An adaptive crosstalk calibration method and system for differential high-speed signals

Abstract

The application provides a kind of adaptive crosstalk calibration method and system of differential high-speed signal, it is related to signal test technical field, method includes: each differential test channel is calibrated, and error model is generated;The crosstalk coupling coefficient between each differential test channel is measured;The differential transmission parameter of each measured differential channel of the sample to be measured is measured, and the differential transmission parameter is corrected in combination with error model and crosstalk coupling coefficient;Periodically insert crosstalk calibration frame, in crosstalk calibration frame, the target channel response of differential crosstalk calibration piece and crosstalk source channel response are measured;In combination with crosstalk source channel response and the theoretical target channel response of differential crosstalk calibration piece, and the error between target channel response and theoretical target channel response, crosstalk coupling coefficient is updated;The differential transmission parameter of subsequent measurement is corrected by updated crosstalk coupling coefficient.The application has the technical effect that: the accuracy of test result is improved.

Description

Technical Field

[0001] This application relates to the field of signal testing technology, specifically to an adaptive crosstalk calibration method and system for differential high-speed signals. Background Technology

[0002] With the rapid development of high-speed digital communication technology, differential signal transmission has been widely used in high-speed serial buses (such as USB, PCIe, HDMI, etc.). Differential signals have advantages such as strong common-mode noise immunity and low electromagnetic radiation, but they are susceptible to crosstalk in high-frequency transmission. Crosstalk can lead to a decrease in signal integrity, affecting measurement accuracy and communication quality. Therefore, accurately measuring and characterizing the transmission parameters of differential high-speed signals and effectively eliminating the effects of crosstalk has become a pressing technical problem to be solved in the field of high-speed signal testing.

[0003] Currently, vector network analyzers are typically used to measure the S-parameters of differential devices, and differential transmission parameters are obtained through mixed-mode S-parameter conversion. To improve measurement accuracy, standard calibration kits (such as SOLT calibrators) are used to calibrate the test channels before testing, establishing an error model to eliminate the systemic errors inherent in the test system itself. Simultaneously, some solutions measure the crosstalk coupling coefficient between channels once before testing and use a fixed crosstalk coupling coefficient to compensate for crosstalk in subsequent tests. While these methods can effectively compensate for crosstalk errors in the initial testing phase, during long-term continuous testing, the operating state of the test system is affected by various factors, causing time-varying drift in the crosstalk coupling characteristics between channels. A fixed crosstalk coupling coefficient cannot reflect this dynamic change, causing the effectiveness of crosstalk compensation to gradually decrease over time, ultimately affecting the accuracy of the test results. Summary of the Invention

[0004] This application provides an adaptive crosstalk calibration method and system for differential high-speed signals to improve the accuracy of test results.

[0005] In a first aspect, this application provides an adaptive crosstalk calibration method for differential high-speed signals, applied to a host computer. The host computer is connected to a vector network analyzer and a switch matrix for communication. The test port of the vector network analyzer is connected to the positive and negative signal lines of multiple differential test channels through the switch matrix. The method includes: The control switch matrix is ​​sequentially switched to each differential test channel to calibrate each differential test channel, generate an error model for each differential test channel, and measure the crosstalk coupling coefficient between each differential test channel. When entering the test mode, the control switch matrix sequentially switches to each measured differential channel of the sample under test. The differential transmission parameters of each measured differential channel are measured by the vector network analyzer. The differential transmission parameters are corrected by combining the error model and crosstalk coupling coefficient to generate the target differential transmission parameters. In the test mode, crosstalk calibration frames are periodically inserted. Within the crosstalk calibration frames, the control switch matrix is ​​switched to the differential crosstalk calibrator. The target channel response and crosstalk source channel response of the differential crosstalk calibrator are measured by a vector network analyzer. Using the crosstalk source channel response as the input signal and the theoretical target channel response of the differential crosstalk calibrator as the desired signal, the crosstalk coupling coefficient is updated based on the error between the target channel response and the theoretical target channel response, and the target crosstalk coupling coefficient is generated. The differential transmission parameters of subsequent measurements are corrected by using the target crosstalk coupling coefficient.

[0006] By adopting the above technical solution, an initial calibration benchmark for the test system is established by calibrating each differential test channel to generate an error model and measuring the crosstalk coupling coefficient in the initial stage of testing. Crosstalk calibration frames are periodically inserted in the test mode. The current target channel response and crosstalk source channel response are measured by the differential crosstalk calibration device. The crosstalk coupling coefficient is updated based on the actual response error using the crosstalk source channel response and the theoretical target channel response as a benchmark to generate the target crosstalk coupling coefficient. This allows the crosstalk coupling coefficient to dynamically track the time-varying crosstalk characteristics of the test system caused by environmental changes and device drift. The target crosstalk coupling coefficient is used to correct subsequent differential transmission parameters, ensuring that crosstalk compensation is always based on the current real state of the system. This overcomes the problem of long-term test accuracy degradation caused by the inability of a fixed crosstalk coupling coefficient to reflect time-varying characteristics, and significantly improves the accuracy of test results.

[0007] Secondly, this application provides an adaptive crosstalk calibration system for differential high-speed signals, the system comprising: a control module, a test module, an insertion module, an input module, and a calibration module; wherein, The control module is used to control the switch matrix to switch sequentially to each differential test channel, calibrate each differential test channel, generate an error model for each differential test channel, and measure the crosstalk coupling coefficient between each differential test channel. The test module is used to control the switch matrix to switch to each differential channel of the sample under test in sequence when entering the test mode, measure the differential transmission parameters of each differential channel under test by a vector network analyzer, and correct the differential transmission parameters by combining the error model and crosstalk coupling coefficient to generate target differential transmission parameters. The insertion module is used to periodically insert crosstalk calibration frames in the test mode. Within the crosstalk calibration frame, the control switch matrix is ​​switched to the differential crosstalk calibration device, and the target channel response and crosstalk source channel response of the differential crosstalk calibration device are measured by the vector network analyzer. The input module is used to take the crosstalk source channel response as the input signal, the theoretical target channel response of the differential crosstalk calibrator as the expected signal, update the crosstalk coupling coefficient according to the error between the target channel response and the theoretical target channel response, and generate the target crosstalk coupling coefficient. The correction module is used to correct the differential transmission parameters of subsequent measurements using the target crosstalk coupling coefficient.

[0008] Thirdly, this application provides an electronic device that adopts the following technical solution: including a processor, a memory, a user interface, and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to enable the electronic device to execute a computer program of any of the above-described adaptive crosstalk calibration methods for differential high-speed signals.

[0009] Fourthly, this application provides a computer-readable storage medium that employs the following technical solution: storing a computer program capable of being loaded by a processor and executing any of the above-mentioned adaptive crosstalk calibration methods for differential high-speed signals.

[0010] In summary, this application includes at least one of the following beneficial technical effects: By calibrating each differential test channel to generate an error model and measuring the crosstalk coupling coefficient in the initial testing phase, an initial calibration benchmark for the test system is established. Crosstalk calibration frames are periodically inserted in the test mode. The current target channel response and crosstalk source channel response are measured using differential crosstalk calibration components. Based on the crosstalk source channel response and the theoretical target channel response, the crosstalk coupling coefficient is updated according to the actual response error to generate a target crosstalk coupling coefficient. This allows the crosstalk coupling coefficient to dynamically track the time-varying crosstalk characteristics of the test system caused by environmental changes and device drift. The target crosstalk coupling coefficient is used to correct subsequent differential transmission parameters, ensuring that crosstalk compensation is always based on the current true state of the system. This overcomes the problem of long-term test accuracy degradation caused by the inability of a fixed crosstalk coupling coefficient to reflect time-varying characteristics, and significantly improves the accuracy of test results. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating an adaptive crosstalk calibration method for differential high-speed signals provided in an embodiment of this application. Figure 2 This is a flowchart of adaptive crosstalk calibration in a differential high-speed signal testing system provided in this embodiment; Figure 3 This is a hardware connection architecture diagram of a differential high-speed signal testing system provided in an embodiment of this application; Figure 4 This is a schematic diagram of a crosstalk propagation mechanism in a differential signal multichannel system provided in an embodiment of this application; Figure 5 This is a flowchart illustrating differential test channel calibration and crosstalk matrix establishment provided in an embodiment of this application; Figure 6 This is a complete test flowchart of adaptive crosstalk calibration for differential high-speed signals provided in an embodiment of this application; Figure 7 This is a detailed flowchart of an embodiment of the present application providing an method for updating crosstalk coupling coefficients based on an adaptive filtering algorithm; Figure 8 This is a schematic diagram of an adaptive crosstalk calibration time axis provided in an embodiment of this application; Figure 9 This is a schematic diagram of the rack deployment of a differential high-speed signal testing system provided in an embodiment of this application; Figure 10 This is a software architecture diagram of a differential high-speed signal adaptive crosstalk calibration and testing system provided in an embodiment of this application; Figure 11 This is a schematic diagram of the structure of an adaptive crosstalk calibration system for differential high-speed signals provided in an embodiment of this application; Figure 12 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0012] Explanation of reference numerals in the attached figures: 1000, electronic device; 1001, processor; 1002, communication bus; 1003, user interface; 1004, network interface; 1005, memory. Detailed Implementation

[0013] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0014] In the description of the embodiments in this application, words such as "illustrative," "for example," or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "illustrative," "for example," or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or designs. Rather, the use of words such as "illustrative," "for example," or "for example" is intended to present the relevant concepts in a specific manner.

[0015] Figure 1 This is a schematic flowchart of an adaptive crosstalk calibration method for differential high-speed signals provided in an embodiment of this application. Figure 1 As shown, the method includes S101-S105: S101, the control switch matrix is ​​switched sequentially to each differential test channel to calibrate each differential test channel, generate the error model of each differential test channel, and measure the crosstalk coupling coefficient between each differential test channel.

[0016] First, the test channels need to be calibrated and crosstalk characteristics measured. This is because test paths such as switch matrices and test cables introduce transmission errors and inter-channel crosstalk. If calibration and crosstalk measurement are not performed, the subsequent measurement results of the sample under test will include these systematic errors and will not reflect the true performance of the sample under test.

[0017] In practice, the host computer sends a switching command to the switch matrix via the communication interface, controlling the switch matrix to connect the first test port of the vector network analyzer to the positive signal line of the first differential test channel, and the second test port to the negative signal line of the first differential test channel. The positive and negative signal lines are two independent transmission lines constituting the differential signal pair, with the positive signal line transmitting the positive-phase signal and the negative signal line transmitting the anti-phase signal. After completing the channel connection, short-circuit calibrators, open-circuit calibrators, load calibrators, and direct-through calibrators are sequentially connected to the test terminals of the differential test channel. The short-circuit calibrator is a standard component that shorts the positive and negative terminals; the open-circuit calibrator is a standard component that opens the positive and negative terminals; the load calibrator is a standard component that connects a standard impedance between the positive and negative terminals; and the direct-through calibrator is a standard component that directly connects the two test ports. After each calibrator is connected, the vector network analyzer measures the mixed-mode scattering parameters of that calibrator. The mixed-mode scattering parameters are a parameter matrix describing the transmission characteristics of differential signals. They include differential mode transmission parameters and common-mode to differential mode conversion parameters. The differential mode transmission parameter Sdd21 represents the transmission coefficient of the differential signal from port 1 to port 2, and the common-mode to differential mode conversion parameter Scd21 represents the conversion coefficient of the common-mode signal to the differential-mode signal.

[0018] After acquiring the mixed-mode scattering parameters of each calibration component, the host computer calculates the error model by combining the standard values ​​of each calibration component. The standard values ​​are the theoretical ideal characteristic parameters of the calibration components; for example, the standard value for a short-circuit calibration component is a reflection coefficient of -1, and the standard value for a through-circuit calibration component is a transmission coefficient of 1. By establishing a set of equations relating the measured values ​​and the standard values, the error model of the differential test channel is obtained. The error model includes error coefficients for differential mode transmission and error coefficients for common-mode to differential mode conversion. These error coefficients quantify the system errors introduced by the test channel itself, such as transmission loss, reflection, and crosstalk. Following the same method, the host computer controls the switch matrix to sequentially switch to the second, third, and finally last differential test channel, completing the generation of error models for all channels.

[0019] After completing the calibration of each channel, the system begins measuring the crosstalk coupling coefficient between the differential test channels. The crosstalk coupling coefficient characterizes the degree of electromagnetic interference generated by one differential test channel to adjacent differential test channels. Specifically, during the measurement, the host computer controls the switch matrix to connect the first and second test ports of the vector network analyzer to the positive and negative signal lines of the first differential test channel, respectively, while simultaneously connecting the third and fourth test ports to the positive and negative signal lines of the second differential test channel, respectively. The vector network analyzer inputs differential excitation signals to the second differential test channel through the third and fourth test ports, and simultaneously measures the coupling response generated by the first differential test channel through the first and second test ports. The coupling response is the induced signal generated by the electromagnetic coupling of the second differential test channel in the first differential test channel without direct excitation. The host computer converts the measured coupling response into a differential mode coupling response, converts the excitation signal of the second differential test channel into a differential mode excitation signal, and then calculates the transfer function between the differential mode coupling response and the differential mode excitation signal. This transfer function is the crosstalk coupling coefficient between the first and second differential test channels. Using the same method, the system sequentially measures the crosstalk coupling coefficient between all adjacent differential test channel pairs.

[0020] Through the aforementioned calibration and crosstalk measurement processes, the system established a complete test channel error model and inter-channel crosstalk coupling coefficient matrix. These parameters provide fundamental data for error de-embedding and crosstalk compensation in subsequent testing phases, enabling the system to accurately separate the true transmission characteristics of the sample under test from the raw measurement data. This effectively eliminates system errors and inter-channel crosstalk introduced by the testing system itself, significantly improving the accuracy of differential high-speed signal testing.

[0021] Based on the above embodiments, as an optional implementation method, in S101, the control switch matrix is ​​sequentially switched to each differential test channel to calibrate each differential test channel and generate an error model for each differential test channel, specifically including S11-S13: S11, the control switch matrix connects the first test port of the vector network analyzer to the positive signal line of each differential test channel, and connects the second test port to the negative signal line of each differential test channel.

[0022] The host computer sends control commands to the switch matrix via a communication interface. The relay switches inside the switch matrix operate according to these commands, connecting the first test port of the vector network analyzer to the positive signal line of the first differential test channel via an RF cable and connector, and simultaneously connecting the second test port to the negative signal line of the first differential test channel. This connection method establishes a two-port test configuration between the vector network analyzer and the differential test channel, enabling the vector network analyzer to excite and measure both lines of the differential signal pair separately, thereby obtaining the complete differential transmission characteristics. Physically, the positive and negative signal lines are typically impedance-matched parallel transmission lines. The positive signal line carries the in-phase signal component, and the negative signal line carries the out-of-phase signal component; the difference between the two constitutes the differential signal.

[0023] S12 connects the short-circuit calibrator, open-circuit calibrator, load calibrator, and through-circuit calibrator sequentially to the positive and negative signal lines of each differential test channel, and measures the mixed-mode scattering parameters of each calibrator using a vector network analyzer.

[0024] Four standard calibration pieces were sequentially connected to the test terminals of the differential test channel using an automatic calibration device. First, the short-circuit calibration piece was connected. This piece directly short-circuits the positive and negative signal lines via a low-impedance metal plate, creating a near-zero ohm short circuit. The vector network analyzer emitted a test signal to the short-circuit calibration piece and measured the characteristics of the reflected signal to obtain the mixed-mode scattering parameters under short-circuit conditions. Mixed-mode scattering parameters are a parameter system specifically designed to describe the characteristics of differential circuits, converting traditional single-ended scattering parameters into a combination of differential and common-mode parameters. For example, Sdd11 represents the reflection coefficient of the differential mode at port 1, Scc11 represents the reflection coefficient of the common mode at port 1, and Scd11 represents the conversion coefficient from common-mode to differential mode at port 1. After removing the short-circuit calibration piece, the operator connected the open-circuit calibration piece. This completely disconnects the positive and negative signal lines, creating an open circuit. The vector network analyzer then measured the mixed-mode scattering parameters under open-circuit conditions. Next, connect the load calibration unit. The load calibration unit connects a precision resistor, typically 100 ohms, between the positive and negative signal lines. This resistor absorbs the incident signal energy to simulate an ideal matched load. The vector network analyzer measures the mixed-mode scattering parameters under matched load conditions. Finally, connect the pass-through calibration unit. The pass-through calibration unit is a standard transmission line structure that directly connects the input and output terminals of the test channel. The vector network analyzer simultaneously excites from the first test port and receives from the second test port, measuring the mixed-mode scattering parameters of the signal during pass-through transmission.

[0025] S13. Based on the mixed-mode scattering parameters of each calibrator and the standard values ​​of each calibrator, calculate the error model of each differential test channel. The error model includes the error coefficient of differential mode transmission and the error coefficient of common-mode to differential-mode conversion.

[0026] The standard values ​​for each calibration component are retrieved from the database. For the short-circuit calibration component, the standard value is: differential mode reflection coefficient Sdd11 = -1 and transmission coefficient Sdd21 = 0. For the open-circuit calibration component, the standard value is: differential mode reflection coefficient Sdd11 = +1 and transmission coefficient Sdd21 = 0. For the load calibration component, the standard value is: differential mode reflection coefficient Sdd11 = 0 and transmission coefficient Sdd21 = 0. For the straight-through calibration component, the standard value is: differential mode reflection coefficient Sdd11 = 0 and transmission coefficient Sdd21 = 1. The host computer establishes a set of mathematical equations for the error model, which describes the relationship between the measured value, the true value, and the error coefficients. The error model is represented using a multi-port error network and includes multiple error terms such as directivity error, source matching error, reflection tracking error, transmission tracking error, load matching error, and crosstalk error. For differential testing, the error model specifically needs to include the error coefficients for differential mode transmission and the error coefficients for common-mode to differential-mode conversion. The error coefficient of differential mode transmission quantifies the amplitude attenuation and phase delay of the differential signal during transmission in the test channel. The error coefficient of common-mode to differential-mode conversion quantifies the degree to which common-mode interference is converted into differential-mode signal. This conversion is usually caused by the asymmetry of the two lines in the test channel.

[0027] The host computer substitutes the measured values ​​and standard values ​​of the four sets of calibration components into the error model equations, forming a linear or nonlinear equation system containing multiple unknown error coefficients. Using matrix solving or least-squares fitting algorithms, the host computer calculates the values ​​of all error coefficients. These error coefficients fully describe the systematic error characteristics of the differential test channel, constituting the error model for that channel. The host computer stores the error model and associates it with the corresponding differential test channel identifier. Following the same process, the host computer controls the switch matrix to sequentially switch to the second, third, and finally the last differential test channel, repeating operations S11 to S13 to generate an independent error model for each differential test channel.

[0028] Based on the above embodiments, as an optional implementation, in S101, measuring the crosstalk coupling coefficient between each differential test channel specifically includes S14-S16: S14, the control switch matrix connects the first and second test ports of the vector network analyzer to the positive and negative signal lines of each differential test channel, respectively, and connects the third and fourth test ports to the positive and negative signal lines of the adjacent differential test channels of each differential test channel, respectively.

[0029] Specifically, the switch matrix connects the first test port of the vector network analyzer to the positive signal line of the first differential test channel, the second test port to the negative signal line of the first differential test channel, the third test port to the positive signal line of the second differential test channel, and the fourth test port to the negative signal line of the second differential test channel. The second differential test channel is an adjacent differential test channel of the first differential test channel. "Adjacent" means that these two differential test channels are physically close together, typically running side-by-side or located at adjacent pin positions on the same connector. This physical proximity results in a non-negligible electromagnetic coupling between the two channels. This four-port connection configuration allows the vector network analyzer to both input signals to the second differential test channel as crosstalk sources through the third and fourth test ports, and monitor the crosstalk signals induced on the first differential test channel through the first and second test ports, thus establishing a complete test link between the crosstalk transmitter and receiver.

[0030] S15, using a vector network analyzer, simultaneously excites each differential test channel and its adjacent differential test channels, measuring the coupling response of each differential test channel under the excitation of the adjacent differential test channels.

[0031] The vector network analyzer simultaneously inputs differential excitation signals of equal amplitude and opposite phase to the second differential test channel through the third and fourth test ports. These differential excitation signals propagate along the second differential test channel and radiate energy into the surrounding space through an electromagnetic field. Since the first and second differential test channels are physically adjacent, the electromagnetic field generated by the second differential test channel couples with the first differential test channel, inducing coupling current and voltage on the positive and negative signal lines of the first differential test channel. The vector network analyzer measures the signal on the first differential test channel in real time through the first and second test ports to obtain the coupling response. The coupling response is the induced signal generated by the electromagnetic coupling of the second differential test channel in the first differential test channel without direct signal excitation. The amplitude and phase characteristics of this signal reflect the crosstalk strength between the two channels. Simultaneously, the vector network analyzer reads back the actual excitation signal of the second differential test channel through the third and fourth test ports, recording the amplitude, phase, and frequency characteristics of the excitation signal, which is the source signal of the crosstalk.

[0032] S16. Based on the coupling response of each differential test channel and the excitation signal of the adjacent differential test channel, calculate the crosstalk coupling coefficient between each differential test channel and the corresponding adjacent differential test channel.

[0033] The host computer first converts the coupling response measured by the first differential test channel into a differential mode coupling response. Specifically, the conversion method involves subtracting the response of the negative signal line from the response of the positive signal line and then dividing by 2 to obtain the pure differential mode component. Similarly, the host computer converts the excitation signal of the second differential test channel into a differential mode excitation signal. The purpose of converting to differential mode is to eliminate the influence of common-mode interference, because the effective information transmitted by the differential signal is carried in the differential mode, while the common-mode component is usually noise and interference. The host computer calculates the transfer function between the differential mode coupling response and the differential mode excitation signal in the frequency domain. The formula for this transfer function is the spectrum of the differential mode coupling response divided by the spectrum of the differential mode excitation signal. The resulting transfer function is the crosstalk coupling coefficient between the first and second differential test channels. This coefficient is a complex frequency domain function, where the amplitude represents the attenuation of the crosstalk signal relative to the excitation signal, and the phase represents the phase delay of the crosstalk signal relative to the excitation signal. For example, if the amplitude of the crosstalk coupling coefficient is -40dB at a certain frequency, it means that the excitation signal of the second differential test channel is attenuated by 10,000 times after being transmitted to the first differential test channel through crosstalk coupling.

[0034] Following the same method, the host computer controls the switch matrix to sequentially establish four-port connections between other differential test channels and their adjacent differential test channels, repeating operations S14 to S16 to measure and calculate the crosstalk coupling coefficient between all adjacent channel pairs. For a test system containing N differential test channels, it is typically necessary to measure the crosstalk coupling coefficient of N-1 pairs of adjacent channel pairs. The host computer organizes all crosstalk coupling coefficients into a crosstalk coupling matrix, where each element represents the crosstalk propagation relationship between a pair of differential test channels. This matrix comprehensively describes the crosstalk network characteristics between all channels in the test system.

[0035] Based on the above embodiments, as an optional implementation, in S16, calculating the crosstalk coupling coefficient between each differential test channel and its corresponding adjacent differential test channel according to the coupling response of each differential test channel and the excitation signal of the adjacent differential test channel specifically includes S161-S162: S161 converts the coupling response of each differential test channel into a differential mode coupling response, and converts the excitation signal of adjacent differential test channels into a differential mode excitation signal.

[0036] Specifically, in step S15, the host computer acquires the coupling response V1+ of the positive signal line of the first differential test channel measured by the vector network analyzer through the first test port, and the coupling response V1- of the negative signal line measured through the second test port. Both responses are single-ended parameters, containing a mixture of differential mode and common-mode components. The host computer calculates the differential mode coupling response Vdd1 according to the definition formula of differential mode, which is Vdd1 equal to the difference between V1+ and V1- divided by 2. This calculation process extracts the difference component between the positive and negative signal lines, eliminates in-phase common-mode interference on the two lines, and obtains a pure differential mode coupled signal. The differential mode coupling response truly reflects the signal strength transmitted from the second differential test channel to the first differential test channel through the differential crosstalk mechanism.

[0037] Simultaneously, the host computer performs the same mode conversion on the excitation signal. The host computer acquires the excitation signal of the second differential test channel read back by the vector network analyzer through the third and fourth test ports, denoted as V2+ and V2-, respectively. The host computer calculates the differential mode excitation signal Vdd2, with the formula: Vdd2 equals the difference between V2+ and V2- divided by 2. The differential mode excitation signal represents the differential mode energy actually applied to the second differential test channel and is the effective source signal for crosstalk. The reason for performing mode conversion is that the main propagation path of crosstalk in differential signal transmission systems is differential mode-to-differential mode coupling. Although common mode-to-differential mode conversion exists, it is usually a second-order effect. By extracting differential mode parameters, we can focus on the main crosstalk mechanism, simplifying subsequent crosstalk modeling and compensation algorithms.

[0038] S162, calculate the transfer function between the differential mode coupling response and the differential mode excitation signal to obtain the crosstalk coupling coefficient between each differential test channel and the corresponding adjacent differential test channel; the crosstalk coupling coefficient characterizes the degree of crosstalk influence of adjacent differential test channels on each differential test channel.

[0039] The host computer first performs a Fast Fourier Transform (FFT) on the differential mode coupling response Vdd1 and the differential mode excitation signal Vdd2, converting the time-domain signal into a frequency-domain spectrum. The frequency domain represents the behavior that can separate different frequency components, which is crucial for analyzing frequency-dependent crosstalk characteristics. This is because crosstalk coupling strength typically varies with frequency, and high-frequency signals, due to their shorter wavelengths and stronger electromagnetic radiation, often generate more significant crosstalk. The host computer calculates the transfer function H(f) at each frequency point. The formula is: H(f) equals the differential mode coupling response spectrum Vdd1(f) divided by the differential mode excitation signal spectrum Vdd2(f). This transfer function H(f) is the crosstalk coupling coefficient between the first and second differential test channels.

[0040] The crosstalk coupling coefficient is a complex function, whose real and imaginary parts describe the amplitude and phase characteristics of crosstalk propagation. The amplitude |H(f)| of the crosstalk coupling coefficient characterizes the degree of crosstalk influence between adjacent differential test channels; a larger value indicates stronger crosstalk. For example, if the amplitude of the crosstalk coupling coefficient is 0.01 at a frequency of 10 GHz, it means that a 1-volt differential excitation signal on the second differential test channel will induce a 0.01-volt crosstalk signal on the first differential test channel. The phase ∠H(f) of the crosstalk coupling coefficient characterizes the phase delay of the crosstalk signal relative to the excitation signal. This phase information is crucial for crosstalk compensation in the time domain, because only by simultaneously correcting the amplitude and phase can the crosstalk effect be completely eliminated. The host computer stores the calculated crosstalk coupling coefficient as a frequency domain complex array, where each element corresponds to the crosstalk propagation characteristics at a specific frequency point.

[0041] Following the same calculation process, the host computer performs mode conversion and transfer function calculation on the measurement data of all adjacent differential test channel pairs, obtaining a complete set of crosstalk coupling coefficients. These crosstalk coupling coefficients accurately quantify the crosstalk propagation relationship between each pair of adjacent channels in the test system, providing a precise mathematical model for the crosstalk compensation algorithm in the subsequent S102 step. Through this crosstalk coefficient extraction method based on transfer function, the system transforms the complex electromagnetic coupling phenomenon into a computable frequency domain transfer function, enabling crosstalk effects to be accurately removed from the measurement results through mathematical operations, significantly improving the inter-channel isolation and measurement accuracy of multi-channel differential signal testing.

[0042] S102, when entering the test mode, the control switch matrix is ​​switched sequentially to each measured differential channel of the sample to be tested. The differential transmission parameters of each measured differential channel are measured by the vector network analyzer. The differential transmission parameters are corrected by combining the error model and crosstalk coupling coefficient to generate the target differential transmission parameters.

[0043] In practice, the operator connects each differential channel of the sample under test to its corresponding differential test channel. The sample under test is a differential signal transmission device or circuit board whose performance needs to be evaluated, and its differential channels are the differential signal transmission paths on the sample under test that need to be tested. The host computer controls the switch matrix to sequentially switch the first and second test ports of the vector network analyzer to the positive and negative terminals of the first differential channel under test of the sample under test. The vector network analyzer inputs a differential test signal to the first differential channel under test and measures the differential transmission parameters of that channel. The differential transmission parameters include differential mode scattering parameters Sdd11, Sdd21, Sdd12, and Sdd22, among which Sdd21 represents the transmission coefficient of the differential signal from the input to the output, which is the core parameter characterizing the transmission performance of the differential channel. Since the test signal needs to pass through the test channel to reach the sample under test, the measured differential transmission parameters are actually a comprehensive result of test channel error, sample characteristics, and inter-channel crosstalk, therefore, a two-step correction process is required.

[0044] First, error de-embedding is performed. The host computer calculates the error de-embedding for the measured differential transmission parameters based on the error model of the first differential test channel. Error de-embedding is the process of establishing mathematical equations using the error coefficients in the error model, and then solving backwards from the original measured values ​​to obtain the true transmission parameters after removing the influence of the test channel. Specifically, the measured differential transmission parameters are used in matrix operations with the differential mode transmission error coefficients and common-mode to differential-mode conversion error coefficients in the error model to eliminate system errors such as transmission loss, phase delay, and impedance mismatch introduced by the test channel itself, resulting in the de-embedding differential transmission parameters. While the de-embedding differential transmission parameters have eliminated the influence of the test channel itself, they still contain the influence of crosstalk from adjacent channels.

[0045] Next, crosstalk compensation is performed. Since multiple differential channels of the sample under test are typically physically adjacent and operate simultaneously, when the vector network analyzer measures the first differential channel, the adjacent second differential channel may transmit signals simultaneously. The signal from the second differential channel will cause crosstalk interference to the first differential channel through electromagnetic coupling. The host computer calculates the crosstalk impact on the first differential channel based on the crosstalk coupling coefficient obtained in step S101 and the signal state of the currently measured adjacent differential channels. The formula for calculating the crosstalk impact is the product of the differential transmission parameter of the adjacent channel and the crosstalk coupling coefficient. The host computer subtracts the crosstalk impact from the de-embedding differential transmission parameters to obtain the target differential transmission parameters after eliminating crosstalk. The target differential transmission parameters are differential transmission parameters that eliminate both the system error of the test channel and the crosstalk between channels, truly reflecting the inherent transmission characteristics of the differential channel under test.

[0046] Following the same method, the host computer controls the switch matrix to sequentially switch to the second, third, and finally last differential channel of the sample under test. For each differential channel, measurement, error de-embedding, and crosstalk compensation are performed to generate the target differential transmission parameters for each channel. Through this calibration process, the system can accurately separate the true performance indicators of each differential channel under test in a multi-channel parallel testing environment, effectively eliminating test system errors and inter-channel interference. This ensures that the measurement results truly reflect the differential signal transmission quality of the sample under test, providing reliable test data for product performance evaluation and quality control.

[0047] Based on the above embodiments, as an optional implementation, in S102, the differential transmission parameters are corrected by combining the error model and the crosstalk coupling coefficient to generate the target differential transmission parameters, specifically including S21-S23: S21. Based on the error model of each differential test channel, the differential transmission parameters of each measured differential channel are de-embedded to obtain the de-embedded differential transmission parameters.

[0048] Specifically, the host computer acquires the original differential transmission parameter Sdd21_meas of the first differential channel under test measured by the vector network analyzer. This parameter contains the superposition effect of the true characteristics of the tested sample and the test channel error. The host computer retrieves the error model of the first differential test channel generated in step S101 from memory. The error model includes the differential mode transmission error coefficient and the common-mode to differential-mode conversion error coefficient of the channel. Based on the mathematical relationship established by the error model, the host computer substitutes the original measured value Sdd21_meas into the error deembedding algorithm. The essence of the error deembedding algorithm is to solve an inverse transfer function. This algorithm mathematically separates the transmission characteristics of the test channel from the measurement results through matrix operations. In specific calculations, the host computer performs matrix division or inverse matrix multiplication operations on the original measured value and the error coefficients in the error model to obtain the deembedded differential transmission parameter Sdd21_deembed. The de-embedding differential transmission parameters eliminate systematic errors such as transmission loss, reflection mismatch, and mode conversion of the test channel itself. Theoretically, they should be equal to the true transmission characteristics of the sample under test in an ideal test environment. However, they still contain interference components introduced by crosstalk coupling between adjacent differential channels under test.

[0049] S22, calculate the crosstalk influence between each differential test channel based on the crosstalk coupling coefficient between each differential test channel.

[0050] The host computer identifies the adjacent differential channel of the first differential channel under test as the second differential channel under test, and retrieves the crosstalk coupling coefficient H12(f) between the two channels measured in step S101 from the memory. Simultaneously, the host computer acquires the measurement signal Sdd21_ch2 of the second differential channel under test. The measurement signal refers to the signal transmitted on the second differential channel while the vector network analyzer is measuring the first differential channel; this signal may originate from excitation from other ports or reflected echoes. The host computer performs a complex multiplication operation in the frequency domain between the measurement signal Sdd21_ch2 of the second differential channel under test and the crosstalk coupling coefficient H12(f). The calculation formula is that the crosstalk effect Xtalk12(f) equals H12(f) multiplied by Sdd21_ch2(f). This multiplication operation simulates the physical process of crosstalk, i.e., the signal on the second differential channel under test leaks to the first differential channel under test through electromagnetic coupling according to the transmission relationship described by the crosstalk coupling coefficient. If the test system contains multiple adjacent channels, the host computer needs to calculate the crosstalk effect for each adjacent channel separately and then perform vector superposition to obtain the total crosstalk effect. The crosstalk effect is a complex function in the frequency domain, which accurately quantifies the amplitude and phase characteristics of the interference signal caused by adjacent channels to the current channel under test.

[0051] S23, remove the crosstalk effect from the de-embedding differential transmission parameters to obtain the target differential transmission parameters; the target differential transmission parameters are differential transmission parameters that have eliminated the effects of test channel error and inter-channel crosstalk.

[0052] The host computer subtracts the crosstalk effect Xtalk12(f) from the deembedded differential transmission parameter Sdd21_deembed. The calculation formula is: the target differential transmission parameter Sdd21_target(f) equals Sdd21_deembed(f) minus Xtalk12(f). This subtraction operation is performed independently at each frequency point in the frequency domain and is a complex subtraction, simultaneously correcting amplitude and phase. The physical meaning of the subtraction operation is to remove the interference components coupled from adjacent channels from the measurement results, restoring the true transmission characteristics of the tested sample in an ideal crosstalk-free environment. The host computer stores the calculated target differential transmission parameter and associates it with the corresponding tested differential channel identifier. The target differential transmission parameter is the final measurement result after two-stage correction processing, eliminating both the systematic errors of the test channel itself and the crosstalk interference from adjacent channels, truly reflecting the differential transmission performance of the tested sample.

[0053] Following the same processing flow, the host computer sequentially performs error de-embedding and crosstalk compensation on the measurement data of all tested differential channels, generating a complete set of target differential transmission parameters. This two-stage correction strategy fully utilizes the error model and crosstalk coupling coefficient established in step S101, decomposing the complex measurement error into two independently processable parts. Error de-embedding eliminates the transmission distortion of a single channel, making the measurement accuracy no longer limited by the device quality and connection status of the test channel. Crosstalk compensation eliminates the coupling interference unique to multi-channel systems, allowing each channel to be measured independently without affecting each other. Through this comprehensive error correction mechanism, the system achieves measurement accuracy close to that of an ideal testing environment, obtaining reliable differential transmission parameters even under high-density multi-channel test configurations, meeting the stringent requirements of high-speed differential signal testing for measurement accuracy and channel isolation.

[0054] S103, in the test mode, periodically inserts crosstalk calibration frames. Within the crosstalk calibration frame, the control switch matrix switches to the differential crosstalk calibrator, and the target channel response and crosstalk source channel response of the differential crosstalk calibrator are measured by the vector network analyzer.

[0055] In practice, the host computer sets a crosstalk calibration counter in the test mode. After a preset number of measurements of the differential channels of the test sample are completed, the counter reaches a preset threshold, and the system automatically triggers a crosstalk calibration frame. The crosstalk calibration frame is a special test period temporarily inserted into the normal test process, specifically used to evaluate and update the crosstalk characteristics of the test system. For example, the system can be set to insert a crosstalk calibration frame after every 100 measurements of the differential channels under test. This ensures timely crosstalk tracking without excessively interrupting the normal test process.

[0056] When the crosstalk calibration frame is triggered, the host computer controls the switch matrix to pause the testing of the sample under test and switches the test port of the vector network analyzer from the sample under test to the differential crosstalk calibrator. The differential crosstalk calibrator is a standard device with known standard crosstalk characteristics. Its design structure ensures controllable and stable crosstalk coupling between the two differential channels. These crosstalk characteristics have been predetermined through precise measurements and stored in the system as a standard reference value. The differential crosstalk calibrator contains two differential channels: a target channel and a crosstalk source channel. The target channel is used to receive crosstalk signals, and the crosstalk source channel is used to generate crosstalk interference.

[0057] The host computer controls the switch matrix to connect the first and second test ports of the vector network analyzer to the positive and negative terminals of the target channel of the differential crosstalk calibrator, respectively, and to the positive and negative terminals of the crosstalk source channel of the differential crosstalk calibrator, respectively. The vector network analyzer inputs differential excitation signals to the crosstalk source channel through the third and fourth test ports, while simultaneously measuring the response signals generated by the target channel through the first and second test ports, and reading back the actual excitation signals of the crosstalk source channel through the third and fourth test ports. The measured target channel response reflects the actual crosstalk impact of the crosstalk source channel on the target channel under the current test system state, while the crosstalk source channel response records the characteristics of the excitation signal used to generate crosstalk.

[0058] Since the crosstalk characteristics of the differential crosstalk calibrator are known and stable standard values, when the crosstalk characteristics of the test system drift, the actual measured target channel response will deviate from the theoretical target channel response calculated based on the standard crosstalk characteristics and the crosstalk source channel response. By comparing the difference between the actual measured target channel response and the theoretical target channel response, the system can detect the changing trend of the crosstalk coupling coefficient. After completing the measurement with the differential crosstalk calibrator, the crosstalk calibration frame ends, and the host computer controls the switch matrix to switch the test port back to the sample under test, restoring the normal test process.

[0059] By periodically inserting crosstalk calibration frames and measuring the response characteristics of the differential crosstalk calibrator, the system establishes a crosstalk characteristic monitoring mechanism, providing real-time reference data for the adaptive update of the crosstalk coupling coefficient in the subsequent S104 step. This periodic calibration strategy enables the test system to dynamically track the time-varying characteristics of crosstalk, ensuring accurate crosstalk compensation capability throughout long-term testing and avoiding the problem of decreased test accuracy caused by crosstalk characteristic drift.

[0060] Based on the above embodiments, as an optional implementation, in S103, the control switch matrix switches to the differential crosstalk calibration device, and the target channel response and crosstalk source channel response of the differential crosstalk calibration device are measured by a vector network analyzer, specifically including S31-S32: S31, the control switch matrix connects the first and second test ports of the vector network analyzer to the positive and negative terminals of the target channel of the differential crosstalk calibrator, respectively, and connects the third and fourth test ports to the positive and negative terminals of the crosstalk source channel of the differential crosstalk calibrator, respectively.

[0061] Specifically, the switch matrix connects the first test port of the vector network analyzer to the positive terminal of the target channel of the differential crosstalk calibrator via an RF cable, and the second test port to the negative terminal of the target channel. The target channel is the differential channel on the differential crosstalk calibrator used to receive crosstalk signals, simulating the differential channel under test affected by crosstalk in actual testing. Simultaneously, the switch matrix connects the third test port to the positive terminal of the crosstalk source channel of the differential crosstalk calibrator, and the fourth test port to the negative terminal of the crosstalk source channel. The crosstalk source channel is the differential channel on the differential crosstalk calibrator used to generate crosstalk interference, simulating the adjacent differential channel transmitting crosstalk in actual testing. The differential crosstalk calibrator is a standard component with known crosstalk characteristics. Its internal structure typically contains two parallel traces or coupled transmission lines with controllable and repeatable electromagnetic coupling between them. The manufacturer has calibrated the crosstalk transmission coefficient of the calibrator at different frequencies through precise measurement or simulation; these standard values ​​are stored as the crosstalk characteristic parameters of the calibrator.

[0062] This four-port connection configuration is identical to the configuration used in step S14 for measuring differential test channel crosstalk, ensuring consistency in measurement conditions. By connecting the differential crosstalk calibrator to the same test path as the actual test sample, the calibrator undergoes the exact same test channel transmission, switch matrix switching, and vector network analyzer port connection process as the test sample. Therefore, the calibrator's measurement results accurately reflect the crosstalk measurement capability under the current test system conditions. If the crosstalk characteristics of the test system change, for example, if temperature variations alter the inter-line coupling of the test cable or if impedance drift at the relay contacts of the switch matrix increases signal leakage, these changes will be reflected in the measurement results of the differential crosstalk calibrator.

[0063] S32, the crosstalk source channel of the differential crosstalk calibrator is excited by a vector network analyzer, and the target channel response and crosstalk source channel response of the differential crosstalk calibrator are measured; the differential crosstalk calibrator is a standard part with known crosstalk characteristics.

[0064] The vector network analyzer inputs a differential excitation signal to the crosstalk source channel of the differential crosstalk calibrator through the third and fourth test ports. This excitation signal forms a differential voltage with equal amplitude and opposite phase between the positive and negative terminals of the crosstalk source channel. The differential excitation signal propagates along the crosstalk source channel, leaking energy to the target channel through electromagnetic field coupling during propagation. Due to the carefully designed stable coupling structure of the two channels inside the differential crosstalk calibrator, the excitation signal induces a crosstalk signal on the target channel according to the standard crosstalk characteristics of the calibrator. The vector network analyzer measures the voltage response of the positive and negative terminals of the target channel in real time through the first and second test ports to obtain the target channel response. The target channel response includes the crosstalk signal coupled from the crosstalk source channel and the superposition of the test system's background noise. This response directly reflects the actual strength and phase characteristics of the crosstalk signal after being transmitted through the current test system.

[0065] Simultaneously, the vector network analyzer reads back the actual signal state of the crosstalk source channel through the third and fourth test ports to obtain the crosstalk source channel response. The crosstalk source channel response includes not only the excitation signal output by the vector network analyzer but also the reflected signal due to impedance mismatch and any possible reverse coupling signal from the target channel. Recording the crosstalk source channel response aims to accurately understand the amplitude and phase of the actual excitation signal acting on the calibration device, as this excitation signal may differ from the ideal output of the vector network analyzer. Only by using the actual excitation signal can the crosstalk transmission coefficient be accurately calculated. The host computer records both the target channel response and the crosstalk source channel response completely; these two sets of data constitute the measured crosstalk characteristics of the differential crosstalk calibration device under the current test conditions.

[0066] S104 uses the crosstalk source channel response as the input signal and the theoretical target channel response of the differential crosstalk calibrator as the desired signal. Based on the error between the target channel response and the theoretical target channel response, the crosstalk coupling coefficient is updated to generate the target crosstalk coupling coefficient.

[0067] In practice, the host computer first calculates the theoretical target channel response based on the known crosstalk characteristics of the differential crosstalk calibrator. The known crosstalk characteristics of the differential crosstalk calibrator are the standard crosstalk transfer function calibrated at the factory, describing the standard attenuation and phase relationship of the crosstalk source channel signal transmitted to the target channel under ideal test conditions. The host computer acquires the crosstalk source channel response measured in step S103, and performs a convolution operation between this response signal and the known crosstalk characteristics of the differential crosstalk calibrator to obtain the theoretical response that the target channel should produce under standard crosstalk conditions. Then, the host computer incorporates the current crosstalk coupling coefficient into the calculation. This crosstalk coupling coefficient represents the actual crosstalk path characteristics between channels in the test system. The crosstalk source channel response is modulated using the crosstalk coupling coefficient and superimposed on the theoretical response to obtain the theoretical target channel response that comprehensively considers both the characteristics of the standard component and the characteristics of the test system.

[0068] Next, the host computer calculates the error signal between the actual measured target channel response and the theoretical target channel response. The error signal reflects the degree of deviation between the current crosstalk coupling coefficient and the actual crosstalk characteristics of the test system. If the crosstalk coupling coefficient is accurate, the theoretical target channel response should be consistent with the actual measured target channel response, and the error signal should be close to zero. Conversely, if the crosstalk coupling coefficient is inaccurate due to system drift, a significant error signal will occur between the two.

[0069] The host computer uses the Normalized Least Mean Square (NMS) algorithm to iteratively update the crosstalk coupling coefficients. The NMS algorithm is an adaptive filtering algorithm whose core idea is to treat the crosstalk coupling coefficients as tap coefficients of an adaptive filter. These coefficients are gradually adjusted by minimizing the mean square value of the error signal, making the filter output approximate the desired signal. Specifically, the host computer first calculates the normalized power of the crosstalk source channel response, which is the energy of the crosstalk source channel response signal plus a small regularization constant. This normalization operation prevents the algorithm from becoming unstable due to excessively large update step sizes when the input signal power is small. Then, the host computer calculates the product of the error signal and the crosstalk source channel response, and divides it by the normalized power to obtain the normalized gradient direction. The host computer then adds the product of the normalized gradient direction and the preset step size factor to the current crosstalk coupling coefficients to obtain the updated crosstalk coupling coefficients. The step size factor is a parameter that controls the update speed. A larger step size factor makes the algorithm converge faster but may cause oscillations, while a smaller step size factor makes the algorithm more stable but converges slower. The system usually sets the step size factor to an empirical value between 0.01 and 0.1.

[0070] The host computer repeatedly executes the error calculation and coefficient update process, adjusting the crosstalk coupling coefficient in each iteration to reduce the error. As the iteration progresses, the theoretical target channel response gradually approaches the actual measured target channel response, and the amplitude of the error signal continuously decreases. When the mean square value of the error signal converges below a preset threshold, it indicates that the crosstalk coupling coefficient accurately reflects the crosstalk characteristics of the current test system. The host computer then stops iterating and determines the crosstalk coupling coefficient at this point as the target crosstalk coupling coefficient. The target crosstalk coupling coefficient is a coefficient parameter that, after adaptive updating, can accurately describe the crosstalk state of the current test system.

[0071] The host computer replaces the previously used crosstalk coupling coefficient with the target crosstalk coupling coefficient. In the subsequent S102 step, the target crosstalk coupling coefficient is used to calculate crosstalk compensation for the differential transmission parameters of the sample under test. Through this adaptive update mechanism, the system achieves dynamic tracking of crosstalk compensation capability. Even when crosstalk characteristics drift due to changes in test environment temperature or equipment aging, the system can still automatically adjust the crosstalk coupling coefficient to match the current actual crosstalk state, ensuring the continuous accuracy of crosstalk compensation during long-term continuous testing and significantly improving the stability and reliability of differential high-speed signal testing.

[0072] Based on the above embodiments, as an optional implementation, in S104, the step of using the crosstalk source channel response as the input signal, using the theoretical target channel response of the differential crosstalk calibrator as the desired signal, and updating the crosstalk coupling coefficient according to the error between the target channel response and the theoretical target channel response to generate the target crosstalk coupling coefficient specifically includes S41-S43: S41. Based on the known crosstalk characteristics and crosstalk source channel response of the differential crosstalk calibrator, and combined with the current crosstalk coupling coefficient, calculate the theoretical target channel response of the differential crosstalk calibrator.

[0073] Specifically, the host computer retrieves the known crosstalk characteristic Hcal_std(f) of the differential crosstalk calibrator from the calibration database. This crosstalk characteristic is a standard value provided by the calibration manufacturer and describes the ideal transfer function between the crosstalk source channel and the target channel of the calibration. The host computer obtains the crosstalk source channel response Vsrc(f) measured in step S32, which represents the excitation signal actually applied to the crosstalk source channel of the calibration. The host computer performs a complex multiplication operation with the current crosstalk coupling coefficient Hxt_current(f), which is the value obtained from the initial measurement in step S101 or the value updated in the previous iteration. The multiplication operation simulates the process of the test system transferring the crosstalk source channel signal to the target channel, and the calculation formula is that the system transfer response Hsys(f) equals Hxt_current(f) multiplied by Vsrc(f). The host computer then performs a convolution operation between the system's transmitted response and the known crosstalk characteristics of the calibration component. The calculation formula is that the theoretical target channel response Vtgt_theory(f) equals the product of Hcal_std(f) and Hsys(f). The theoretical target channel response represents the target channel signal that the differential crosstalk calibration component should theoretically produce under the test system state described by the current crosstalk coupling coefficient. This calculation process combines the standard characteristics of the calibration component and the current state model of the test system to establish a theoretical prediction of the complete transmission link from the crosstalk source to the target channel.

[0074] S42, calculate the error between the measured target channel response and the theoretical target channel response.

[0075] The host computer acquires the target channel response Vtgt_meas(f) obtained from the actual measurement in step S32. This response is the signal read by the vector network analyzer from the target channel of the calibration component under real test conditions. The host computer calculates the error between the measured value and the theoretical value at each frequency point in the frequency domain. The calculation formula is that the error function E(f) equals Vtgt_meas(f) minus Vtgt_theory(f). This error function is a complex frequency domain function, and its real and imaginary parts reflect the amplitude error and phase error, respectively. The error arises from the deviation between the current crosstalk coupling coefficient and the actual crosstalk characteristics of the test system. If the current crosstalk coupling coefficient accurately describes the state of the test system, the theoretical target channel response should perfectly match the measured value, with the error close to zero. If the error is large, it indicates that the crosstalk characteristics of the test system have deviated from the state at the initial measurement in step S101. For example, the ambient temperature may have increased from 25 degrees Celsius at the initial calibration to the current 35 degrees Celsius, causing a change in the dielectric constant of the test cable and altering the inter-line coupling strength. Alternatively, the contact surface oxidation of the relays in the switch matrix after thousands of switching cycles may have increased the contact impedance and introduced additional crosstalk paths. The host computer performs full-band analysis on the error function and calculates the root mean square value of the error as the overall error index.

[0076] S43, using the crosstalk source channel response as the input signal and the error as the feedback signal, the crosstalk coupling coefficient is iteratively updated. When the error converges to a preset threshold, the target crosstalk coupling coefficient is obtained. The target crosstalk coupling coefficient is used to compensate for the crosstalk drift caused by the test system with changes in time and environment.

[0077] The host computer employs an adaptive filtering algorithm to adjust the crosstalk coupling coefficient. This algorithm uses the crosstalk source channel response Vsrc(f) as the input signal and the error function E(f) as the feedback signal, calculating the update amount of the crosstalk coupling coefficient using the minimum mean square error criterion. Specifically, the host computer calculates the updated crosstalk coupling coefficient Hxt_new(f) according to the iterative formula of the adaptive algorithm. Hxt_new(f) is equal to Hxt_current(f) plus the step size factor μ multiplied by the conjugate product of the error E(f) and the input signal Vsrc(f). The step size factor μ is a parameter controlling the convergence speed, typically ranging from 0.001 to 0.1. A larger step size factor allows the algorithm to converge quickly but may cause oscillations, while a smaller step size factor makes the algorithm stable but converges more slowly. The physical meaning of this iterative formula is to adjust the crosstalk coupling coefficient along the direction of error gradient descent based on the current prediction error, so that the adjusted coefficient can more accurately predict the target channel response of the calibration component.

[0078] The host computer uses the updated crosstalk coupling coefficient Hxt_new(f) as the new current value and returns to step S41 to recalculate the theoretical target channel response, forming a closed-loop iteration. In each iteration, the error function gradually decreases, and the crosstalk coupling coefficient gradually approaches the optimal value that accurately describes the true state of the test system. The host computer continuously monitors the root mean square (RMS) value of the error, and determines that the algorithm has converged when this value drops below a preset threshold. The preset threshold is usually set according to the test accuracy requirements, for example, requiring the RMS value of the error to be less than 1% of the measured signal amplitude, corresponding to an error level of approximately -40dB. When the algorithm converges, the host computer stores the final crosstalk coupling coefficient as the target crosstalk coupling coefficient Hxt_target(f), which reflects the true crosstalk propagation characteristics of the test system under the current temperature, humidity, and device conditions.

[0079] S105 corrects the differential transmission parameters of subsequent measurements using the target crosstalk coupling coefficient.

[0080] In practice, the system resumes the test procedure of step S102, and the host computer controls the switch matrix to sequentially switch to each differential channel of the sample under test for measurement. After the vector network analyzer measures the differential transmission parameters of a certain differential channel under test, the host computer first performs error de-embedding processing on the measurement data according to the error model of the differential test channel to obtain the de-embedding differential transmission parameters. Next, when performing crosstalk compensation calculation, the host computer no longer uses the initial crosstalk coupling coefficient measured in step S101, but instead uses the target crosstalk coupling coefficient updated in step S104. The host computer calculates the crosstalk impact on the current differential channel under test based on the target crosstalk coupling coefficient and the measurement signals of adjacent differential channels under test. Since the target crosstalk coupling coefficient has been adaptively corrected using the actual measurement data of the differential crosstalk calibrator, it can accurately quantify the degree of crosstalk from adjacent channels to the current channel under the current test system state. Therefore, the calculated crosstalk impact is closer to the actual crosstalk interference than when using the initial crosstalk coupling coefficient. The host computer subtracts this more accurate crosstalk effect from the de-embedding differential transmission parameters to obtain the corrected target differential transmission parameters.

[0081] The system continuously uses the target crosstalk coupling coefficient for crosstalk compensation until the next crosstalk calibration frame is triggered. When the system re-enters step S103 to insert a new crosstalk calibration frame, it remeasures the response of the differential crosstalk calibrator and further updates the crosstalk coupling coefficient in step S104. If the crosstalk characteristics of the test system continue to drift during this period, a new round of adaptive updates will generate a new target crosstalk coupling coefficient to track this change. In this way, the system forms a closed-loop cycle of "normal test - insertion of calibration frame - adaptive update - application of updated coefficient - normal test" throughout the entire testing process, and the crosstalk coupling coefficient always remains synchronized with the current state of the test system.

[0082] By using a target crosstalk coupling coefficient for correction, the system achieves dynamic compensation for the time-varying crosstalk characteristics of the test system. Even during continuous testing lasting several hours or even days, even if the ambient temperature rises from 20 degrees Celsius in the morning to 30 degrees Celsius in the afternoon, causing thermal expansion and contraction of the test cable and altering the coupling characteristics between channels, or if the contact impedance of the test connector changes slightly due to repeated insertion and removal, the system can promptly capture these changes through periodic adaptive updates and adjust the crosstalk coupling coefficient accordingly. This adaptive crosstalk calibration mechanism makes the measurement accuracy no longer dependent on the stability of the test system state, effectively overcoming the accuracy degradation problem of traditional fixed calibration methods during long-term testing. It ensures that all test samples obtain consistent and accurate measurement results from the start to the end of the test, significantly improving the long-term stability and measurement reliability of the differential high-speed signal test system.

[0083] Figure 2This embodiment presents a flowchart of adaptive crosstalk calibration in a differential high-speed signal testing system, demonstrating the complete algorithm flow from system startup to the completion of crosstalk coupling coefficient updates. The entire process employs an iterative optimization design approach, achieving dynamic tracking and compensation of the time-varying characteristics of the testing system through periodic measurements of standard calibration components.

[0084] After the process starts at the initiation node, it first enters the system initialization phase. In this phase, the system completes three key preparatory tasks: organizing the test steps to ensure the orderly execution of the test process, starting the vector network analyzer to put it into working condition, opening the switch matrix to establish signal paths, and preheating the equipment to bring the test system to a stable operating temperature. These initialization operations lay the hardware foundation for subsequent precision measurements.

[0085] After initialization, the system enters the first judgment node to determine whether the test has started or the calibration cycle has been reached. This judgment constitutes the core triggering mechanism of the entire adaptive calibration algorithm. If the judgment result is "yes," it means that the crosstalk calibration process needs to be executed, and the system enters the calibration branch on the right. In this branch, the system first measures the response of the standard calibration components, specifically including comprehensive error model calibration to measure the transmission error coefficient of each test channel, and measuring the initial crosstalk coupling coefficient to establish a basic model of inter-channel crosstalk. This measurement process uses standard short-circuit, open-circuit, load, and through calibration components, and simultaneously excites adjacent channels and measures the coupling response through a four-port configuration, establishing a complete error feature library for the test system.

[0086] If the result of the first judgment node is "No", it means that the system is in normal testing state and no calibration is required. The process then enters the normal testing branch on the left. In this branch, the system executes the conventional sample testing procedure, reads the measurement results of the channel under test, and then performs two-level correction on the measurement data by combining the previously established error model and crosstalk coupling coefficient to generate the final target differential transmission parameters.

[0087] Regardless of whether it's the calibration branch or the testing branch, the process converges to the next processing stage, which specifically handles the measurement of crosstalk calibration frames. The system initiates the crosstalk calibration process module, calculating the current cumulative test frame count n plus 0. This addition of 0 actually retrieves the current frame count value. Next, the system judges the crosstalk calibration frames, comparing the cumulative frame count n with the set calibration interval G.

[0088] The process then proceeds to detailed iterative calculation steps. Step A reads Sdd21 and Sdd31 data from the VNA, which are the raw mixed-mode scattering parameters obtained from the vector network analyzer. Step B calculates the theoretical response raw(f) based on the acquired data and the current crosstalk coupling coefficient Hcouple. This calculation establishes a theoretical prediction of the calibrator's response based on the current state model of the test system. Step C calculates the error by subtracting the theoretical value raw(f) from the measured crosstalk coupling coefficient comp(f), obtaining the error function. The magnitude and phase characteristics of this error reflect the deviation between the current crosstalk coupling coefficient and the actual system state. Step D updates the crosstalk coupling coefficient using an adaptive filtering algorithm, multiplying the error signal by a step size factor μ and then adding it to the current coefficient to achieve gradient descent optimization. Step E increments the iteration counter n by 1, preparing for the next iteration.

[0089] After step E, the process enters the iteration termination condition judgment stage, which checks whether the current iteration number n is equal to the preset maximum iteration number Q. If the judgment result is "no", it means that further iteration and optimization are needed, and the process returns to step A to reread the data and calculate the error, forming a closed-loop feedback. If the judgment result is "yes", it means that a sufficient number of iterations have been completed or the error has converged to an acceptable range, and the process enters the result output stage on the right.

[0090] During the results output phase, the system performs a series of post-processing operations. Step F1 exports the filter parameters, saving the iteratively optimized crosstalk coupling coefficient as the target crosstalk coupling coefficient. Step F2 records the crosstalk calibration trigger time, providing a time reference for the next calibration cycle. Step F3 initializes the crosstalk calibration frame counter, resetting it to zero to prepare for the next test cycle. Step F4 uses the updated crosstalk coefficient to compensate for crosstalk in subsequent measurements, ensuring measurement accuracy. Step F5 executes data logging, persistently storing the calibration results and measurement data. Step F6 calculates the cumulative measurement progress and inputs it into the system, updating the test progress information. Step F7 organizes the data, preparing for the final report.

[0091] After all post-processing operations are completed, the process enters the final judgment node to determine whether all measurements are complete. If the judgment result is "No," it means there are still samples to be tested, and the process returns to the initial calibration cycle judgment node to continue executing the loop of "normal test - insert calibration frame - adaptive update - apply update coefficients." If the judgment result is "Yes," it means all test tasks have been completed, the process reaches the end node, and the entire test process terminates.

[0092] This flowchart, through the organic combination of decision nodes, processing modules, and iterative loops, realizes a complete algorithm for adaptive crosstalk calibration. Its core innovation lies in the periodic insertion of standard calibration components for measurement as a system state monitoring method. Error feedback between measured and theoretical responses drives the adaptive updating of the crosstalk coupling coefficient, enabling the test system to dynamically track and compensate for changes in crosstalk characteristics caused by time-varying factors such as temperature drift and device aging. This maintains stable high measurement accuracy during long-term continuous testing, effectively solving the problem of accuracy degradation over time in traditional static calibration methods.

[0093] Figure 3 This is a hardware connection architecture diagram of a differential high-speed signal testing system provided in an embodiment of this application, showing the connection relationship between the measuring instrument, the signal switching device and the test object.

[0094] The top layer of the system houses the core instrumentation area, containing a 4-port vector network analyzer (VNA). This instrument serves as the signal source and measurement core of the test system. Ports 1 and 2 form the first pair of differential ports, while ports 3 and 4 form the second pair. This four-port configuration enables the system to simultaneously excite and measure two independent differential channels, which is crucial for crosstalk measurements because it's necessary to monitor the coupling response of the other channel while excitation is being applied to one channel.

[0095] Ports 1 and 2 of the vector network analyzer are connected to switch matrix A of the dual-port switch matrix, and ports 3 and 4 are connected to switch matrix B. The dual-port switch matrix is ​​the core of the signal routing, internally employing a switching network composed of high-frequency relays to flexibly switch the test ports to different test objects according to instructions from the host computer. The dual-matrix design ensures that the positive and negative lines of the differential signal pair can switch synchronously, maintaining the integrity and signal balance of the differential channel.

[0096] The output of the switch matrix connects to the adapter panel area. The adapter panel provides standardized RF connectors, and the central differential measurement panel module enables signal distribution from the switch matrix to multiple test ports, branching out into multiple independent differential channel ports through multiple transmission lines.

[0097] The system's bottom layer is the load area, containing all test objects. On the far left is the differential crosstalk calibration device, a standard device with known crosstalk characteristics, containing four ports: positive and negative ports for the target channel and the crosstalk source channel. During crosstalk calibration, a switch matrix connects the four ports of the vector network analyzer to the four ports of the calibration device. The actual crosstalk response of the calibration device is obtained through four-port measurements, providing a reference for adaptively updating the crosstalk coupling coefficient.

[0098] To the right of the calibration kit are arranged the test samples DUT 1 to DUT n, which are the target objects whose differential transmission performance needs to be characterized, such as high-speed connectors, PCB differential pairs, or differential cable assemblies. Each DUT is connected via an adapter panel, but at any given time only the selected DUT is connected to the vector network analyzer via the switch matrix.

[0099] During system operation, the host computer controls the switch matrix to switch to the designated test object. When measuring the DUT, the matrix connects ports 1 and 2 to the differential positive and negative terminals of the DUT. The vector network analyzer transmits a signal from port 1, which passes through the DUT and is received at port 2, completing the differential transmission parameter measurement. After measuring one DUT, the system switches to the next, sequentially completing the testing of all samples. During crosstalk calibration, the matrix establishes a four-port configuration. Ports 3 and 4 input excitation to the crosstalk source channel of the calibration component, while ports 1 and 2 monitor the coupling response of the target channel to obtain the crosstalk transmission characteristics.

[0100] The core advantage of this architecture lies in its multi-channel multiplexing achieved through a switch matrix, allowing a single four-port instrument to serve multiple DUTs, thus reducing system costs. The differential crosstalk calibration component shares the same test path with the DUT, ensuring that calibration conditions match actual test conditions, enabling adaptive crosstalk correction based on the calibration component to accurately reflect the true system state. The modular design facilitates maintenance and upgrades, and standardized interfaces allow for easy replacement of different types of DUTs. The overall architecture provides a reliable hardware foundation for high-precision differential signal testing.

[0101] Figure 4 This is a schematic diagram illustrating the crosstalk propagation mechanism in a differential signal multi-channel system according to an embodiment of this application, showing the signal transmission path and crosstalk coupling relationship between two differential channels. Legend: Solid lines represent normal differential mode transmission, and dashed lines represent crosstalk coupling paths.

[0102] The diagram contains two differential channels. The differential input of Channel 1 on the left represents the signal input of the first differential channel, and the differential output of Channel 1 in the upper right represents the signal output of that channel. The input and output of Channel 1 are connected by a solid line labeled "Main Transmission," indicating that the differential signal propagates normally along the designed transmission path within Channel 1; this is the intended transmission path for the signal. The differential input of Channel 2 in the middle and the differential output of Channel 2 in the lower right constitute the second differential channel, also connected by a solid line labeled "Main Transmission," representing the normal signal transmission path of Channel 2.

[0103] Crosstalk coupling paths are represented by dashed lines. A dashed line labeled "NEXT Backward Crosstalk" extends from the differential input of channel 1 to the differential input of channel 2. NEXT stands for Near-End Crosstalk. This path indicates that the signal at the input of channel 1 leaks to the input of channel 2 through electromagnetic coupling, causing interference at the same end. Similarly, a dashed line labeled "NEXT Backward Crosstalk" also extends from the differential input of channel 2 to the differential input of channel 1, indicating that the crosstalk is bidirectional.

[0104] A dashed line labeled "FEXT Forward Crosstalk" extends from the differential input of channel 1, pointing towards the differential output of channel 2. FEXT stands for Far-End Crosstalk. This path indicates that the signal at the input of channel 1 will couple to the output of channel 2 after traveling a certain distance, causing cross-end interference. Similarly, a dashed line labeled "FEXT Forward Crosstalk" also extends from the differential input of channel 2, pointing towards the differential output of channel 1.

[0105] This diagram clearly illustrates the two main forms of crosstalk in a multi-channel differential system. Near-end crosstalk (NEXT) occurs at the same end of the signal, with the interfering signal in the opposite direction to the excitation signal, and is typically more significant at high frequencies. Far-end crosstalk (FEXT) occurs at the opposite end of the signal, with the interfering signal in the same direction as the excitation signal, and its intensity accumulates with increasing transmission distance. In actual testing, when a signal is transmitted through channel 1, both the input and output of channel 2 will be affected by crosstalk. These crosstalk components will be superimposed on the signal of channel 2 itself, leading to distorted measurement results.

[0106] This diagram illustrates why crosstalk calibration and compensation are necessary. The system quantifies the transmission strength of these dashed paths by measuring the crosstalk coupling coefficient. Then, in actual measurements, it calculates the crosstalk effect based on the signals and crosstalk coefficients of adjacent channels. Subtracting these interference components from the measurement results yields the true transmission characteristics of each channel. This crosstalk compensation mechanism is a key technological foundation for achieving high-precision multi-channel differential signal testing.

[0107] Figure 5 This is a flowchart of differential test channel calibration and crosstalk matrix establishment provided in an embodiment of this application, which shows how the error model establishment of each differential channel and the crosstalk coefficient measurement between channels are completed sequentially during the system initialization phase.

[0108] After the process starts at the initial node, it first enters the system warm-up phase, executing the offline hybrid mode calibration mode. System warm-up allows the test equipment to reach a stable operating temperature. Entering the offline hybrid mode calibration mode means that the system will perform standard multi-port error calibration on the test channels, establishing the measurement capability of hybrid mode parameters (differential mode and common mode).

[0109] After warm-up, the process enters the initialization step, setting the differential logic port index i to 1. This index is used to traverse all differential channels in the system, ensuring that each channel undergoes a complete calibration process. A differential logic port refers to the system-defined differential test channel number, typically corresponding to a physical pair of positive and negative signal lines.

[0110] Next, we proceed to the decision node to determine whether the current port i is a valid differential port. This decision is necessary because the test system may be configured with multiple physical ports, but not all ports are defined for use as differential test channels, or some ports may be disabled due to faults. The decision is typically based on the port definition table in the system configuration file or the port status detection results.

[0111] If the judgment result is "yes", it means that the current port is a valid differential channel, and the process enters the left branch to start the calibration process of that channel. First, the switch matrix is ​​switched to the i-th differential port. The host computer sends a control command to activate the relay of the switch matrix, connecting the test port of the vector network analyzer to the positive and negative signal lines of the i-th differential channel.

[0112] After establishing the connection, the process proceeds to the VNA to perform a mixed-mode SOLT full-port calibration step. SOLT is an abbreviation for Short-Open-Load-Thru, representing four standard calibration devices: short-circuit, open-circuit, load, and through. The operator connects these four calibration devices sequentially to the ports of the differential channel, and the vector network analyzer measures the mixed-mode scattering parameters of each device. These mixed-mode parameters include differential-to-differential transmission coefficients and common-mode-to-differential conversion coefficients, comprehensively characterizing the transmission and mode-switching characteristics of the differential channel.

[0113] After completing the calibration measurements, the process proceeds to the step of saving the hybrid mode calibration coefficients for that port. The host computer calculates the error coefficients for the differential channel using an error model algorithm based on the measured values ​​and standard values ​​of the four calibration components. These coefficients include multiple parameters such as directivity error, source matching error, reflection tracking error, and transmission tracking error. These error coefficients constitute the error model for that channel, are saved in memory, and associated with port index i.

[0114] After saving the error model, the process executes an auto-increment operation on the port index, where i equals i plus 1, and then returns to the decision node to repeat the calibration process for the next differential channel. This loop continues until all valid differential ports have completed SOLT calibration and established their respective error models.

[0115] When a node encounters a port index i that is not a valid differential port, the result is "No," indicating that all differential channels requiring calibration have been processed. The process then proceeds to the right branch, executing the SOLT calibration completion step for all differential ports, marking the end of the single-channel error model establishment phase.

[0116] The next step in the process is to measure the crosstalk parameters between all differential channel combinations. The system needs to measure the crosstalk coupling coefficient between each pair of adjacent differential channels, which requires establishing a four-port test configuration. The host computer controls the switch matrix to sequentially establish connections between different channel pairs, connecting ports 1 and 2 of the vector network analyzer to a differential channel as a crosstalk receiver, and ports 3 and 4 to adjacent differential channels as crosstalk transmitters. The vector network analyzer inputs differential excitation to the crosstalk source channel through ports 3 and 4, while simultaneously measuring the coupling response of the target channel through ports 1 and 2, recording the complete four-port scattering parameter matrix.

[0117] After crosstalk measurements for all channel pairs are completed, the process proceeds to the step of calculating and generating the initial crosstalk coupling matrix. The host computer performs mode conversion on the measurement data, converting single-ended parameters into differential mode parameters, and then calculates the crosstalk coupling coefficients between each pair of channels, i.e., the transfer function between the differential mode coupling response and the differential mode excitation signal. These crosstalk coefficients are organized into a matrix structure, where rows and columns correspond to different differential channels, and matrix elements represent the crosstalk strength between the corresponding channel pairs. The initial crosstalk coupling matrix fully describes the inter-channel crosstalk network characteristics of the test system in its initial state.

[0118] The final step of the process is to store the crosstalk matrix and establish a crosstalk baseline. The host computer saves the initial crosstalk coupling matrix to non-volatile memory as the system's crosstalk baseline reference. This baseline data will be used for comparison in subsequent adaptive crosstalk calibration. When a differential crosstalk calibrator is inserted for measurement, the system compares the measured crosstalk response of the calibrator with the response predicted based on this baseline and updates the crosstalk coupling coefficients according to the deviation. After the matrix storage is completed, the process reaches the end node, and the entire initialization calibration process is completed.

[0119] This flowchart illustrates the two core tasks of system initialization: first, establishing error models for each differential channel through SOLT standard calibration to eliminate transmission errors within a single channel; and second, establishing a crosstalk coupling matrix between channels through four-port crosstalk measurement to quantify the coupling characteristics of the multi-channel system. These two sets of parameters provide the mathematical basis for error de-embedding and crosstalk compensation in subsequent testing, and are prerequisites for achieving high-precision differential signal testing.

[0120] Figure 6This is a complete test flowchart of adaptive crosstalk calibration for differential high-speed signals provided in the embodiments of this application. It shows the entire loop process from system startup to the completion of testing of all samples, including a complete closed loop of normal testing, periodic crosstalk calibration and adaptive coefficient update.

[0121] After the process starts from the start node, the initialization of the measurement counter n is performed, which is equal to 0. This counter is used to track the number of completed measurement frames and serves as the basis for triggering crosstalk calibration.

[0122] The next step in the process is to switch the switch matrix to the differential channel of the target DUT. The host computer controls the switch matrix to connect the vector network analyzer to the differential channel of the current test sample and establish a test connection.

[0123] Then, the VNA performs differential mode parameter measurement and raw data acquisition steps. The vector network analyzer inputs a test signal to the differential channel under test, measures differential transmission parameters such as Sdd21, and obtains raw measurement data including test channel error and crosstalk effects.

[0124] After measurement, the process proceeds to the steps of crosstalk source extraction and calculation of valid measurement results. The host computer performs error de-embedding on the original data according to the error model of the differential channel, eliminating transmission errors in the test channel. Simultaneously, it calculates the crosstalk influence based on the crosstalk coupling coefficient and the signals of adjacent channels, subtracts the crosstalk component from the de-embedding data, and obtains the corrected target differential transmission parameters.

[0125] Next, the steps of storing data and updating the UI interface in real time are performed. The host computer saves the calibrated measurement results to the data file and displays the test progress and parameter curves in real time on the user interface, which makes it easy for operators to monitor the test process.

[0126] Then, the counter is incremented by n, which equals n plus 1, and a measurement is recorded as complete.

[0127] The process enters the first decision node, which checks whether the remainder of n divided by 15 is equal to Q. Here, 15 represents the insertion interval of the crosstalk calibration frame, and Q is a specific trigger value. This decision implements a periodic triggering mechanism, for example, inserting crosstalk calibration once after every 15 samples are tested.

[0128] If the result is "No", it means the calibration cycle has not yet been reached. The process skips the calibration step and directly proceeds to the second judgment node to determine whether all DUT channels have been completed. If there are still samples to be tested, the process returns to the switch matrix step to continue measuring the next sample.

[0129] If the first judgment result is "yes", it means that the calibration cycle has been reached, and the process enters the crosstalk calibration branch on the right. First, the step of switching to online crosstalk calibration mode is executed, and the system switches from normal test mode to crosstalk calibration mode.

[0130] Then, the steps of pausing DUT measurement and switching to the calibration device are executed. The host computer controls the switch matrix to disconnect the current test sample and switch to the differential crosstalk calibration device to establish a four-port test configuration.

[0131] Next, the full-channel crosstalk matrix is ​​measured, and the differential-mode mixed parameters are executed. The vector network analyzer simultaneously excites the crosstalk source channels of the calibration device through four ports and measures the response of the target channel to obtain the actual crosstalk characteristic data of the calibration device under the current test system state.

[0132] After measurement, the system proceeds to the steps of calculating health and updating the database. The host computer calculates the theoretical target channel response based on the known crosstalk characteristics of the calibration component and the crosstalk source channel response, and compares this calculation with the measured response to obtain the error. Adaptive iteration is performed using the crosstalk source response as input and the error as feedback to update the crosstalk coupling coefficient until the error converges. The updated target crosstalk coupling coefficient is saved to the database, replacing the previous coefficient for crosstalk compensation in subsequent measurements. The health assessment reflects the stability of the test system's crosstalk characteristics.

[0133] Finally, the DUT measurement mode recovery step is executed, the system returns from crosstalk calibration mode to normal test mode, the switch matrix is ​​reconnected to the test sample, and the test process continues.

[0134] After the calibration branch is completed, the process proceeds to the second decision node to determine whether all DUT channels have been tested. If the decision result is "No," it means there are still untested samples, and the process returns to the switching matrix step to continue testing using the updated crosstalk coupling coefficient. If the decision result is "Yes," it means all samples have been tested, and the process reaches the end node.

[0135] The core feature of this flowchart is the seamless integration of crosstalk calibration into the normal testing process. Periodic triggering is achieved through counters and modulus judgments. The system periodically inserts calibration frames during continuous testing, automatically updating the crosstalk coefficients without manual intervention. Each updated coefficient is immediately applied to subsequent measurements, ensuring that crosstalk compensation is always based on the current true state of the test system. This effectively addresses the effects of time-varying factors such as temperature drift and device aging, achieving long-term stable and high-precision measurements.

[0136] Figure 7 This is a detailed flowchart of an embodiment of the present application that provides an adaptive filtering algorithm for updating the crosstalk coupling coefficient. It shows how the crosstalk coupling coefficient can be gradually converged to the optimal value that can accurately describe the current state of the test system through iterative optimization.

[0137] After the process starts from the start node, the first step is to switch the switch matrix to the differential crosstalk standard calibration kit. The host computer controls the switch matrix to establish a four-port test configuration, connecting ports 1 and 2 of the vector network analyzer to the positive and negative ends of the target channel of the differential crosstalk calibration kit, and ports 3 and 4 to the positive and negative ends of the crosstalk source channel.

[0138] Next, the VNA is used to acquire the NEXT / FEXT between channels and obtain the full crosstalk matrix. The vector network analyzer inputs differential excitation signals to the crosstalk source channels of the calibration device through ports 3 and 4, while simultaneously measuring the coupling response of the target channel through ports 1 and 2, and reading back the actual signals of the crosstalk source channels through ports 3 and 4. The system measures the complete four-port scattering parameter matrix of near-end crosstalk (NEXT) and far-end crosstalk (FEXT) to obtain the full crosstalk transmission characteristics of the calibration device.

[0139] Then, the step of reading the current value matrix w_i is executed. The host computer retrieves the current crosstalk coupling coefficient matrix from memory. This matrix may be the baseline value obtained from the initial measurement, or the value updated in the previous iteration. This current value serves as the starting point for the iterative algorithm.

[0140] Next, the initialization of iteration parameters and the setting of differential values ​​are performed. The host computer sets the control parameters of the adaptive algorithm, including the step size factor μ to control the convergence speed, the maximum number of iterations Q to prevent infinite loops, and the convergence threshold to determine whether the error is small enough. At the same time, the iteration counter and the error accumulation variable are initialized.

[0141] Then, the core calculation step begins, calculating the difference between the expected output and the actual standard output to obtain the error signal e(n). The host computer, based on the known crosstalk characteristics of the differential crosstalk calibrator and the measured crosstalk source channel response, combined with the current crosstalk coupling coefficient w_i, calculates the theoretical target channel response through complex multiplication. This theoretical response represents the target channel signal that the calibrator should produce under the system state described by the current coefficients. The host computer subtracts the theoretical response from the actual measured target channel response to obtain the error signal e(n). The error signal is a complex frequency domain function, and its amplitude and phase reflect the degree of deviation of the current crosstalk coupling coefficient.

[0142] After obtaining the error signal, the NLMS algorithm's normalized step size calculation step is executed. NLMS is an abbreviation for Normalized Least Mean Square. This algorithm introduces a normalization factor based on the traditional LMS algorithm, enabling adaptive adjustment of the step size and improving the algorithm's stability when signal power changes. The host computer calculates the normalized step size using the formula μ divided by the input signal power plus a small positive number to prevent division by zero.

[0143] Then, the iterative update step of the crosstalk filter weights w_i is performed. The host computer updates the crosstalk coupling coefficients according to the core formula of the adaptive filtering algorithm. The calculation formula is that w_i+1 equals w_i plus the normalized step size multiplied by the product of the error signal e(n) and the conjugate of the input signal. This update formula adjusts the coefficients along the error gradient descent direction, so that the adjusted coefficients can more accurately predict the response of the calibrated component. The updated coefficient w_i+1 becomes the new current value.

[0144] After an update is completed, the process enters the judgment node, which determines whether the error e(n) is less than a preset threshold. The host computer calculates the root mean square value or maximum amplitude of the error signal and compares it with the preset threshold. The preset threshold is set according to the measurement accuracy requirements, for example, requiring the error to be less than 1% of the measured signal.

[0145] If the judgment result is "no", it means that the error is still large and the coefficients have not yet converged to the optimal value. The process returns to the step of calculating the difference between the expected output and the actual standard output. The system uses the updated coefficients w_i+1 to recalculate the theoretical response and error for the next iteration. This loop continues, and the crosstalk coupling coefficient is gradually adjusted in each iteration, so that the error gradually decreases.

[0146] If the judgment result is "yes", it means that the error has converged to an acceptable range or the maximum number of iterations has been reached, and the algorithm terminates. The process executes the convergence and locking of the final weight matrix. The host computer marks the final crosstalk coupling coefficient matrix as converged and locks it.

[0147] The next step is to calculate the channel health index. The host computer assesses the degree of change in the crosstalk coupling coefficient relative to the initial baseline and calculates the health index. The health index reflects the degree of drift in the crosstalk characteristics of the test system. A small change indicates that the system is stable, while a large change may indicate that maintenance or a complete recalibration is required.

[0148] Then, the step of writing the new values ​​and health status to the database is executed. The host computer saves the converged target crosstalk coupling coefficient and health status index to the database, replacing the previous values. These updated coefficients will be used for crosstalk compensation in subsequent normal tests to ensure that the compensation is based on the current real state of the test system.

[0149] Finally, the process reaches the end node, completing a full adaptive update process for the crosstalk coupling coefficient.

[0150] Figure 8 This is a schematic diagram of an adaptive crosstalk calibration time axis provided in an embodiment of this application, illustrating how normal sample testing and periodic crosstalk calibration are alternated during long-term continuous testing, and how the crosstalk coupling coefficient is dynamically updated over time to track the time-varying characteristics of the test system.

[0151] The horizontal axis of the graph represents the timeline, progressing from left to right. Above the timeline are a series of test events, labeled from left to right: Test Point 1, Test Point 2, Save Point (task count + cumulative coefficient mapping), Test Point 3, and the fault interruption (power outage / temperature interruption) on the far right. These event points represent critical moments in the test process.

[0152] Test points 1, 2, and 3 represent the normal sample testing phase, during which the system sequentially measures the differential transmission parameters of each sample according to the standard procedure. During these phases, the system uses the current crosstalk coupling coefficient to perform error de-embedding and crosstalk compensation on the measurement data, generating the corrected target differential transmission parameters.

[0153] A savepoint event occurred between test point 2 and test point 3, which is the trigger moment for the crosstalk calibration frame periodically inserted by the system. At this moment, the system pauses normal sample testing and switches to the differential crosstalk calibrator for measurement. The system collects the target channel response and the crosstalk source channel response of the calibrator and executes an adaptive iterative algorithm to update the crosstalk coupling coefficient. The savepoint event is marked with "task count + cumulative coefficient mapping", indicating that the system not only updates the crosstalk coefficient at this time, but also records the current test task progress and coefficient mapping relationship, and persists this data to the database.

[0154] A dashed line extends downwards from the save point, pointing to the lower right corner, from the frame annotation point of test point 3. The line is labeled "Repair Link (Dashed Path)," indicating that this is a logical connection path rather than a physical transmission path. This dashed line illustrates the evolution of the crosstalk coupling coefficient, showing that the crosstalk coefficient updated from the save point will be used in subsequent measurements at test point 3 and beyond.

[0155] A vertical arrow in the lower right corner points to the "Load Last Save Point" label, which corresponds to the fault recovery mechanism. When the system experiences a fault interruption after test point 3, such as a sudden power outage or ambient temperature exceeding the operating range causing the system to pause, the test process is forced to stop. After restarting, the system can load the data from the most recent save point from the database, including the test progress, crosstalk coupling coefficient, and system status information at that time, resuming the test from the interrupted position without having to start from the beginning.

[0156] The upper right corner of the diagram is marked with an arrow indicating "System Restart," signifying the system restart action after a failure. After restarting, the system executes the operation of loading the last save point, forming a complete closed loop from failure to recovery.

[0157] The lower left corner of the diagram contains a key data frame labeled "Key Databases: 1. Task Progress Table 2. Update Parameter Table." This indicates that the system uses two core data tables to support adaptive calibration and fault recovery functions. The Task Progress Table records the execution status of each test task, the number of completed measurements, and the current test location, enabling the system to track test progress. The Update Parameter Table stores the updated crosstalk coupling coefficient, error convergence status, and system health indicators after each crosstalk calibration, preserving historical evolution data of the crosstalk coefficient.

[0158] This timeline diagram clearly illustrates the dynamic process of adaptive crosstalk calibration. At the start of the test (test points 1 and 2), the system uses the initially measured crosstalk coupling coefficients. As time progresses, crosstalk characteristics may drift when conditions such as system temperature and device states change. The system periodically inserts calibration frames (save points) to detect and correct this drift, generating updated crosstalk coupling coefficients. These updated coefficients are immediately applied to subsequent tests (test point 3 and beyond), ensuring that crosstalk compensation is always based on the current system state.

[0159] The dashed path vividly illustrates the propagation and application relationship of crosstalk coefficients. The coefficients updated from the save point propagate backward along the time axis, affecting the crosstalk compensation accuracy of all subsequent measurements. The mechanism of loading the last save point demonstrates the robustness of the system; even in the event of an unexpected interruption, it can recover from the most recent stable state, avoiding the waste of time on repeated tests and the loss of crosstalk calibration state.

[0160] This design of periodic updates and state preservation enables the test system to maintain stable high accuracy during continuous testing for hours or even days, effectively addressing the impact of environmental changes and system aging on crosstalk characteristics, and achieving true long-term stable measurement capability.

[0161] Figure 9 This is a rack deployment diagram of a differential high-speed signal testing system provided in an embodiment of this application, illustrating the connection relationships and signal flow between the various functional modules of the system. The entire system adopts a layered architecture, from the control layer to the measurement layer and then to the switching layer, realizing automated testing and adaptive crosstalk calibration functions.

[0162] At the top of the diagram is the host computer (PC / server) module, which is the control center and data processing core of the entire testing system. The host computer runs the test software, responsible for arranging the test process, controlling the actions of each hardware module, acquiring and processing measurement data, executing error de-embedding and crosstalk compensation algorithms, and generating test reports. The host computer connects to the vector network analyzer and switch matrix via communication interfaces such as GPIB, LAN, or USB, sending control commands and reading measurement results.

[0163] The host computer is connected downwards to the Vector Network Analyzer (VNA) module via a solid arrow, indicating bidirectional transmission of control commands and data. The Vector Network Analyzer is the core measuring instrument of the system, capable of generating high-frequency test signals and precisely measuring the amplitude and phase of reflected and transmitted signals. It supports a four-port measurement configuration to characterize differential parameters and crosstalk characteristics.

[0164] The host computer is connected to the calibration module on the right side via a dashed arrow. The dashed line indicates a logical connection rather than a physical signal connection, meaning the host computer stores the standard parameter data of the calibration module. When executing the adaptive crosstalk calibration algorithm, these standard values ​​are compared with the measured values. The calibration modules include two types: one is the SOLT standard calibration module (short circuit, open circuit, load, shoot-through) used to establish the error model of the test channel; the other is the differential crosstalk calibration module used for periodic measurements to update the crosstalk coupling coefficient.

[0165] The vector network analyzer is connected downwards to the switch matrix chassis module via a solid arrow, indicating the transmission path of the RF signal. The four test ports of the vector network analyzer are connected to the input ports of the switch matrix via high-quality RF cables. The switch matrix routes these ports to different output ports according to instructions from the host computer.

[0166] The vector network analyzer is connected to the DUT (Device Under Test) module on the right side via a solid arrow, but this connection is actually relayed through a switch matrix, indicating that the measurement signal is ultimately transmitted to the sample under test. The DUT is the differential device to be characterized, such as a high-speed connector, PCB differential pairs, or differential cable assembly.

[0167] The switch matrix chassis also features a solid arrow pointing to the DUT module on the right, more accurately illustrating the signal path: the test signal originates from the vector network analyzer, is routed through the switch matrix, and finally reaches the selected DUT for measurement. The switch matrix enables one-to-many connection multiplexing, allowing a single vector network analyzer to serve multiple test objects.

[0168] Although the calibration module in the upper right corner is drawn independently in physical location, it is actually connected to the vector network analyzer via a switch matrix during actual use. When the system needs to perform crosstalk calibration, the switch matrix is ​​switched to the calibration module to establish a four-port measurement configuration. After calibration is completed, it is switched back to the DUT to continue normal testing.

[0169] The overall system workflow is as follows: The host computer sends commands to control the switch matrix to switch to the specified test object, and simultaneously configures the measurement parameters of the vector network analyzer. The vector network analyzer outputs a test signal, which is routed through the switch matrix to the DUT or calibration device. The measurement results are returned to the vector network analyzer, and then read and processed by the host computer. The host computer corrects the original measurement data based on the stored error model and crosstalk coupling coefficient, generating the final target differential transmission parameters.

[0170] This rack-mount deployment scheme achieves modularity and scalability of the test system. The various functional modules are interconnected through standard interfaces, facilitating maintenance and upgrades. Software control by the host computer automates the entire testing process, reducing manual operation and improving testing efficiency and consistency. The introduction of a switch matrix enables resource sharing and reduces system costs. The integration of standard calibration components provides a reference benchmark for adaptive crosstalk calibration, ensuring long-term accuracy and stability. The overall architecture provides a complete hardware and software support platform for high-precision, high-efficiency differential high-speed signal testing.

[0171] Figure 10 This is a software architecture diagram of a differential high-speed signal adaptive crosstalk calibration test system provided in an embodiment of this application, illustrating the layered design of the host computer software and the interaction relationships between the modules. The entire architecture adopts a classic four-layer structure, from top to bottom: UI layer, business logic layer, device abstraction layer, and data persistence layer.

[0172] The top layer is the UI layer (QWidgets), responsible for interface display, human-computer interaction, parameter configuration, and result visualization. This layer uses the QWidgets component of the Qt framework to build a graphical user interface, providing operators with an intuitive test control panel. Users can configure test parameters such as frequency range, power level, and crosstalk calibration interval through the interface, start or stop the test process, view measurement progress and parameter curves in real time, and export test reports. The UI layer transmits user operation commands to the business logic layer through calls or command issuance, while simultaneously receiving data feedback and status reports from the business logic layer and the data persistence layer, dynamically updating the interface display content.

[0173] The middle layer is the business logic layer, containing two core functional modules. The test scheduling module on the left is responsible for process control and task scheduling, orchestrating test sequences to determine when to measure the DUT, when to insert calibration frames, and how to traverse multiple test channels. It manages the test state machine to ensure the process executes in the predetermined order, monitors test progress, and triggers appropriate processing upon completion or exception. The calibration engine module on the right is responsible for algorithm calculation and error calibration. It implements an error de-embedding algorithm to eliminate transmission errors in test channels based on the error model, executes a crosstalk compensation algorithm to remove crosstalk effects between channels based on the crosstalk coupling coefficient, and runs an adaptive iterative algorithm to update the crosstalk coupling coefficient to track the time-varying characteristics of the system. These two modules work together: the test scheduling module controls the execution order of the process, and the calibration engine module precisely corrects the data from each measurement, jointly achieving high-precision adaptive testing functionality.

[0174] The business logic layer interacts downwards with the device abstraction layer (VISA / serial port). The device abstraction layer encapsulates the hardware interface, providing a unified device operation API to the upper layers and shielding them from the communication protocol differences between various hardware devices. This layer interacts with the vector network analyzer via the VISA communication protocol, sending SCPI commands to configure measurement parameters, trigger measurements, and read measurement data. It also communicates with the switch matrix via the RS232 serial port, sending switching commands to control relay actions to route test ports to specified DUTs or calibration pieces. The device abstraction layer is also responsible for device driver initialization, connection status monitoring, and error handling to ensure the normal operation of the hardware devices.

[0175] Both modules in the business logic layer interact with the data persistence layer (SQLite). The data persistence layer uses the lightweight SQLite database to implement test data storage, configuration saving, log reading / writing, and data querying functions. This layer stores historical versions of calibration data, including error model coefficients and crosstalk coupling coefficients; saves test results, including differential transmission parameters for each DUT at various frequency points; records test progress information for fault recovery; and saves user configuration parameters for repeated testing. The test scheduling module reads the progress and configuration of the last test from the database, writes the current state to the database during the test, and can recover from the most recent save point in case of unexpected system interruption. The calibration engine module reads the current error model and crosstalk coupling coefficients from the database for data correction and writes the adaptively updated coefficients to the database for subsequent measurements.

[0176] The data persistence layer provides data query services to the UI layer and business logic layer through the data return path, and pushes system status information to the UI layer through the status reporting path, so that the user interface can reflect the test progress and system health in real time.

[0177] The entire software architecture achieves module decoupling and separation of responsibilities through a clear layered design. The UI layer focuses on user experience, the business logic layer implements core algorithms and process control, the device abstraction layer isolates hardware differences, and the data persistence layer provides reliable storage. Well-defined interfaces between the layers ensure good maintainability and scalability. When new test instruments need to be supported, only the device abstraction layer needs to be extended; when calibration algorithms need to be optimized, only the calibration engine module needs to be modified, without affecting other modules. This architecture provides a robust software foundation for complex adaptive crosstalk calibration test systems, supporting the entire test process from user interaction to hardware control to data processing.

[0178] Based on the above method, this application also discloses an adaptive crosstalk calibration system for differential high-speed signals, such as... Figure 11 As shown, Figure 11 This is a schematic diagram of an adaptive crosstalk calibration system for differential high-speed signals provided in an embodiment of this application. The system includes: a control module, a test module, an insertion module, an input module, and a calibration module. The control module controls a switch matrix to sequentially switch to each differential test channel, calibrates each differential test channel, generates an error model for each differential test channel, and measures the crosstalk coupling coefficient between each differential test channel. The test module, when entering test mode, controls the switch matrix to sequentially switch to each differential channel of the sample under test, measures the differential transmission parameters of each differential channel using a vector network analyzer, and combines the error model and crosstalk coupling coefficient to calibrate the differential transmission... The system performs parameter correction to generate target differential transmission parameters; an insertion module periodically inserts crosstalk calibration frames in the test mode. Within the crosstalk calibration frame, the control switch matrix switches to the differential crosstalk calibrator, and the target channel response and crosstalk source channel response of the differential crosstalk calibrator are measured by a vector network analyzer; an input module uses the crosstalk source channel response as the input signal and the theoretical target channel response of the differential crosstalk calibrator as the desired signal. Based on the error between the target channel response and the theoretical target channel response, the crosstalk coupling coefficient is updated to generate the target crosstalk coupling coefficient; a correction module corrects the subsequently measured differential transmission parameters using the target crosstalk coupling coefficient.

[0179] Please see Figure 12 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 12 As shown, the electronic device 1000 may include: at least one processor 1001, at least one network interface 1004, a user interface 1003, a memory 1005, and at least one communication bus 1002.

[0180] The communication bus 1002 is used to realize the connection and communication between these components.

[0181] The user interface 1003 may include a display screen and a camera. Optionally, the user interface 1003 may also include a standard wired interface and a wireless interface.

[0182] The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0183] The processor 1001 may include one or more processing cores. The processor 1001 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 1005, and by calling data stored in the memory 1005. Optionally, the processor 1001 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 1001 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content to be displayed on the screen; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 1001 and may be implemented as a separate chip.

[0184] The memory 1005 may include random access memory (RAM) or read-only memory. Optionally, the memory 1005 may include a non-transitory computer-readable storage medium. The memory 1005 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 1005 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 1005 may also be at least one storage device located remotely from the aforementioned processor 1001. Figure 12 As shown, the memory 1005, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an application program for an adaptive crosstalk calibration method for differential high-speed signals.

[0185] exist Figure 12 In the electronic device 1000 shown, the user interface 1003 is mainly used to provide an input interface for the user and to obtain the user input data; while the processor 1001 can be used to call the application program of an adaptive crosstalk calibration method for differential high-speed signals stored in the memory 1005. When executed by one or more processors, the electronic device performs one or more of the methods described in the above embodiments.

[0186] An electronic device readable storage medium stores instructions that, when executed by one or more processors, cause the electronic device to perform one or more of the methods described in the above embodiments.

[0187] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0188] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0189] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some service interfaces; indirect couplings or communication connections between devices or units may be electrical or other forms.

[0190] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0191] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0192] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0193] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Other embodiments of this disclosure will be readily apparent to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described herein. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. An adaptive crosstalk calibration method for differential high-speed signals, characterized in that, This method is applied to a host computer, which communicates with a vector network analyzer and a switch matrix. The test ports of the vector network analyzer are connected to the positive and negative signal lines of multiple differential test channels via the switch matrix. The control switch matrix is ​​sequentially switched to each differential test channel to calibrate each differential test channel, generate an error model for each differential test channel, and measure the crosstalk coupling coefficient between each differential test channel. When entering the test mode, the control switch matrix sequentially switches to each measured differential channel of the sample under test. The differential transmission parameters of each measured differential channel are measured by the vector network analyzer. The differential transmission parameters are corrected by combining the error model and crosstalk coupling coefficient to generate the target differential transmission parameters. In the test mode, crosstalk calibration frames are periodically inserted. Within the crosstalk calibration frames, the control switch matrix is ​​switched to the differential crosstalk calibrator. The target channel response and crosstalk source channel response of the differential crosstalk calibrator are measured by a vector network analyzer. Using the crosstalk source channel response as the input signal and the theoretical target channel response of the differential crosstalk calibrator as the desired signal, the crosstalk coupling coefficient is updated based on the error between the target channel response and the theoretical target channel response, and the target crosstalk coupling coefficient is generated. The differential transmission parameters of subsequent measurements are corrected by using the target crosstalk coupling coefficient.

2. The adaptive crosstalk calibration method for differential high-speed signals according to claim 1, characterized in that, The control switch matrix sequentially switches to each differential test channel, calibrates each differential test channel, and generates an error model for each differential test channel, including: The control switch matrix connects the first test port of the vector network analyzer to the positive signal line of each differential test channel, and connects the second test port to the negative signal line of each differential test channel. Connect the short-circuit calibrator, open-circuit calibrator, load calibrator and through-circuit calibrator sequentially to the positive and negative signal lines of each differential test channel, and measure the mixed-mode scattering parameters of each calibrator using a vector network analyzer; Based on the mixed-mode scattering parameters of each calibrator and the standard values ​​of each calibrator, the error model of each differential test channel is calculated. The error model includes the error coefficient of differential mode transmission and the error coefficient of common-mode to differential mode conversion.

3. The adaptive crosstalk calibration method for differential high-speed signals according to claim 1, characterized in that, The measurement of the crosstalk coupling coefficient between each differential test channel includes: The control switch matrix connects the first and second test ports of the vector network analyzer to the positive and negative signal lines of each differential test channel, respectively, and connects the third and fourth test ports to the positive and negative signal lines of the adjacent differential test channels of each differential test channel, respectively. The coupling response of each differential test channel under the excitation of adjacent differential test channels is measured by simultaneously exciting each differential test channel and each differential test channel with a vector network analyzer. Based on the coupling response of each differential test channel and the excitation signal of the adjacent differential test channel, the crosstalk coupling coefficient between each differential test channel and the corresponding adjacent differential test channel is calculated.

4. The adaptive crosstalk calibration method for differential high-speed signals according to claim 3, characterized in that, The step of calculating the crosstalk coupling coefficient between each differential test channel and its corresponding adjacent differential test channel based on the coupling response of each differential test channel and the excitation signal of the adjacent differential test channel includes: The coupled response of each differential test channel is converted into a differential mode coupled response, and the excitation signal of adjacent differential test channels is converted into a differential mode excitation signal. The transfer function between the differential mode coupling response and the differential mode excitation signal is calculated to obtain the crosstalk coupling coefficient between each differential test channel and its corresponding adjacent differential test channel; the crosstalk coupling coefficient characterizes the degree of crosstalk influence of adjacent differential test channels on each differential test channel.

5. The adaptive crosstalk calibration method for differential high-speed signals according to claim 1, characterized in that, The step of correcting the differential transmission parameters by combining the error model and crosstalk coupling coefficient to generate the target differential transmission parameters includes: Based on the error model of each differential test channel, the differential transmission parameters of each measured differential channel are de-embedded to obtain the de-embedded differential transmission parameters. Calculate the crosstalk influence between each differential test channel based on the crosstalk coupling coefficient between each differential test channel. The target differential transmission parameters are obtained by removing crosstalk from the de-embedding differential transmission parameters; the target differential transmission parameters are differential transmission parameters that have eliminated the effects of test channel error and inter-channel crosstalk.

6. The adaptive crosstalk calibration method for differential high-speed signals according to claim 1, characterized in that, The control switch matrix is ​​switched to the differential crosstalk calibration device, and the target channel response and crosstalk source channel response of the differential crosstalk calibration device are measured by a vector network analyzer, including: The control switch matrix connects the first and second test ports of the vector network analyzer to the positive and negative terminals of the target channel of the differential crosstalk calibrator, respectively, and connects the third and fourth test ports to the positive and negative terminals of the crosstalk source channel of the differential crosstalk calibrator, respectively. The crosstalk source channel of the differential crosstalk calibrator is excited by a vector network analyzer, and the target channel response and crosstalk source channel response of the differential crosstalk calibrator are measured; the differential crosstalk calibrator is a standard part with known crosstalk characteristics.

7. The adaptive crosstalk calibration method for differential high-speed signals according to claim 1, characterized in that, The process of using the crosstalk source channel response as the input signal and the theoretical target channel response of the differential crosstalk calibrator as the desired signal, updating the crosstalk coupling coefficient based on the error between the target channel response and the theoretical target channel response, and generating the target crosstalk coupling coefficient includes: Based on the known crosstalk characteristics and crosstalk source channel response of the differential crosstalk calibrator, and combined with the current crosstalk coupling coefficient, the theoretical target channel response of the differential crosstalk calibrator is calculated. Calculate the error between the measured target channel response and the theoretical target channel response; Using the crosstalk source channel response as the input signal and the error as the feedback signal, the crosstalk coupling coefficient is iteratively updated. When the error converges to a preset threshold, the target crosstalk coupling coefficient is obtained. The target crosstalk coupling coefficient is used to compensate for the crosstalk drift caused by the test system with changes in time and environment.

8. An adaptive crosstalk calibration system for differential high-speed signals, characterized in that, The system includes: a control module, a testing module, an insertion module, an input module, and a calibration module; wherein, The control module is used to control the switch matrix to switch sequentially to each differential test channel, calibrate each differential test channel, generate an error model for each differential test channel, and measure the crosstalk coupling coefficient between each differential test channel. The test module is used to control the switch matrix to switch to each differential channel of the sample under test in sequence when entering the test mode, measure the differential transmission parameters of each differential channel under test by a vector network analyzer, and correct the differential transmission parameters by combining the error model and crosstalk coupling coefficient to generate target differential transmission parameters. The insertion module is used to periodically insert crosstalk calibration frames in the test mode. Within the crosstalk calibration frame, the control switch matrix is ​​switched to the differential crosstalk calibration device, and the target channel response and crosstalk source channel response of the differential crosstalk calibration device are measured by the vector network analyzer. The input module is used to take the crosstalk source channel response as the input signal, the theoretical target channel response of the differential crosstalk calibrator as the expected signal, update the crosstalk coupling coefficient according to the error between the target channel response and the theoretical target channel response, and generate the target crosstalk coupling coefficient. The correction module is used to correct the differential transmission parameters of subsequent measurements using the target crosstalk coupling coefficient.

9. An electronic device, characterized in that, The device includes a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer program is stored that can be loaded by a processor and executed as described in any one of claims 1-7.

Images