3).

[0166] S15-2: The probe controller controls the laser probe to move downwards.

[0167] In this embodiment, the probe controller controls the three-dimensional displacement platform to move along the x and y axes, moving the laser probe above the corresponding coordinate point, and then controls the z-axis of the three-dimensional displacement platform to move downward, causing the laser probe to move downward.

[0168] S15-3: When the probe inside the laser probe touches the board under test, the test signal at the actual test point is sent to the oscilloscope.

[0169] In this embodiment, the laser sensor is an optical signal sensor that sends a signal to the host when it receives an optical signal.

[0170] In this embodiment, when the probe inside the laser probe contacts the test point on the board under test, the two beams of laser emitted by the pump source inside the probe are split by the beam splitter and have equal optical path lengths. Due to the frequency doubling effect, the BBO crystal emits a second harmonic with a wavelength of 400nm to the 400nm laser sensor. After the sensor detects the second harmonic, it sends a signal to the host. The host controls the EDM controller to terminate the movement of the z-axis and complete the point occlusion. When the point occlusion is completed, the signal at the test point will be transmitted to the oscilloscope along the probe.

[0171] S15-4: The oscilloscope generates the waveform of the test signal based on the test signal.

[0172] In this embodiment, the test signal waveform diagram is a waveform diagram generated based on the signal under test, which shows the image features of the signal under test.

[0173] In this embodiment, after receiving the signal under test, the oscilloscope generates a test signal waveform based on the received signal. The test signal waveform is analyzed and saved. Based on the waveform, the signal quality and whether there are any problems with the link can be determined, thus completing the signal testing of the board under test.

[0174] In this embodiment, a fully automated signal testing system is used to achieve the signal testing. The actual coordinates of the test point are determined by a specially constructed test point coordinate reading network. The test board is then pierced by a specially designed laser probe to obtain the test signal waveform. This enables fast and accurate signal testing of the test board.

[0175] In another embodiment of this application, a laser probe is chosen for piercing the test point because it is highly sensitive and largely ensures that the test board will not be damaged. When designing the laser probe, various components were selected based on its principle, and the principle of the laser probe was proven. The following is the proof process of the principle of the laser probe.

[0176] refer to Figure 4 , Figure 4 This is a schematic diagram of the second harmonic generation process according to an embodiment of this application, as shown below. Figure 4As shown, according to the principle of nonlinear optics, assuming that several waves with frequency ω are incident on a nonlinear medium (such as the BBO crystal of this application), due to the second harmonic effect, a second-order nonlinear polarization intensity with a frequency of 2ω will be generated. This polarization intensity acts as an excitation source, generating second-harmonic radiation with a frequency of 2ω, which is output by the medium. This is the process of generating the second harmonic, also known as the frequency doubling process. At the incident end z of the distance sensor, a thin medium with a thickness of dz generates a second-harmonic photoelectric field dE2ω at the output end, which is:

[0177] dE 2ω ∝cos[2ω(tt′)-2k1z]dz (1)

[0178] Where K1 is the angular wave number, and we have:

[0179]

[0180] Where t' is the time required for the second harmonic of frequency 2ω to propagate a distance Lz in the medium.

[0181]

[0182] Where v2 is the propagation speed of the second harmonic in the medium, and by substituting t' into formula (7) and integrating, the second harmonic photoelectric field can be obtained as follows:

[0183]

[0184] Where Δk=2k1-k2, the radiation intensity of the second harmonic at the dielectric output terminal can be obtained as follows:

[0185]

[0186] Normally, Δk≠0, and the value of z affects the direction of dE2ω. The second harmonic radiation generated by dz at different z locations cannot be superimposed in phase, and the total second harmonic power output is extremely small. When Δk is zero, the direction of dE2ω is independent of z, and formula (11) can be simplified to I2ω∝L2. At this time, the second harmonic intensity is only proportional to the square of the thickness L of the nonlinear medium. The second harmonic radiation emitted by the thin layer at different coordinates can be superimposed in phase at the output end, so that the total second harmonic power output reaches the maximum value. Δk=0 is called phase matching. For the negative uniaxial crystal BBO, its Type I phase matching condition is:

[0187]

[0188] Where no and ne are the refractive indices of the o-ray and e-ray calculated according to the dispersion formula, θm is the angle between the wave vector and the optical axis, and the refractive index n of the e-ray is known. e The relationship between (θ) and direction θ is:

[0189]

[0190] Therefore, for the second harmonic:

[0191]

[0192] For a negative uniaxial crystal, when the type 1 phase matching condition is satisfied, the following equation should also be satisfied:

[0193]

[0194] According to formula (9), under the condition of satisfying phase matching, when the wavelength of the incident light of the BBO crystal is 800nm, there exists a unique phase matching angle (29.3°) that allows it to output frequency-doubled light with a wavelength of 400nm. Figure 4 When the optical axis cutting angle of the BBO crystal is 29.3°, there exists a unique crystal rotation angle that makes the optical path lengths of the two incident beams equal, enabling it to emit a second harmonic with a wavelength of 400nm.

[0195] This embodiment demonstrates the accuracy of laser probes in piercing the test board by proving their principle. Compared to the controllable force of hand-held probes, mechanically controlled probes cannot control the force, only the distance. Excessive piercing distance can cause the particles to be too large, damaging both the probe and the board, while insufficient piercing distance can result in poor contact and failure to capture the test signal. Laser sensors have high precision and sensitivity, and the laser does not interact with the electronic components on the test board, thus avoiding signal quality interference and affecting test results. Therefore, this embodiment finds a unique phase matching angle based on the principle of the laser probe. At this angle, the laser probe can effectively detect the test signal of the test board, ensuring that the test board is not damaged while still accurately detecting the test signal.

[0196] In another embodiment of this application, the training steps of the network for reading the coordinates of the point to be measured include:

[0197] S21: Collect multiple PCB diagrams with labeled theoretical test points and photos of multiple test boards without labeled theoretical test points corresponding to the PCB diagrams.

[0198] In this embodiment, when training the network for reading the coordinates of the points to be tested, firstly, multiple PCB diagrams with labeled theoretical points to be tested and multiple photos of the boards to be tested without labeled theoretical points to be tested corresponding to the PCB diagrams are collected. The photos of the boards to be tested and the PCB diagrams are then compared. Figure 1 The ratio of the PCB diagram to the photograph of the board under test is 1:1.

[0199] S22: Overlay the multiple PCB diagrams and the multiple test board photos corresponding to the PCB diagrams to obtain multiple training images.

[0200] In this embodiment, each PCB diagram in multiple PCB diagrams is superimposed with a photo of its corresponding board under test to obtain multiple training images.

[0201] For example, using sizes of 10 4 ×10 4 The PCB diagram of the board with labeled theoretical test points (×3) and the photo of the board without labeled theoretical test points are superimposed as the input to the neural network (size 10). 4 ×10 4 ×6), where the PCB diagram and the actual board under test are kept at a 1:1 ratio.

[0202] S23: For each of the multiple training images, the actual test point coordinates of the test board corresponding to the training image are used as the label of the training image to obtain multiple sets of training data.

[0203] In this embodiment, a corresponding label is added to each training image. The label content is the actual coordinate point of the test board, which is obtained in advance on the test board. Each training image labeled with the actual coordinate point is a set of training data. By labeling all training images, multiple sets of training data are obtained.

[0204] S24: Train the pre-constructed convolutional neural network using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

[0205] In this embodiment, the multiple sets of training data are divided into a training set, a validation set, and a test set in an 8:1:1 ratio. The training and validation sets are used for hyperparameter selection and architecture optimization, while the test set is used to verify the model's reconstruction capability and generalization ability. The convolutional neural network is first trained using the training and validation sets, and then validated using the validation set, ultimately resulting in a network for reading the coordinates of the points to be tested.

[0206] In this embodiment, the proposed convolutional neural network consists of convolutional layers and fully connected layers. The convolutional layers extract image features from the PCB diagram and the photograph of the board under test through convolutional kernels. By connecting with the fully connected layers, the latent vectors obtained from the image feature encoding are mapped to the actual coordinates of the test points on the board under test. In the multi-layer convolution operation, the data x of size H×W×C is convolved with the convolution kernel of size f1×f2 in the Lth convolutional layer to obtain the feature map z of the c-th channel, following the formula:

[0207]

[0208] In the formula k is the pixel value in the convolution kernel, and H' and W' are the sizes of the feature map after convolution.

[0209]

[0210]

[0211] Where f is the kernel size, s is the kernel stride, and p is the number of padding layers in the input. After the convolution operation, a feature map with size H'×W'×C' is obtained for channel C', and the activation value a is obtained by applying the non-linear activation function ReLU. c :

[0212] a c =g(z) c ) = max(z c ,0) (13)

[0213] Activation value a c It will be used as input in the next convolution until all convolution operations are completed and a size of H is obtained. n ×W n ×C n The feature map.

[0214] The feature maps obtained from the convolutional layers are converted into one-dimensional vectors and then input into fully connected layers. Fully connected layers consist of single-layer perceptrons containing multiple neurons, with neurons tightly connected between layers to form a fully connected network. A set of units in each layer calculates a weighted sum of the inputs from the previous layer and passes the result through the non-linear activation function ReLU, as shown in the following equation:

[0215]

[0216] In the formula z (l) Represents the relationship with the weight matrix W (l) The bias vector b of the l-th layer (l) and the activation vector a from the previous layer (l-1) Linearly correlated hidden vectors. Adjustable parameter (W) (l) ,b (l) Weights, also known as inputs or outputs, are key to achieving optimal input-output functionality in neural networks.

[0217] To correctly adjust the weight vector, a backpropagation learning algorithm is introduced to propagate the computational error of the objective function from the output layer back to the input layer. To optimize the trainable parameters, a mean squared error loss function is introduced as the convergence objective.

[0218]

[0219] In the formula x (i) y (i) and ...

[0220] In this embodiment, step S24 specifically includes the following steps:

[0221] S24-1: Divide the multiple sets of training data into N subsets on average, and use a subset of the N subsets as the training set and another subset as the validation set.

[0222] In this embodiment, when training the convolutional neural network, 10-fold cross-validation and hold-out are used as methods to adjust and evaluate the neural network structure and hyperparameters. Hold-out involves dividing multiple sets of training data into N subsets, such as ten subsets. During training, nine subsets are selected as the training set and one subset is selected as the validation set for training. The performance index of the model is calculated to determine the optimal combination of hyperparameters of the neural network in the target task.

[0223] S24-2: Train the convolutional neural network sequentially using the training set and the validation set to determine the optimal structure and optimal hyperparameters of the convolutional neural network, thereby obtaining a convolutional neural network with determined structural parameters.

[0224] In this embodiment, the convolutional neural network is trained sequentially using a training set and a validation set to determine the optimal structure and optimal hyperparameters of the convolutional neural network, thus obtaining a convolutional neural network with determined structural parameters.

[0225] In this embodiment, reference Figure 5 , Figure 5 This is a schematic diagram of the optimal neural network architecture proposed in an embodiment of this application. Where f is the kernel size, s is the stride, is the padding, c is the channel of output, FC is the fully connected layer, Conv is the convolutional layer, Kernel is the convolution kernel, Conv operation is the convolution operation, H... n ×W n ×C n For the Cth n The dimensions after performing a convolution operation on each channel. Figure 5 In the middle, in the size of 10 4 ×104 After the superimposed image of the ×6 PCB diagram and the photo of the board under test is input into the convolutional neural network, the convolutional neural network performs feature extraction and multi-layer convolution operations to obtain the actual coordinates of the test point and the actual ground point of the board under test.

[0226] In this embodiment, the exploration of the optimal model architecture and hyperparameter space is divided into two parts. For convolutional layers, considering the dimension and size of the input data, a five-layer architecture is started, and the depth of the convolutional layers is gradually increased to improve learning ability until a saturation point is reached. While increasing the convolutional depth, different combinations of kernel size, strides, padding, and channel number are explored, and the parameters at the time of verification error convergence are recorded. The fully connected layers are selected in the same way in terms of the number of unit layers. Unlike the architecture exploration, considering the limitations of computing resources and time, the hyperparameters are not explored in a mesh format, but rather explored intuitively from a small range of values ​​for each hyperparameter. The learning rate value is tried from 10. -3 Up to 10 -6 Other hyperparameter attempts were also carried out in the same way.

[0227] In this embodiment, the Adam algorithm is used instead of the traditional gradient descent algorithm to achieve convergence of the loss function. The Adam algorithm performs well in training neural networks and outperforms other stochastic optimization methods. After the loss function converges, the optimal structure and hyperparameters of the neural network are determined.

[0228] S24-3: The convolutional neural network with determined structural parameters is retrained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

[0229] In this embodiment, after obtaining a convolutional neural network with determined structural parameters, the convolutional neural network with determined structural parameters is retrained using a complete training set, i.e., multiple sets of training data, to obtain a network for reading the coordinate points to be tested.

[0230] For example, the network for reading the coordinates of the target point is implemented using Python 3.7 (programming language) on a GPU server with a Tesla H100 (architecture), and the Keras (neural network library) framework is used to support its construction.

[0231] In this embodiment, to eliminate the differences between the theoretical and actual coordinates of the test points caused by the differences between the test board and the PCB layout during the printing process, as well as the differences generated when the test board is placed on the test bench, a pre-designed test point coordinate reading network analyzes the PCB layout and the test board photograph. The test point coordinate reading network calculates the actual test point coordinates through convolution operations. When training the convolutional neural network, the optimal architecture and parameters of the convolutional neural network are found through 10-fold cross-validation and hold-out method. After determining the optimal architecture and parameters of the convolutional neural network, the convolutional neural network is retrained using training data. Finally, the test point coordinate reading network can quickly obtain the actual test point coordinates of the test board.

[0232] Based on the same inventive concept, one embodiment of this application provides a signal testing device. (Reference) Figure 6 , Figure 6 This is a schematic diagram of a signal testing apparatus 600 according to an embodiment of this application. Figure 6 As shown, the device includes:

[0233] The image sampling module 601 is used to place the board under test on the board test bench, take images of the board under test, and obtain a board photo of the board under test.

[0234] The theoretical test point annotation module 602 is used to perform test point analysis on the PCB diagram of the board under test, and to annotate the theoretical test points and theoretical ground points on the PCB diagram to obtain the annotated PCB diagram.

[0235] The actual test point acquisition module 603 is used to input the board photo and the labeled PCB diagram into a pre-trained test point coordinate reading network to obtain the actual test point coordinates and actual ground point coordinates of the board under test.

[0236] The coordinate sending module 604 is used to send the actual coordinates of the test point and the actual grounding point to the probe controller on the board test bench.

[0237] The test signal waveform acquisition module 605 is used to control the laser probe through the probe controller to perform a test on the board under test according to the actual test point coordinates and the actual grounding point coordinates, and obtain the test signal waveform at the actual test point.

[0238] Optionally, the training steps of the network for reading the coordinates of the point to be measured include:

[0239] Collect PCB diagrams with multiple theoretical test points marked and photos of test boards with multiple unmarked theoretical test points corresponding to the PCB diagrams;

[0240] The multiple PCB diagrams and the photos of the multiple boards under test corresponding to the PCB diagrams are superimposed to obtain multiple training images;

[0241] For each of the multiple training images, the actual test point coordinates of the test board corresponding to the training image are used as the label of the training image to obtain multiple sets of training data.

[0242] The pre-constructed convolutional neural network is trained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

[0243] Optionally, training the pre-constructed convolutional neural network using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured includes:

[0244] The multiple sets of training data are divided into N subsets on average. A subset of the N subsets is used as the training set, and another subset is used as the validation set.

[0245] The convolutional neural network is trained sequentially using the training set and the test set to determine the optimal structure and optimal hyperparameters of the convolutional neural network, thereby obtaining a convolutional neural network with determined structure and parameters.

[0246] The convolutional neural network with determined structural parameters is retrained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

[0247] Optionally, the imaging sampling module includes:

[0248] The board under test placement submodule is used to place the board under test on the board test platform according to the preset limit coordinates;

[0249] The board photo acquisition submodule is used to capture and sample the board under test using a camera on the board test bench to obtain the board photo.

[0250] Optionally, the board photo acquisition submodule includes:

[0251] The local board under test acquisition submodule is used to take local images of the board under test by means of the camera according to the preset sampling coordinates, and obtain multiple local images of the board under test.

[0252] The photo stitching submodule is used to stitch together the multiple partial photos of the board under test to obtain the photo of the board under test.

[0253] Optionally, the theoretical test point annotation module includes:

[0254] The theoretical test point coordinate determination submodule is used to determine the theoretical test point coordinates and ground point coordinates in the PCB diagram based on the signal type and link of the signal under test in the PCB diagram.

[0255] The theoretical test point coordinate annotation submodule is used to annotate the theoretical test point and the grounding point in the PCB diagram according to the theoretical test point coordinates and the grounding point coordinates, so as to obtain the annotated PCB diagram.

[0256] Optionally, the test signal waveform acquisition module includes:

[0257] The first probe control submodule is used by the probe controller to control the laser probe to move above the actual coordinates of the point to be measured;

[0258] The second probe control submodule is used by the probe controller to control the laser probe to move downward;

[0259] The signal transmission submodule is used to send the test signal at the actual test point to the oscilloscope when the probe inside the laser probe contacts the board under test.

[0260] The test signal waveform generation submodule is used by the oscilloscope to generate the test signal waveform based on the test signal.

[0261] Based on the same inventive concept, another embodiment of this application provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the signal testing method as described in any of the above embodiments of this application.

[0262] Based on the same inventive concept, another embodiment of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the steps of the signal testing method described in any of the above embodiments of this application.

[0263] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0264] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0265] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0266] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0267] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0268] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0269] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0270] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0271] The signal testing method, apparatus, device, and storage medium provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A signal testing method, characterized in that, The method includes: The board under test is placed on the board test bench, and the board under test is photographed to obtain a board photo. The test points are analyzed on the PCB diagram of the board under test, and the theoretical test points and theoretical ground points are marked on the PCB diagram to obtain the marked PCB diagram. The circuit board photo and the labeled PCB diagram are input into a pre-trained test point coordinate reading network to obtain the actual test point coordinates and actual ground point coordinates of the circuit board under test. The actual coordinates of the test point and the actual grounding point are sent to the probe controller on the board test bench. The laser probe is controlled by the probe controller to perform a test on the board under test based on the actual coordinates of the test point and the actual grounding point, thereby obtaining the test signal waveform at the actual test point.

2. The method according to claim 1, characterized in that, The training steps of the network for reading the coordinates of the point to be measured include: Collect PCB diagrams with multiple theoretical test points marked and photos of test boards with multiple unmarked theoretical test points corresponding to the PCB diagrams; The PCB diagrams with the labeled theoretical test points and the photos of the multiple test boards corresponding to the PCB diagrams are superimposed to obtain multiple training images; For each of the multiple training images, the actual test point coordinates of the test board corresponding to the training image are used as the label of the training image to obtain multiple sets of training data. The pre-constructed convolutional neural network is trained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

3. The method according to claim 2, characterized in that, The step of training a pre-constructed convolutional neural network using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured includes: The multiple sets of training data are divided into N subsets on average. A subset of the N subsets is used as the training set, and another subset is used as the validation set. The convolutional neural network is trained sequentially using the training set and the validation set to determine the optimal structure and optimal hyperparameters of the convolutional neural network, thereby obtaining a convolutional neural network with determined structure and parameters. The convolutional neural network with determined structural parameters is retrained using the multiple sets of training data to obtain the network for reading the coordinates of the point to be measured.

4. The method according to claim 1, characterized in that, The step of placing the board under test on the board test bench and taking pictures of the board under test to obtain a picture of the board under test includes: The board to be tested is placed on the board test platform according to the preset limit coordinates; The test board is photographed by a camera on the test board to obtain a photograph of the test board.

5. The method according to claim 4, characterized in that, The step of taking pictures of the board under test by the camera on the board test bench to obtain the board photo includes: Using the camera, the test board is partially sampled according to preset sampling coordinates to obtain multiple partial photos of the test board. The multiple partial photos of the board under test are stitched together to obtain the photo of the board under test.

6. The method according to claim 1, characterized in that, The step involves analyzing the test points on the PCB layout of the board under test, marking the theoretical test points on the PCB layout, and obtaining the marked PCB layout, including: Based on the signal type and link of the signal under test in the PCB diagram, determine the theoretical coordinates of the test point and the ground point in the PCB diagram; Based on the theoretical test point coordinates and the grounding point coordinates, the theoretical test point and the grounding point are marked on the PCB diagram to obtain the marked PCB diagram.

7. The method according to claim 1, characterized in that, The process involves controlling a laser probe via the probe controller to perform a penetration test on the board under test based on the actual coordinates of the test point and the actual grounding point, thereby obtaining a test signal waveform at the actual test point, including: The probe controller controls the laser probe to move above the actual coordinates of the point to be measured; The probe controller controls the laser probe to move downwards; When the probe inside the laser probe contacts the board under test, the test signal at the actual test point is sent to the oscilloscope. The oscilloscope generates a waveform diagram of the test signal based on the test signal.

8. A signal testing device, characterized in that, The device includes: The image sampling module is used to place the board under test on the board test bench, take images of the board under test, and obtain a picture of the board under test. The theoretical test point annotation module is used to perform test point analysis on the PCB diagram of the board under test, and to annotate the theoretical test points and theoretical ground points on the PCB diagram to obtain the annotated PCB diagram; The actual test point acquisition module is used to input the board photo and the labeled PCB diagram into a pre-trained test point coordinate reading network to obtain the actual test point coordinates and actual ground point coordinates of the board under test. The coordinate sending module is used to send the actual coordinates of the test point and the actual grounding point to the probe controller on the board test bench. The test signal waveform acquisition module is used to control the laser probe through the probe controller to perform a test on the board under test according to the actual test point coordinates and the actual grounding point coordinates, and obtain the test signal waveform at the actual test point.

9. A signal testing system, characterized in that, The system includes: Test bench, main unit, oscilloscope; The test bench is used to place the board under test and acquire the test signals of the board under test, and includes: N three-dimensional displacement platforms, an adjustable horizontal optical platform, and an electrical discharge controller; The N three-dimensional displacement platforms are used to perform tack testing on the test board placed on the adjustable horizontal optical platform; The N three-dimensional displacement platforms are installed side by side in the slide rail on one side of the adjustable horizontal optical platform; The N three-dimensional displacement platforms are connected to the EDM controller via serial cables; Each of the N three-dimensional displacement platforms is equipped with a laser probe, and a camera is installed on the first of the N three-dimensional displacement platforms. The laser probe is used to receive the test signal from the board under test and send the test signal to the host and the oscilloscope; The host computer is used to control the test bench and the oscilloscope. The host computer is connected to the test bench and the oscilloscope via a data cable. The oscilloscope is used to generate a test signal waveform diagram corresponding to the received test signal.

10. A laser probe, characterized in that, The laser probe is used in the signal testing system of claim 9, comprising: Pump source, used to emit near-infrared laser with a wavelength of the first wavelength; A reflector for reflecting the near-infrared laser emitted by the pump source; A beam splitter is used to split a reflected near-infrared laser into a first beam and a second beam. An optical path compensation system is used to increase the optical path of the first beam; A condenser lens is used to receive the first beam and the second beam, and to focus the first beam and the second beam onto the BBO crystal; BBO crystal is used to generate a second harmonic based on the first beam and the second beam, and to send the second harmonic to the laser sensor. A laser sensor is used to send a signal to the host computer after detecting the second harmonic.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1 to 7.

Images