High-throughput probe-type magnetic sensor full-band test system based on matrix switch

Abstract

The application discloses a high-flux probe type magnetic sensor full-band test system based on a matrix switch and belongs to the technical field of magnetic sensor testing, which comprises a task queue management module, a switching protection and air window control module, a stability detection and judgment module, a full-band test and quality control module and a spin torque ferromagnetic resonance analysis module. The application solves the technical problems in the prior art that due to the limited low-temperature or strong magnetic field test environment space, the sample test interval is deep and the space is compact, the traditional manual sample replacement, probe adjustment and multi-sample switching operation are extremely difficult to implement, it is difficult to balance the automation and high-precision requirements, and when the matrix switch switches the signal path, the dynamic adjustment of physical and electrical parameters easily leads to signal instantaneous distortion and unstable high-frequency measurement, thereby affecting the test result precision and reliability.

Description

A Full-Band Testing System for High-Throughput Probe-Type Magnetic Sensors Based on Matrix Switches Technical Field

[0001] This invention relates to the field of magnetic sensor testing technology, and in particular to a high-throughput probe-type magnetic sensor full-band testing system based on matrix switches. Background Technology

[0002] Existing magnetic sensor testing methods typically begin by applying varying magnetic fields, temperatures, or angles to the sensor under test using probes, sample holders, or mechanical structures to create diverse test conditions. Then, under these varying conditions, the sensor's electrical response (e.g., voltage and current changes) is acquired in real-time, and its output signal is recorded. By processing the test data, the sensor's performance stability, durability, and response characteristics under different conditions can be evaluated. In some specialized applications, cyclic loading is also required to determine the sensor's fatigue life, ensuring its reliability over long-term use. The core of this testing method lies in the precise control of changes in the test environment and the real-time acquisition of electrical signals. By comparing and analyzing the sensor's response under different conditions, the performance of the magnetic sensor can be comprehensively evaluated.

[0003] For example, Chinese invention patent CN102866374B discloses a probe-based magnetic sensor testing method. The method involves fabricating a probe card based on the circuitry of the magnetic sensor under test, placing the sensor on a test bench, setting at least one magnetic coil on the X-axis, and controlling the direction and intensity of the magnetic field on the X-axis of the test bench by controlling the magnitude and direction of the current in the magnetic coil. Based on the output of the magnetic sensor under test, the bias output and / or sensitivity of the X-axis of the sensor are calculated. A motor controls the contact between the probe and the magnetic sensor on the test bench until all magnetic sensors on the test bench have been tested.

[0004] Current magnetic sensor testing methods primarily rely on manual operation for sample switching and frequency band adjustment. Manual adjustments to the testing equipment are required between each sample test, leading to inefficiency. While these methods effectively meet the needs of small-sample, low-throughput testing, technological advancements and increasing application demands, particularly in new material development, industrial manufacturing, and IoT applications, necessitate higher requirements for testing efficiency and accuracy. These scenarios demand automated, high-throughput, multi-frequency testing of numerous different types of magnetic sensors to meet requirements for rapid screening, mass production, and real-time monitoring. Especially in new material development, the process of testing numerous samples to screen for optimal materials requires automated, high-throughput testing systems. Similarly, rapid testing of each batch of products in industrial production demands efficient methods. Furthermore, sensors in IoT applications require periodic testing in various environments. Traditional manual operation methods are clearly inefficient in addressing these needs. To solve these problems, a matrix switch capable of automatically switching samples and frequency bands and managing multiple signal paths is used as an automated control system to reduce manual intervention and meet the demands of multi-sample, high-throughput, and high-frequency testing. However, when a matrix switch switches signal paths, while the relay physically switches the signal path, the electrical path adjustment still requires a certain amount of time to stabilize. Due to minute changes in the relay contacts during switching, impedance mismatch in the signal path, and dynamic adjustments in current flow, this transition process can cause momentary signal distortion and reflection. Specifically, contact changes during relay switching lead to instability in the electrical path impedance, resulting in signal reflection or attenuation; simultaneously, when current switches from one path to another, electrical parameters (such as impedance and capacitance) experience brief fluctuations, causing momentary signal distortion. Even after the relay completes the mechanical switching, the electrical transition process of the signal path still requires time to stabilize. Especially in high-frequency (GHz or THz band) testing, due to the short wavelength of the signal, these minute electrical fluctuations can lead to significant signal errors, affecting the stability and accuracy of the test results. Although the transition time is extremely short, these changes are sufficient to cause measurement instability at high frequencies, affecting the reliability and accuracy of the data.

[0005] Furthermore, with the increasing research demands in cutting-edge fields such as quantum devices, more and more magnetic sensor tests are required to be conducted in extreme environments such as ultra-low temperatures and strong magnetic fields. In such extreme environments, the sample testing area is typically deep and the space is extremely limited, making it difficult to complete high-precision electrical tests within this confined space using traditional manual methods. Especially in plug-in cryogenic platforms, a probe-based automated docking method is necessary to achieve rapid sample replacement and reliable contact. However, existing probe-based cryogenic magnetic field equipment typically has limitations in terms of achievable temperature and magnetic field strength, making it difficult to simultaneously meet the dual requirements of extreme environments and batch automated testing. Summary of the Invention

[0006] To address the technical problems in existing technologies, such as limited space in low-temperature or strong magnetic field testing environments, deep and compact sample testing areas, making it extremely difficult to manually change samples, adjust probes, and switch between multiple samples, and to simultaneously meet automation and high-precision requirements, as well as the problem of instantaneous signal distortion and high-frequency measurement instability due to dynamic adjustments of physical and electrical parameters when switching signal paths using matrix switches, thus affecting the accuracy and reliability of test results, this invention provides a high-throughput probe-type magnetic sensor full-band testing system based on a matrix switch. By integrating probe-type operation into a low-temperature insertion rod structure, it achieves automatic and precise contact and switching of multiple samples in extreme environments. Combined with matrix switches and automated signal management, it realizes fully automatic batch switching of samples, stable acquisition of high-frequency signals, and full-process test task scheduling, effectively improving testing efficiency, data consistency, and environmental adaptability. Specifically, it includes: a task queue management module, a switching protection and window control module, a stability detection and judgment module, a full-band testing and quality control module, and a ferromagnetic resonance analysis module.

[0007] The task queue management module is used to establish a test task queue. The test task queue includes at least the sample number, probe position, test parameter settings, and port mapping relationship between the matrix switch input port and the coaxial output port. The task queue is executed one by one while keeping the probe in contact.

[0008] The switching protection and window control module is used to control the RF output to enter a silent state and the detection acquisition to enter a hold state before the matrix switch switches to the target sample channel for each task. It determines the hold duration of the window period associated with the current test frequency band and issues a matrix switch switching command. After switching, it enters the window period and prohibits the acquisition results from being written into the formal test data during the window period.

[0009] The stability detection and determination module is used to perform stability detection and acquisition in the target sample channel after the window period ends to determine stability. When the stability is determined to be stable, the RF output is restored and the hold state is released to enter the formal test. When the stability is determined to be unstable, the waiting period is extended and the stability detection and determination are repeated until the stability is achieved or the timeout occurs.

[0010] The full-band testing and quality control module is used to perform frequency sweep and / or field sweep acquisition for the current test task during formal testing, perform preheating and formal sampling for each test point, output the sampled test data of each test point, determine the quality label of the current test task, and trigger retest when the quality label does not meet the requirements and replace the corresponding test data with the retest result, thereby forming the confirmation test data of the current test task.

[0011] The spin torque ferromagnetic resonance analysis module is used to perform spin torque ferromagnetic resonance analysis on the confirmation test data of each test task in the test task queue, output the analysis results of each test task, and associate and store the analysis results of each test task with the confirmation test data.

[0012] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0013] 1. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches provided by this invention integrates modules such as automatic matrix switch switching, switching protection and window control, stability detection and judgment, full-band testing and quality control, and spin torque ferromagnetic resonance analysis. This enables multi-sample, full-band, automated, and batch testing of magnetic sensors. It can automatically and efficiently switch test samples and parameters according to the task queue without manual lead replacement or probe movement, improving testing efficiency and batch data consistency. The switching protection and window control module solves the problems of instantaneous signal distortion and high-frequency measurement instability during relay switching, ensuring signal integrity and data reliability under high-frequency and full-band conditions. The stability detection and judgment module ensures that the signal path fully recovers stability after each switch, reducing the probability of false detection and distortion. The full-band testing and quality control module supports automatic frequency sweep and / or field sweep testing and introduces a quality judgment and automatic retest mechanism to eliminate abnormal data and compensate for occasional faults, ensuring the scientific validity and accuracy of the data. The spin-torque ferromagnetic resonance analysis module can automatically fit data and extract physical parameters from batch test results, and directly output key spintronics parameters such as resonance field, linewidth, and symmetry / antisymmetry components, supporting diverse needs such as materials research and development, performance screening, and basic research.

[0014] 2. This invention improves test reliability and data consistency after each channel switch by performing stability detection and acquisition in the target sample channel after the window period ends to determine stability. It ensures that after mechanical switching and electrical parameter recovery, transient interference caused by factors such as relay contact jitter and electrical instability is eliminated, avoiding the impact of switching transients on formal data in high-frequency testing environments. Therefore, subsequent formal data acquisition and analysis are only initiated after the signal channel state is determined to be stable, guaranteeing the accuracy, repeatability, and traceability of acquired data in high-throughput automated testing scenarios, further enhancing the practical value and engineering applicability of the testing system under multi-sample batch and full-frequency band conditions.

[0015] 3. This invention automates the quality assessment of the current testing task by determining its quality label, enabling real-time hierarchical management of data quality for each test point and the overall task in high-throughput batch testing scenarios. By scientifically evaluating the ratio of qualified to unqualified test points, it can promptly screen out data anomalies or acquisition errors, preventing invalid or low-quality data from entering subsequent analysis stages and improving the reliability and usability of test results. Furthermore, the quality labeling mechanism provides a clear basis for subsequent automated retesting, process review, and manual intervention, enabling the entire testing process to possess self-inspection, self-repair, and quality closed-loop capabilities under conditions of large-scale, multi-sample, and full-frequency acquisition, thereby enhancing the intelligence level and engineering application value of the testing system. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 is a framework diagram of a full-band test system for a high-throughput probe-type magnetic sensor based on a matrix switch provided in an embodiment of this application.

[0018] Figure 2 is a schematic diagram of the 6×24 matrix switch structure involved in this embodiment;

[0019] Figure 3 is a schematic diagram of the automatic switching test matrix for the multi-sample four-end method;

[0020] Figure 4 shows a schematic diagram of the ST-FMR test circuit connection;

[0021] Figure 5 is a block diagram of the Bias-T principle;

[0022] Figure 6 shows the appearance of the anti-static transfer box and its internal circuit structure.

[0023] Figure 7 is a schematic diagram of the structure of the low-temperature testing component involved in this embodiment;

[0024] Figure 8 is a partial enlarged view of part A in the structural schematic diagram of the low-temperature test assembly;

[0025] Figure 9 shows the S-parameter curves of the low-temperature sample rod in the 80 GHz range;

[0026] Figure 10 is a schematic diagram of a partial structure of the probe-type ST-FMR;

[0027] Figure 11 shows the array layout of the probe-type ST-FMR device;

[0028] Figure 12 shows typical data and fitting results of the ST-FMR test;

[0029] Figure 13 shows typical data and fitting results of FMR differential spectrum;

[0030] Figure 14 shows the resonant field-frequency dispersion curve;

[0031] Figure 15 shows the frequency-linewidth curve.

[0032] Reference numerals: 1. Shock-absorbing spring; 2. Z-axis motion stage; 3. X-axis motion stage; 4. Y-axis motion stage; 5. Probe section; 6. Observation window / low-temperature objective / optical fiber optic cable; 7. Low-temperature frame; 8. Z-axis piezoelectric motor; 9. Y-axis piezoelectric motor; 10. High-frequency probe; 11. Rotary piezoelectric motor; 12. X-axis piezoelectric motor; 13. Stage; 14. Micro / nano-fabricated STFMR device; 15. Low-temperature high-frequency cable. Detailed Implementation

[0033] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0034] Embodiment 1 of the present invention: As shown in Figure 1, the framework diagram of the high-throughput probe-type magnetic sensor full-band test system based on matrix switch includes a task queue management module, a switching protection and window control module, a stability detection and judgment module, a full-band test and quality control module, and a spin torque ferromagnetic resonance analysis module.

[0035] Figure 2 shows a schematic diagram of the 6×24 matrix switch structure involved in this embodiment, illustrating a matrix relay switch array with 6 rows and 24 columns. Horizontally, there are 6 coaxial output ports, and vertically, there are 24 input ports. Each node is a dual-channel low-noise relay capable of simultaneously switching signal lines and shielding lines, enabling arbitrary connection between inputs and outputs. By selecting and controlling different relays, automatic connection between the output terminal and any input terminal (i.e., any sample under test) can be achieved, effectively supporting high-throughput, multi-channel electrical testing. Figure 3 shows a schematic diagram of the multi-sample four-terminal automatic switching test matrix, using four-terminal resistance testing as an example to demonstrate the application of the M6×24 matrix switch. Horizontally, each column corresponds to a magnetic sensor under test (e.g., Sample1, Sample2…), and vertically, each row corresponds to a functional channel (e.g., positive current terminal I+, positive voltage terminal V+, negative voltage terminal V-, negative current terminal I-). By controlling the corresponding matrix node switches, the instruments (current source, voltmeter) can be automatically connected to the test points of the target sample, achieving rapid and uninterrupted switching and testing of multiple samples.

[0036] In the specific implementation process, the M6×24 matrix switch shown in Figure 2 is used as the core of automated signal switching. Before testing, the magnetic device under test is connected sequentially to the 24 input ports of the matrix switch, and each input port is connected to the four-terminal test pins or multi-pin leads of the sample under test. The instruments (such as current sources and voltmeters) are connected to the 6 output ports of the matrix switch to achieve unified management of signal sources and measuring instruments. The establishment of the test task queue includes the following steps: First, according to the actual wiring of the sample under test, the connection relationship of each input port is numbered and recorded. According to the required test method (such as four-terminal measurement), the functional allocation relationship between each instrument output port and input port is clarified, and the corresponding switch closing combination is set in the queue form. The continuity of each path is tested one by one using a low-frequency signal. The contact validity is determined by measuring the circuit resistance. If the resistance of a certain path is lower than the preset threshold, it is recorded as a valid channel; otherwise, the cable or probe is adjusted until continuity is achieved. After all valid connection information is completed, test tasks are generated one by one, including sample number, input port number, output port number, required test parameters, and switch combination method. During batch testing, the matrix switches are sequentially controlled to close designated relays according to the test task queue, automatically connecting the instrument to the target sample. After testing one sample, the system automatically switches to the channel combination for the next sample to begin the next round of testing. Throughout the entire process, there is no need to manually change leads or move probes; all channel switching and testing procedures are automated by the software, greatly improving testing efficiency and batch data consistency.

[0037] In the implementation of this invention, to achieve high-throughput automated full-band testing of magnetic sensors, standardized interconnection and signal path configuration of various key instruments and equipment are required. The construction of the testing system not only includes core matrix relay switches and automated channel management, but also covers standardized connections between various devices such as high-frequency radio frequency signal sources, lock-in amplifiers, sample rods, and probe stations. Figure 4 shows a schematic diagram of the ST-FMR test circuit connection, including the FMR host (B300B), lock-in amplifier, Bias-T, coaxial cables L140 and B140B-1, and the top terminal of the sample rod. The radio frequency (RF) port of the FMR host B300B is connected to the RF port of the Bias-T via coaxial cable L140. The RF terminal of the Bias-T is directly connected to the top terminal of the sample rod for high-frequency excitation of the sample. In the electrical response signal generated after sample excitation, the DC or low-frequency component is separated from the DC terminal of the Bias-T and connected to the SIGNALINPUT A port of the lock-in amplifier via a coaxial cable B140B-1 to achieve the acquisition and detection of the DC or low-frequency signal. The SINE OUT output terminal of the lock-in amplifier is connected to the Input port of the FMR host via a signal line to achieve reference signal synchronization and system coordinated operation. Figure 5 further illustrates the principle block diagram of the Bias-T. The Bias-T is a circuit module that separates radio frequency (RF) signals and DC signals. The Bias-T module receives the RF signal through its input terminal. After separation, the DC signal is delivered to the sample through a separate path, while the RF signal continues to be transmitted along the RF path. This design ensures the independence of DC and RF signals, and allows a DC bias to be applied to the sample without affecting the accuracy of the RF signal. Figure 6 shows the appearance of the anti-static adapter box and its internal circuit structure. This anti-static adapter box has a multi-channel measurement interface, which can realize the safe connection of different instruments to the sample. The internal circuitry incorporates high-impedance current-limiting protection (such as a 10-megohm current-limiting resistor) and features floating / ground switching capabilities, effectively suppressing electrostatic discharge and external interference, thus improving signal purity and testing safety. All test signals pass through an adapter box before entering the matrix switch, further enhancing system stability and anti-interference capabilities during high-throughput, multi-sample automatic switching. Through the networking and signal path construction of the aforementioned testing system, end-to-end electrical continuity, signal quality, and safety are ensured, laying the foundation for efficient and stable operation of subsequent automatic switching tests, full-band data acquisition, and spin-torque ferromagnetic resonance analysis. The circuit design of this invention includes reserved interfaces or insertion points for RC, LC, and π-type filters, allowing for the selection of different types of filter units according to actual testing needs. This filters out high-frequency noise, electromagnetic interference, and other non-ideal signal components, improving the system's signal-to-noise ratio and ensuring the accuracy and stability of high-frequency and low-noise tests.

[0038] To avoid interference from transient signals generated during matrix switch channel switching on the formal test data, the following switching protection and window control process can be adopted: When the signal path needs to be switched from the current sample channel to the target sample channel, a command is first sent to the RF signal source to set the RF output to a silent state. The silent state can be achieved by reducing the RF output power to an extremely low value or by directly shutting down the RF source. Simultaneously, a command is sent to put the signal acquisition device into a hold state, pausing data acquisition or real-time updates of integration. After the above protection state is established, the maximum frequency among all frequencies covered by the current test task is obtained based on the full-band scanning range of the test task. Using this maximum frequency as the unique lookup key, the corresponding window supplementation time for that frequency range is obtained from a pre-set frequency band-window duration correspondence table. Generally, the higher the signal frequency, the shorter the wavelength of the signal in the transmission path, and the higher the requirements for signal path impedance matching and electrical parameter stability. High-frequency signals are more susceptible to factors such as contact instability, impedance abrupt changes, and parasitic parameter variations during relay switching, which can easily lead to signal reflection, energy loss, and even transient distortion, resulting in abnormal fluctuations or distortions in measurement data. To ensure that the high-frequency signal obtains sufficient stabilization recovery time within the path after matrix switch switching and to avoid the impact of switching transients on the accuracy of formal test data, the set window compensation time should be extended accordingly as the maximum frequency of the test task increases. In other words, the higher the maximum test frequency, the longer the required window compensation time should be to fully guarantee the stability of the signal path and the reliability of the data under high-frequency test conditions.

[0039] Furthermore, the hardware specifications of the matrix switch used were examined to obtain the time required for the relay's mechanical action, the maximum duration of relay contact bounce, and the time required for the signal path to achieve electrical parameter stabilization after channel switching. The maximum value among these three was taken as the base window duration for this switching operation. The time required for the relay's mechanical action refers to the time required for the relay to fully close or open its internal mechanical structure after receiving the switching command; this process involves the actual movement of mechanical components. The maximum duration of relay contact bounce refers to the multiple small bounces that may occur at the relay's contact points during mechanical switching due to elasticity and micro-friction, leading to brief signal discontinuities or instability; glitches or jumps may occur in the signal output during this stage. The time required for the signal path to achieve electrical parameter stabilization after channel switching refers to the self-adjustment and stabilization process that parameters such as impedance, capacitance, and inductance in the signal path need to undergo after the physical switching of the relay is completed. Only when these electrical parameters tend to stabilize can the complete transmission and acquisition of high-frequency signals be ensured. The reason for considering these three durations simultaneously is that the final stability of the signal path depends on the complete recovery of both the mechanical and electrical aspects. In actual handover, the three processes mentioned above may partially overlap or be connected in series, and the prolongation of any one of them may affect the overall signal stability. Therefore, in order to ensure that the signal path remains completely stable even under the worst-case scenario, the longest of the three processes is used as the basic window duration required for this handover, so as to fully cover all potential instability factors.

[0040] The required window duration for this channel switching is obtained by adding the basic window duration to the corresponding window supplement duration for the frequency band interval. Subsequently, a channel switching command is sent to the matrix switch to control the coaxial output port to switch from the current input port to the input port corresponding to the target sample. After the switching command is issued, the window duration timer is started. During the window duration, data writing operations from the acquisition device are blocked, prohibiting any signals acquired during this stage from being written into the formal test data, thereby preventing erroneous data caused by signal path instability from entering the main dataset. This ensures the accuracy and traceability of test data during high-frequency, full-band automated testing.

[0041] After the matrix switch channel switching and idle period, the output signal of the target sample channel needs to be tested for stability to determine whether the signal meets the requirements of subsequent formal testing. To this end, the following method can be used: First, based on a pre-set sampling window length and sampling frequency, the output signal is continuously acquired on the target sample channel to obtain voltage and current signal sequences over a period of time. For each set of acquired data, the mean and standard deviation of the voltage and the mean and standard deviation of the current are calculated to form characteristic parameters reflecting the stable state of the signal. Then, the voltage mean is subtracted from a preset voltage reference value in the system to obtain the voltage mean offset; similarly, the current mean is subtracted from the corresponding preset current reference value to obtain the current mean offset. The first stability criterion is that the voltage standard deviation is less than a preset voltage standard deviation threshold; the second stability criterion is that the current standard deviation is less than a preset current standard deviation threshold; the third stability criterion is that the absolute value of the voltage mean offset is less than a preset voltage mean offset threshold; and the fourth stability criterion is that the absolute value of the current mean offset is less than a preset current mean offset threshold. The stability characteristic parameters acquired in this test are compared sequentially with the four criteria mentioned above. If all criteria are met, the current signal path is considered stable. At this point, the RF output is de-silent, the hold state of the acquisition device is released, and the test proceeds to the subsequent formal testing phase. If any criterion is not met, the signal is considered unstable. Based on a pre-set waiting time, the waiting period is extended, and the signal is reacquired and the stability determination process is repeated until the signal is determined to be stable or the cumulative waiting time exceeds the preset timeout limit. Through the above operations, it is ensured that the signal path after each matrix switch can fully complete stability verification before entering the formal test, thereby effectively improving the data reliability and consistency of the fully automated batch testing process.

[0042] If the stability determination stops due to the accumulated waiting time exceeding the preset timeout limit, the current test task is marked, a prompt message is generated, and the time of the anomaly, task number, and determination basis are recorded in detail. The process then switches to the next test task and continues the subsequent automated testing process. This ensures that even with persistent instability in some channels, the overall high-throughput testing process can still be completed efficiently and continuously, and guarantees that all anomalies are traceable and can be subsequently reviewed.

[0043] Once the target sample channel passes the stability test, the formal testing process can begin. First, the system automatically controls the matrix switch according to the port mapping relationship set in the test task queue, connecting the coaxial output port to the input port corresponding to the current sample under test. Simultaneously, a recovery command is sent to the RF signal source to restore its output power to the preset test state, and the hold command of the detection and acquisition device is released, allowing it to resume normal data acquisition. Then, based on the pre-set parameters for this test, including the start and end frequencies of the sweep frequency, step interval, magnetic field range of the sweep field, and stepping mode, a parameter set covering all target test points is automatically generated. The output frequency of the RF signal source or the magnetic field of the detection device is adjusted point by point accordingly, switching sequentially to each test point. At each test point, according to the preheating sampling duration set in the test parameters, a preheating data segment is first collected for sufficient convergence and stabilization of the instrument state; this preheating data is not included in the final test result. After preheating, the system collects formal measurement data according to the formal sampling duration, writing the collected test voltage and current sequences as the main measurement data for that test point into the sampling test data for that point. For each test point, the formal segment sampling can be performed cyclically according to the set number of repeated samplings to obtain multiple independent test voltage and current sequences. All data are recorded under the corresponding sampled test data item. Following the above process, the test acquisition of all preset frequency points or magnetic field points is completed sequentially until the entire frequency band or magnetic field range involved in this test task is covered. Finally, all test data acquired under the current test task is output.

[0044] Frequency sweep and field sweep are two fundamental testing modes for comprehensively characterizing the performance of magnetic sensors. Frequency sweep refers to automatically controlling a radio frequency signal source to sequentially output a series of excitation signals at different frequencies under a given fixed magnetic field condition, acquiring the electrical response of the sample at each set frequency point. During this process, the magnetic field parameters remain constant, focusing on analyzing the sensor's resonant behavior or signal response changes under different microwave frequencies, thereby obtaining the sample's spectral response curve across the entire frequency range. Field sweep complements frequency sweep, referring to automatically adjusting the magnetic field source (such as an electromagnet or permanent magnet) under a constant microwave excitation frequency, causing the magnetic field to gradually change from its initial value in predetermined steps, acquiring electrical signals at each magnetic field point. Through field sweep, the ferromagnetic resonance characteristics, impedance changes, or spin-related effects of the sample under different magnetic field strengths can be characterized in detail. Common ferromagnetic resonance tests (such as ST-FMR) often require point-by-point scanning of the magnetic field at a fixed frequency to obtain accurate physical parameters such as resonant field displacement and linewidth. This invention allows for flexible execution of frequency sweep, field sweep, or a combination of both, depending on actual testing needs. In the test parameter settings, users can specify whether to perform a two-dimensional parameter scan using frequency sweep, field sweep, or both simultaneously. Based on the automatic control of the matrix switch and instrument switching, signal acquisition is completed one by one at each frequency and magnetic field combination point, enabling batch response measurements of target samples across the entire frequency and magnetic field range. Through these functions, efficient evaluation of the magnetic and electrical properties of multiple samples across the entire parameter space can be achieved, meeting diverse testing needs such as new material screening, device mass production, and fundamental physics research.

[0045] As shown in Figures 7 and 8, the low-temperature testing component involved in this embodiment is illustrated in the structural schematic diagram and a partial enlarged view of part A in the structural schematic diagram. From top to bottom, the component includes: a shock-absorbing spring 1 for providing mechanical damping, which isolates external vibrations from the low-temperature testing system and ensures testing accuracy. Below the shock-absorbing spring 1 is a Z-axis motion stage 2, which can precisely adjust the height of the sample or probe in the vertical direction to achieve accurate contact with the test object. Below the Z-axis motion stage 2 are the X-axis motion stage 3 and the Y-axis motion stage 4, which are used to achieve fine-tuning of the sample or probe in the horizontal direction (X, Y), respectively, to meet multi-point testing or positioning requirements. At the bottom is the probe part 5, which is used for precise contact with the electrode of the sample under test for signal input and output.

[0046] To ensure excellent high-frequency signal transmission capabilities under cryogenic and high magnetic field conditions, simulation and theoretical loss analysis were conducted on the microwave transmission components of the cryogenic sample rod at high frequencies (20~60GHz). Table 1 shows the high-frequency insertion loss statistics for each component of the cryogenic sample rod. The insertion loss parameters of key components such as the vacuum interface, flexible cable, adapter, semi-rigid cable, waveguide, and coplanar waveguide (CPW) were statistically analyzed at typical frequency bands. Based on the actual assembly quantity, the total loss of the entire rod at different frequency points was estimated (11.2dB at 20GHz, 17.3dB at 40GHz, and 22.0dB at 60GHz). Simulation results show that the CPW structure using a high-performance Rogers 4003 substrate can effectively extend the transmission frequency to above 70GHz, while controlling the overall insertion loss within -22dB, meeting the requirements for high-frequency and high-sensitivity measurements. As shown in Figure 9, the microwave simulation results further present the S-parameter curves of the cryogenic sample rod in the 80 GHz range, where S12 reflects the forward insertion loss and S11 reflects the return loss. The results show that the system possesses excellent high-frequency transmission characteristics within the target frequency band, providing a crucial foundation for high-frequency ST-FMR, FMR, and related magnetic measurements. In summary, by rationally optimizing the parameters and connection methods of each component, the cryogenic sample rod achieves low-loss transmission over an ultra-wide frequency band, improving the high-frequency signal integrity and measurement accuracy of the system under extreme physical environments, and laying a solid foundation for subsequent high-throughput automated testing of multiple samples.

[0047] Table 1. Statistics on insertion loss of various components in the low-temperature sample holder at high frequencies.

[0048]

[0049] Figure 10 shows a schematic diagram of a probe-type ST-FMR structure, illustrating the core local layout of a low-temperature, multi-dimensionally adjustable probe structure for spin-torque ferromagnetic resonance (ST-FMR) testing. It demonstrates the core structure and main components of the ST-FMR testing device. The external structure is a low-temperature frame 7, with an observation window / low-temperature objective / optical fiber optic cable 6 at its top for easy observation or optical measurement. The entire device is connected to external instruments via a low-temperature high-frequency cable 15 for high-frequency signal transmission. Multiple piezoelectric motors are installed internally, including an X-axis piezoelectric motor 12, a Y-axis piezoelectric motor 9, and a Z-axis piezoelectric motor 8, distributed at different locations within the device. These piezoelectric motors are used for sample positioning and clamping operations with micron-level precision. A high-frequency probe 10 is installed inside the device for applying and acquiring high-frequency signals to the micro / nano-fabricated STFMR device 14. The device sample is placed on a stage 13, which also integrates a rotatable piezoelectric motor 11, facilitating multi-angle rotation and adjustment of the sample. The entire system precisely adjusts the relative position of the probe and the sample using multiple sets of piezoelectric motors, achieving automated ST-FMR testing at low temperatures, high frequencies, and multiple degrees of freedom. Figure 11 shows the layout of the probe-type ST-FMR device array, illustrating the planar layout of an ST-FMR device array integrated on a micro / nano-fabricated chip. Each sub-unit is an independently designed ST-FMR testing device, capable of being tested individually or in batches using external high-frequency probes. The devices on the chip are arranged in a regular pattern, facilitating high-throughput automated testing. This array structure significantly improves testing efficiency, enabling fully automated batch scanning and data acquisition of multiple samples in a single setup, providing crucial support for material screening and large-scale parameter testing. Based on the aforementioned probe-type ST-FMR testing structure and device array, the testing system fully leverages its advantages in automation, batch processing, and high-precision measurement. By integrating a multi-channel high-frequency signal source, matrix switch, and automated acquisition system on the periphery of this structure, coupled with piezoelectric motor drive, precise positioning and signal on / off control of each sample are achieved. In actual testing, the high-frequency probe can be automatically switched to the corresponding ST-FMR device according to the task queue settings, realizing automated frequency and field sweep measurements for multiple samples and multiple frequency bands. Each test follows a preset procedure, including switching protection, signal stability determination, data acquisition, and quality inspection, ultimately outputting high-quality, traceable full-band ST-FMR test data, which can be further analyzed and physical parameter extracted.

[0050] For each test point under each test task, the quality of the test data must be assessed. Specifically, for multiple sets of repeated sampling data set for that test point, voltage and current sequences are obtained separately. First, for each set of test voltage and test current sequences, their standard deviation and effective sampling percentage are calculated. The effective sampling percentage refers to the proportion of sampled values ​​in the sequence that are within the instrument's range or a pre-defined effective interval. During the quality assessment process, if the standard deviation of a set of test voltage sequences is less than a preset voltage standard deviation threshold, and its effective sampling percentage is higher than a preset effective sampling percentage threshold, then that set of test voltage data is considered qualified. Similarly, for test current sequences, when its standard deviation is less than the current standard deviation threshold and its effective sampling percentage is higher than the corresponding threshold, then that set of test current data is considered qualified. If a set of test data satisfies both voltage and current data qualification, then the entire set of test data is recorded as a qualified set. After completing the above assessment, all sampling sets for the current test point are traversed, and the number of qualified sets and the total number of test sets are counted. The number of qualified sets divided by the total number of test sets is taken as the percentage of qualified test sets for that test point. Furthermore, all test points under the current test task are traversed, and the percentage of qualified test groups for each test point is calculated. If the percentage of qualified test groups for a test point is greater than a preset threshold, the test point is determined to be a qualified test point; otherwise, it is recorded as a unqualified test point. Finally, the total number of qualified test points in this test task is counted and combined with the total number of test points to obtain the percentage of qualified test points (the number of qualified test points divided by the total number of test points). Based on this percentage of qualified test points, the overall test quality label of the current test task is confirmed for subsequent data validity screening, retest determination, or automatic archiving operations.

[0051] If the percentage of qualified test points is less than the preset threshold, it indicates that the data quality of most test points in the current test task is substandard, with widespread anomalies, noise, or signal distortion. This results in an excessively low proportion of effective data for the entire test task, making it unsuitable for subsequent physical parameter analysis or result reliability. Therefore, the test quality label for the current test task is marked as severely unqualified. A procedural retest will be performed: First, the relevant module states will be reset, and the switching protection and window control, stability detection and judgment, and full-band testing and quality label confirmation processes will be executed again sequentially, re-collecting a complete set of test data for the entire test task. If the quality label for the current test task meets the qualification standard after the procedural retest, the original sampled test data will be replaced with all test point data obtained from this retest to form new confirmation test data. If the quality label remains severely unqualified or unqualified after the procedural retest, a prompt message will be generated, and the test task will be skipped to ensure that the overall high-throughput batch testing process is not interrupted by a single abnormal task.

[0052] If the percentage of qualified test points is greater than or equal to the preset threshold for qualified test point percentage but not equal to one, it indicates that the data quality of most test points in the current test task meets the requirements, with only some test points exhibiting anomalies or sporadic failures. The overall data has some reference value, but there is still a risk to the data integrity of some local points. Therefore, the test quality label for the current test task is recorded as unqualified. An automatic retest operation is triggered. For all locations in the current test task that are determined to be unqualified, data acquisition for the corresponding test points is re-executed according to the original test parameters, obtaining retest sampling test data for each unqualified test point. Subsequently, the quality of the retested data is re-evaluated. If all unqualified test points meet the standard for qualified test points after retesting, the sampling test data for the corresponding unqualified test points in the original test task is replaced with the retest data, thus forming the final confirmed test data for the current test task. If there are still test points that fail to meet the qualified standard after retesting, relevant prompts are generated, and the test task is automatically skipped, proceeding to the next test task.

[0053] If the percentage of qualified test points is equal to one, it means that all test points of the current test task have passed the quality judgment, and the test data of each group have reached the preset standards in terms of volatility and sampling integrity. The overall test results are highly reliable and there is no need for retesting or resampling. Therefore, the test quality label of the current test task is recorded as qualified, and the sampled test data of each test point of the current test task is directly used as the confirmation test data of the current test task.

[0054] In this embodiment, a mechanism for triggering retests due to quality non-compliance is introduced primarily to ensure the reliability of test data, eliminate occasional interference factors, and improve overall testing efficiency. When a test quality label indicates non-compliance, it typically means that only a few test points failed to meet the preset quality standard, while most test point data performed normally. In this case, retesting the non-compliant test points allows for the rapid location and correction of local anomalies caused by occasional noise, short-term instrument drift, or environmental disturbances, maximizing the preservation of the valid portion of the original test data and avoiding resource waste caused by full-task retesting. If, after point retesting, all non-compliant test points meet the quality standard, it indicates that the anomaly is occasional and has been effectively eliminated. Replacing the retested data with the corresponding test points then forms confirmatory test data, ensuring the integrity and accuracy of the overall test results. If test points still fail after point retesting, it indicates a more serious or systemic anomaly in the test task, which is difficult to repair through simple re-collection. In this case, a prompt message is generated, and the test task is skipped, avoiding ineffective retries and wasting system resources. For test quality labels indicating severe non-compliance, this often manifests as most test point data failing to meet standards. In such cases, a process-wide retest is required. This involves re-executing the switching protection, window control, stability assessment, and the entire testing and quality verification process, essentially restarting the automated testing workflow for that task to eliminate potential systematic errors and recalibrate each step. If the retest yields satisfactory results, the original data is replaced to form the verification test data. If the retest still fails to meet the quality standards, a warning message is generated, and the test task is skipped to prevent the system from entering an invalid loop.

[0055] After completing high-throughput full-band acquisition for each test task, the voltage and current sequences obtained at each test point need to be processed. Typically, for each frequency or magnetic field point, the voltage and current signals obtained from multiple repeated samplings at that test point are averaged or medianized to reduce the influence of occasional noise. Taking frequency scanning mode as an example, the average voltage values ​​at all frequency points can be arranged in ascending order of frequency to form the voltage-frequency response curve of the sample; if magnetic field scanning is used, a voltage-magnetic field response curve is obtained. This response curve can comprehensively reflect the changes in the electrical output of the sample under different external excitation conditions. The sliding window method is used to calculate the first derivative and find the interval with the largest rate of change in the curve. A typical approach is to calculate the average slope within each window (e.g., 5-10 points) and find the point where the absolute value of the slope is maximized, thus initially determining its resonance center; then, using the maximum point as the center, the boundary points where the slope changes sign are found on both sides, serving as the boundaries of the resonance interval. For scenarios with low signal-to-noise ratios or atypical resonance patterns, the rate of change threshold and window width can be set based on engineer experience, supporting semi-automatic adjustment to obtain a reasonable resonance range. Within the identified resonance range, a Lorentzian or a mixture of symmetric and antisymmetric Lorentzian models is used to perform nonlinear fitting of the resonance signal. The specific fitting function can be: In this model, V(H) represents the response of the mixed signal obtained from the measurement to the magnetic field H, S represents the coefficient of the symmetric term (corresponding to the spin pumping effect component), A represents the coefficient of the antisymmetric term (corresponding to the AMR component), k represents the coefficient of the linear term, H represents the magnetic field, H0 represents the magnetic field value of the resonant field, ΔH represents the resonant linewidth (half-width at half maximum), and C represents the constant term. The aforementioned constructed magnetic device response curve and fitting function are input into the calculation process of the fitting algorithm. It is recommended to use the commonly used nonlinear least squares method for curve fitting; in this embodiment, the Levenberg-Marquardt algorithm is preferred. Such algorithms are readily available in mainstream scientific computing environments such as MATLAB and Python. During the fitting process, the magnetic field (or frequency) of the response curve is used as the independent variable, the actually measured voltage or current signal is used as the dependent variable, and the resonant field, resonant linewidth, coefficient of the symmetric term, coefficient of the antisymmetric term, coefficient of the linear term, and constant term in the fitting function are used as variables to be optimized. Subsequently, the fitting algorithm automatically adjusts the values ​​of these parameters to minimize the error (e.g., root mean square error) between the fitting function and the actual measured data. After the fitting converges, the optimal estimates of each physical parameter can be obtained: the resonance field represents the magnetic field value corresponding to the resonance peak of the fitted curve, directly reflecting the ferromagnetic resonance condition of the sample; the resonance linewidth reflects the width of the resonance peak; the coefficient of the symmetry term reflects the intensity of the spin pumping effect or the spin current correlation response; the coefficient of the antisymmetry term reflects the contribution of effects such as anisotropic magnetoresistance (AMR); and the linear term and constant term are used to compensate for DC drift, background noise, or signal baseline changes in the measurement system. After completing the above parameter fitting, all key physical parameters are organized according to the test task number, and the structured output is used as the analysis results, which are stored one-to-one with the original acquisition data of the test task. The analysis results can be used for subsequent research on material physical properties, magnetic parameter calibration, and batch data statistics and comparative analysis, significantly improving the efficiency and reliability of physical quantity extraction under automated high-throughput testing.

[0056] Figure 12 shows typical data and fitting results of ST-FMR testing at a fixed frequency. This figure illustrates typical experimental data and fitting results obtained by testing magnetic devices using the ST-FMR method under fixed frequency (8 GHz) and temperature (300 K) conditions. In the figure, green dots and squares represent experimental measurement data for different samples, with the horizontal axis representing the applied magnetic field strength (kOe) and the vertical axis representing the corresponding voltage signal (μV). It can be seen that the voltage signal exhibits a clear resonant response curve with changes in the magnetic field, containing both symmetric and antisymmetric components. By performing nonlinear curve fitting on the experimental data (the solid black line in the figure), key parameters such as the signal's resonance position, linewidth, and component amplitude can be accurately characterized. The fitted curve closely matches the experimental points, indicating that the physical model used can effectively describe the magnetodynamic response of the sample under ST-FMR conditions.

[0057] It should be noted that since there are many physical effects that generate ST-FMR signals, such as anisotropic magnetoresistance and spin Hall magnetoresistance, and these signals are often mixed with spin pump and inverse spin Hall voltage signals, ST-FMR measurement also requires rotation angle testing or comparative testing of multiple samples. Furthermore, various theories and methods for ST-FMR data analysis have been developed in recent years, and this invention will not introduce further data analysis.

[0058] In Embodiment 2 of this invention, based on Embodiment 1, after determining instability, the waiting time can be determined not only using a preset value but also through the following steps: Dividing the voltage standard deviation by a voltage standard deviation threshold yields a voltage standard deviation deviation coefficient; dividing the current standard deviation by a current standard deviation threshold yields a current standard deviation deviation coefficient; dividing the absolute value of the voltage mean deviation by a voltage mean deviation threshold yields an absolute voltage mean deviation deviation coefficient; dividing the absolute value of the current mean deviation by a current mean deviation threshold yields an absolute current mean deviation deviation coefficient; setting all deviation coefficients less than 1 among the voltage standard deviation deviation coefficient, current standard deviation deviation coefficient, absolute voltage mean deviation deviation coefficient, and absolute current mean deviation deviation coefficient to zero, then calculating the average of the non-zero values, which is recorded as the stability deviation coefficient. Using the stability deviation coefficient as the lookup key, the corresponding waiting time is retrieved from a preset stability deviation coefficient-waiting time mapping table. A larger stability deviation coefficient indicates a larger remaining margin in the distance criterion threshold of the acquired signal characteristics, meaning there is still considerable room for determining the signal distance as stable. Since we are far from a stable state at this point, a longer waiting time is needed to ensure that the signal path can fully recover to stability. Conversely, if the stability deviation coefficient is small, it means that the current acquired feature is close to the criterion threshold and is close to a stable state, requiring only a shorter waiting time.

[0059] In Embodiment 3 of this invention, based on Embodiments 1 or 2, it can also be used for ferromagnetic resonance (FMR) testing. The ferromagnetic resonance absorption information of the sample is obtained by detecting the microwave output intensity using a diode detector. Combined with AC modulated magnetic field and lock-in amplifier measurements, the differential spectrum of the sample's ferromagnetic resonance absorption intensity with respect to the magnetic field can be directly obtained. The fitting function used is the same as that in Embodiment 1. Typical data and fitting results of the FMR differential spectrum at a fixed frequency are shown in Figure 13. In the figure, the blue hollow circle represents the relationship between the experimentally measured detection voltage and the applied magnetic field (in kOe), and the red solid line is the curve processed using the fitting function. It can be seen that the experimental data exhibits a pair of distinct positive and negative peaks in the resonance region, corresponding to the differential characteristics of the ferromagnetic resonance absorption signal. The fitting curve can well describe the trend of the experimental data. By fitting the differential spectrum data, key physical parameters such as the resonance field and resonance linewidth of the sample can be accurately extracted, providing a reliable experimental basis for further analysis of the material's magnetodynamic properties and damping mechanism. This figure demonstrates the application effect of the testing system of this invention in ferromagnetic resonance absorption spectrum testing and quantitative analysis of physical parameters.

[0060] In general, FMR testing needs to be conducted at multiple frequencies to obtain more information. This example uses a typical in-plane anisotropic thin film. By measuring the ferromagnetic resonance spectrum of the sample at different frequencies, we can obtain the relationship between the resonant magnetic field and the frequency. Fitting this spectrum using the Kittel equation, we can obtain the gyromagnetic ratio of the material. Further linear fitting and analysis of the frequency and resonant linewidth yields the Gilbert damping factor of the material. Figure 14 shows the resonant field-frequency dispersion curve, describing the relationship between the ferromagnetic resonance frequency and the applied magnetic field. The horizontal axis represents the resonant magnetic field (unit: kOe), and the vertical axis represents the frequency (unit: GHz). The blue dots represent the experimentally measured resonant magnetic field data points at various frequencies, and the red solid line is the theoretical fitting curve obtained from the fitting function. As can be seen from the figure, the experimental data and the theoretical fitting curve are in high agreement, indicating that the selected fitting model can accurately describe the variation of the ferromagnetic resonance frequency with the resonant magnetic field. By fitting this relationship, key physical parameters such as the effective magnetization of the material can be further extracted, providing a basis for subsequent analysis of dynamic parameters such as the Gilbert damping factor. The fitting function is as follows: Where f is the resonant frequency in GHz. The gyromagnetic ratio (an intrinsic parameter of magnetic materials), measured in radians / ( ), Vacuum permeability ( H0 is the resonant magnetic field, with units of kOe (kilo-Oersted). M effEffective magnetization (an intrinsic parameter of the material), measured in emu / cm. 3 (Electromagnetic unit / cubic centimeter), 4 M eff It is often presented in the form of an equivalent magnetic field, with the unit corresponding to Oe.

[0061] Figure 15 shows the frequency-linewidth curve, describing the relationship between the ferromagnetic resonance linewidth and frequency. The resonance linewidth reflects the degree of resonance loss. The horizontal axis represents the resonance frequency (in GHz), and the vertical axis represents the corresponding linewidth (in Oe, Ørsted). Experimental data are represented by blue dots, and the red solid line represents the linear fitting result based on the fitting function. It can be seen that the experimental data points are basically distributed along the fitted straight line, indicating a good linear relationship between the linewidth and frequency. The slope obtained through fitting directly reflects the Gilbert damping factor (…). ), as shown in the figure The value is 0.007. This linear relationship not only reveals the magnetic loss characteristics of the material, but also provides a reliable basis for subsequent extraction of damping parameters and evaluation of ferromagnetic dynamic performance. The fitting function is: Where ΔH is the resonance linewidth, and the unit is Oe (Oersted). Gilbert damping factor (a material loss parameter, marked in the figure) =0.007), dimensionless, f is the resonant frequency in GHz, ΔH0: intrinsic linewidth (linewidth at zero frequency, inherent loss of the material), in Oe.

[0062] It should be emphasized that this example is only a simple reference for data processing. In actual situations, due to the different anisotropies of the tested samples, the applicable Kittel equation needs to be analyzed in combination with specific requirements.

[0063] In Embodiment 4 of this invention, based on Embodiments 1, 2, or 3, it can also be used to test and analyze spin-pumped and inverse spin Hall effect (ISHE) signals. For different types of testing devices and magnetic device structures, either an FMR sample bar or an ST-FMR sample bar can be selected for measurement according to actual needs. For example, when using an FMR sample bar, the thin film sample to be tested can be attached to a coplanar waveguide (CPW) structure, and the ISHE voltage signal can be acquired in real time by scanning with an external magnetic field to analyze the spin injection and detection characteristics of the ferromagnetic-nonmagnetic interface. When using an ST-FMR sample bar, the CPW can be integrated with the sample through micro-nano fabrication to achieve high-precision measurement of various spin dynamic parameters of microstructured samples. During the above ISHE testing process, the corresponding physical model and fitting algorithm will be automatically called according to the sample type, wiring method, and preset parameters of each test task to uniformly process the data collected from each test point in batches, output and store the analysis results of physical parameters such as ISHE signal intensity and spin injection efficiency, achieving seamless integration with the ST-FMR analysis process.

[0064] The various features and processes described above can be used independently of each other or can be combined in various ways. All possible combinations and sub-combinations are intended to fall within the scope of this disclosure. Furthermore, certain method or process blocks may be omitted in some embodiments. The methods and processes described herein are not limited to any particular order, and the blocks or states associated with them may be performed in other suitable orders. For example, the described blocks or states may be performed in an order different from the order specifically disclosed, or multiple blocks or states may be combined in a single block or state. Example blocks or states may be performed serially, in parallel, or in some other manner. Blocks or states may be added to or removed from the disclosed example embodiments. The exemplary systems and components described herein may be configured differently from those described. For example, elements may be added to, removed from, or rearranged compared to the disclosed example embodiments.

[0065] In this specification, multiple instances may implement components, operations, or structures described as single instances. Although individual operations of one or more methods are shown and described as separate operations, one or more of the separate operations may be performed simultaneously and do not need to be performed in the order shown. Structures and functions presented as separate components in the example configuration may be implemented as composite structures or components. Similarly, structures and functions presented as single components may be implemented as separate components. These and other variations, modifications, additions, and improvements fall within the scope of this document.

Claims

1. A high-throughput probe-type magnetic sensor full-band testing system based on matrix switches, used to perform frequency sweep and / or field sweep tests on multiple magnetic devices under test within a preset frequency band to conduct spin torque ferromagnetic resonance tests, characterized in that, include: The task queue management module is used to establish a test task queue. The test task queue includes at least the sample number, probe position, test parameter settings, and port mapping relationship between the matrix switch input port and the coaxial output port. The task queue is executed one by one while maintaining probe contact. The switching protection and window control module is used to control the RF output to enter a silent state and the detection acquisition to enter a hold state before the matrix switch switches to the target sample channel for each task. The module determines the hold duration of the window period associated with the current test frequency band and issues a matrix switch switching command. After switching, the task enters the window period, and the acquisition results are prohibited from being written into the formal test data during the window period. The stability detection and judgment module is used to perform stability detection and acquisition on the target sample channel after the window period ends to determine stability. When the stability is determined to be stable, the RF output is restored and the hold state is released to enter the formal test. When the stability is determined to be unstable, the waiting period is extended and the stability detection and acquisition and stability judgment are repeated until stability is achieved or a timeout occurs. The full-band test and quality control module is used to perform frequency sweep and / or field sweep acquisition on the current test task during the formal test. It performs preheating segment sampling and formal segment sampling on each test point, outputs the sampled test data of each test point, determines the quality label of the current test task, and triggers retest when the quality label does not meet the requirements and replaces the corresponding test data with the retest result to form the confirmation test data of the current test task. The spin torque ferromagnetic resonance analysis module is used to perform spin torque ferromagnetic resonance analysis on the confirmation test data of each test task in the test task queue, output the analysis results of each test task, and associate and store the analysis results of each test task with the confirmation test data.

2. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 1, characterized in that: The switching protection and idle window control module includes: upon receiving a switching request for the target sample channel, before the matrix switch performs the switching, sending a control command to the radio frequency signal source to silence the radio frequency output and simultaneously controlling the detection and acquisition device to enter a hold state to pause the current acquisition or integration update; determining the idle window hold duration based on the frequency band interval corresponding to the current test task; after determining the idle window hold duration, issuing a matrix switch switching command to switch the coaxial output port of the matrix switch from the current input port to the input port corresponding to the target sample, and entering the idle window period after the switching command is issued; during the idle window period, blocking the data writing operation of the detection and acquisition device output, and prohibiting any signals acquired during the idle window period from being written into the formal test data, so as to avoid the influence of the matrix switch switching transient on the formal test results.

3. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 2, characterized in that: The process of determining the window duration based on the frequency band interval corresponding to the current test task specifically includes: obtaining the highest frequency within the covered test frequency band corresponding to the current test task, using this as a unique query key, and querying the window supplementation duration from the preset frequency band-window duration table; obtaining the time required for relay mechanical action, contact bounce time, and electrical path transition stabilization time, and taking the maximum value of the three as the basic window duration; and obtaining the window duration based on the basic window duration and the window supplementation duration.

4. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 1, characterized in that: The stability detection and determination module specifically includes: after the end of the idle period, continuously acquiring the output signal of the target sample channel with a preset sampling window and sampling frequency to obtain stability characteristic parameters reflecting the stable state of the signal, and thus determining the stability; when the stability determination result is stable, controlling the RF output to return to the normal test state and releasing the hold state of the detection and acquisition device, and entering the formal test stage; when the stability determination result is unstable, determining to extend the waiting time, and re-performing stability detection and acquisition after the extended waiting time to determine the stability; if the stability determination result is still unstable, repeating the extension of waiting and stability determination until the stability determination result is stable or the cumulative extension time exceeds the preset timeout limit.

5. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 4, characterized in that: The stability determination process involves the following steps: continuously acquiring the output signal of the target sample channel using a preset sampling window and sampling frequency to obtain a voltage sequence and a current sequence; calculating stability characteristic parameters including the voltage mean, voltage standard deviation, current mean, and current standard deviation; and performing difference processing between the voltage mean and current mean and their corresponding preset reference values ​​to obtain the voltage mean offset and current mean offset. The voltage standard deviation being less than a preset voltage standard deviation threshold is used as the first stability criterion. The standard deviation of the current is less than the preset standard deviation threshold of the current, which is used as the second stability criterion. The absolute value of the voltage mean offset is less than a preset voltage mean offset threshold, which is used as the third stability criterion. The absolute value of the current mean offset is less than the preset current mean offset threshold as the fourth stability criterion; the first stability criterion, the second stability criterion, the third stability criterion and the fourth stability criterion are combined as the stability criterion set; if there is no stability detection parameter that does not satisfy the stability criterion set, the stability determination result is recorded as stable; if there is any stability detection parameter that does not satisfy the stability criterion set, the stability determination result is recorded as unstable.

6. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 1, characterized in that: The full-band testing and quality control module specifically includes: after the stability determination result is stable, according to the port mapping relationship of the test task queue, automatically connecting the coaxial output port to the input port corresponding to the current sample under test through a matrix switch, controlling the RF output to return to normal, and releasing the hold state of the detection and acquisition device to enter the formal testing process; automatically generating the frequency sweep parameter group or field sweep parameter group for this test according to the preset test parameter settings, and determining all frequency points or magnetic field points covered by this test; for each frequency point or magnetic field point, first setting the RF signal source and detection device according to the test parameters, switching to the target frequency band or magnetic field of the current test point, and preparing for sampling; at each test point, executing the sampling based on the preheating period in the test parameter settings. The preheating sampling process involves collecting a segment of preheating data. This preheating data is not included in the formal measurement results and is only used to ensure the instrument's state fully converges. After preheating, formal sampling is performed based on the formal sampling duration set in the test parameters. The resulting test voltage and test current sequences are used as the main measurement data for that test point and written into the sampled test data for that test point. For the same test point, formal sampling is performed cyclically according to the number of repetitions set in the test parameters to obtain multiple sets of test voltage and test current sequences, all of which are written into the sampled test data for that test point. The above test steps are then performed sequentially on all preset frequency points or magnetic field points until all frequency bands or magnetic field ranges required for the current task are covered. Finally, the sampled test data for the current test task is output.

7. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 1, characterized in that: The full-band testing and quality control module further includes: for a test point in the current testing task, based on each set of test voltage sequences and each set of test current sequences at that test point, calculating the standard deviation of each voltage, the effective sampling percentage of each voltage, the standard deviation of each current, and the effective sampling percentage of each current, respectively; if the standard deviation of a certain test voltage sequence is less than a preset voltage standard deviation threshold and the effective sampling percentage of the voltage is greater than a preset effective sampling percentage threshold, then the test voltage data of that set is deemed qualified; if the standard deviation of a certain test current sequence is less than a preset current standard deviation threshold and the effective sampling percentage of the current is greater than a preset effective sampling percentage threshold, then the test current data of that set is deemed qualified; if a certain set of test voltages simultaneously exists... If both the data and the test current data are qualified, the test data for that group is deemed qualified. The test voltage and current sequences for each test point are iterated through to obtain the number of qualified test groups, and the percentage of qualified test groups for that test point is calculated. For the current test task, each test point is iterated through to obtain the percentage of qualified test groups for each test point. Test points with a percentage greater than the threshold for the percentage of qualified test groups are recorded as qualified test points, while those with a percentage less than or equal to the threshold are recorded as unqualified test points. The number of qualified test points is counted, and the percentage of qualified test points is calculated based on the total number of test points. The quality label for the current test task is confirmed based on the percentage of qualified test points.

8. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 7, characterized in that: The method of determining the quality label of the current test task based on the percentage of qualified test points specifically includes: if the percentage of qualified test points is less than a preset threshold for the percentage of qualified test points, the test quality label of the current test task is recorded as seriously unqualified; if the percentage of qualified test points is greater than or equal to the preset threshold for the percentage of qualified test points but not equal to one, the test quality label of the current test task is recorded as unqualified; if the percentage of qualified test points is equal to one, the test quality label of the current test task is recorded as qualified, and the sampled test data of each test point of the current test task is directly used as the confirmation test data of the current test task.

9. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 1, characterized in that: The full-band testing and quality control module further includes: triggering point retesting when the test quality label is unqualified; resampling the corresponding test points for all unqualified test points in the current test task to obtain retest sampling test data for each unqualified test point; re-evaluating the pass / fail status of each unqualified test point; if all unqualified test points are qualified after retesting, replacing the sampling test data of the unqualified test points in the current test task with the retest sampling test data of the corresponding unqualified test points to form the confirmation test data for the current test task; if there are still unqualified test points that need to be retested, generating a prompt message. The system will skip the current test task. If the test quality label is "severely unqualified," a procedural retest will be triggered, resetting all states and re-executing the switching protection and window control module, the stability detection and judgment module, and the test and quality label confirmation process of this module. If the quality label of the current test task is qualified after the retest, the original sampled test data of each test point will be replaced with the sampled test data of each test point obtained after the retest, thus forming the confirmation test data of the current test task. If the quality label of the current test task is still "severely unqualified" or "unqualified" after the retest, a prompt message will be generated and the current test task will be skipped.

10. The high-throughput probe-type magnetic sensor full-band testing system based on matrix switches as described in claim 1, characterized in that: The spin-torque ferromagnetic resonance analysis module includes: constructing signal response curves for each test task under different frequencies or magnetic field conditions based on the voltage and current sequences of each test point in the test task queue; identifying resonance intervals and extracting resonance signals based on the signal response curves; using a curve fitting algorithm to fit the parameters of the resonance signals to obtain the spin-torque ferromagnetic resonance physical parameters of the corresponding test task; packaging the physical parameters of each test task into the analysis results of each test task and storing them in association with the corresponding confirmed test data.

Images