A method and system for evaluating the performance of a silicon carbide chip

Abstract

The application provides a performance evaluation method and evaluation system of a silicon carbide chip, relates to the field of semiconductor testing, separates a nonlinear disturbance component from dynamic response data; constructs a link transmission compensation function based on the nonlinear disturbance component, determines dynamic characteristic parameters through the link transmission compensation function and the dynamic response data; performs multi-dimensional spectrum analysis on the dynamic characteristic parameters to obtain high-frequency performance parameters, and then derives energy efficiency of the silicon carbide chip according to the high-frequency performance parameters to obtain energy efficiency trend and energy efficiency thermal coupling characteristics; deviates from abnormal values to perform multi-working-condition pulse excitation on the silicon carbide chip to obtain a multi-dimensional compensation matrix; and performs adaptive coupling compensation on the energy efficiency trend and the energy efficiency thermal coupling characteristics based on the multi-dimensional compensation matrix, and then obtains a performance evaluation data set of the silicon carbide chip. The application can realize multi-dimensional adaptive coupling compensation driven by abnormal values to improve the reliability of performance evaluation of the silicon carbide chip.

Description

A performance evaluation method and system for silicon carbide chips

[0001] This application relates to the field of semiconductor testing, and more specifically, to a performance evaluation method and system for silicon carbide chips. Background Technology

[0002] To ensure the safe and efficient operation of semiconductor devices in different application scenarios, semiconductor testing technology has gradually become a key link in the semiconductor industry chain. Common semiconductor tests include static parameter measurement, dynamic characteristic analysis, power loss assessment, and environmental adaptability testing. The purpose is to reproduce the actual working state of the device under experimental conditions and to quantitatively evaluate its performance through test results.

[0003] Silicon carbide (SiC) chips are a new generation of wide-bandgap semiconductor devices, possessing advantages such as high voltage resistance, high frequency, low on-resistance, and high thermal conductivity. They are widely used in new energy vehicles, power electronics, and smart grids. Performance evaluation of SiC chips typically includes measuring their static electrical parameters, testing switching losses under high-frequency conditions, and analyzing junction temperature and energy efficiency changes at different temperatures. This provides partial performance information for the chip under a single operating condition. However, existing SiC chip performance evaluation methods still have the following shortcomings: First, traditional tests often rely on filtering or averaging, which can easily confuse the chip's intrinsic nonlinear disturbances with test noise. Second, they lack comprehensive evaluation capabilities across operating conditions and multiple dimensions, making it difficult to test the true response of SiC chips in complex environments. Finally, the lack of adaptive mechanisms leads to insufficient consistency of evaluation results across power and temperature conditions. Therefore, how to achieve outlier-driven multidimensional adaptive coupling compensation to improve the reliability of SiC chip performance evaluation is a challenge facing the industry. Summary of the Invention

[0004] This application provides a performance evaluation method and system for silicon carbide chips, which can realize multi-dimensional adaptive coupling compensation based on outlier-driven methods to improve the reliability of silicon carbide chip performance evaluation.

[0005] In a first aspect, this application provides a performance evaluation method for a silicon carbide chip, the performance evaluation method comprising the following steps:

[0006] Collect dynamic response data of silicon carbide chip under stable operating conditions, and separate nonlinear disturbance components from the dynamic response data;

[0007] A link transmission compensation function is constructed based on the nonlinear disturbance component, and dynamic characteristic parameters are determined by the link transmission compensation function and the dynamic response data.

[0008] Multidimensional spectrum analysis is performed on the dynamic characteristic parameters to obtain high-frequency performance parameters. Then, based on the high-frequency performance parameters, the thermal efficiency of the silicon carbide chip is derived to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels.

[0009] The deviation anomaly value of the silicon carbide chip during the electromagnetic interference process is collected, and the silicon carbide chip is subjected to multi-condition pulse excitation based on the deviation anomaly value to obtain a multi-dimensional compensation matrix;

[0010] Based on the multidimensional compensation matrix, adaptive coupling compensation is performed on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels, thereby obtaining the performance evaluation dataset of silicon carbide chips.

[0011] In this embodiment, dynamic response data of the silicon carbide chip under stable operating conditions are collected through electromagnetic shielding testing.

[0012] In this embodiment, separating the nonlinear disturbance component from the dynamic response data specifically includes:

[0013] The dynamic response data is baseline-fitted to obtain the residual sequence;

[0014] The residual sequence is subjected to time-frequency joint analysis to obtain the nonlinear perturbation component.

[0015] In this embodiment, constructing the link transmission compensation function based on the nonlinear disturbance component specifically includes:

[0016] Construct the initial transfer function model for the silicon carbide chip test link;

[0017] The nonlinear disturbance component is mapped to the initial transmission function model for iterative correction, thereby obtaining the link transmission compensation function.

[0018] In this embodiment, determining the dynamic characteristic parameters through the link transmission compensation function and the dynamic response data specifically includes:

[0019] The dynamic response data is corrected for amplitude and phase frequency using the link transmission compensation function to obtain the compensation response signal;

[0020] The compensation response signal is subjected to joint time-domain and frequency-domain analysis, and then the switching transient characteristics, parasitic response characteristics and power loss characteristics of the silicon carbide chip are obtained by separation and extraction.

[0021] By analyzing the dynamic changing trends of the switching transient characteristics, the parasitic response characteristics, and the power loss characteristics, a set of dynamic characteristic parameters is obtained.

[0022] In this embodiment, the high-frequency performance parameters obtained by performing multi-dimensional spectral analysis on the dynamic characteristic parameters specifically include:

[0023] The dynamic characteristic parameters are subjected to Fourier transform and wavelet decomposition to obtain the joint time-frequency spectrum;

[0024] High-frequency harmonic components, parasitic oscillation components, and transient jitter characteristics are extracted from the time-frequency joint spectrum.

[0025] The high-frequency performance parameter set of the silicon carbide chip under high-frequency drive is determined by the high-frequency harmonic components, the parasitic oscillation components, and the transient jitter characteristics.

[0026] In this embodiment, the thermal efficiency of the silicon carbide chip is derived based on the high-frequency performance parameters, and the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels are obtained, specifically including:

[0027] The power loss of the silicon carbide chip at different power levels is determined by the aforementioned high-frequency performance parameters.

[0028] A structural thermal model of a silicon carbide chip is obtained, and then all power losses are input into the structural thermal model to obtain junction temperature change curves and power loss curves.

[0029] Based on the junction temperature change curve and the power loss curve, the energy efficiency trend under different power levels is derived.

[0030] Junction temperature fluctuations and energy efficiency degradation are extracted from all energy efficiency trends, and then the energy efficiency thermal coupling characteristics are determined through the junction temperature fluctuations and energy efficiency degradation.

[0031] In this embodiment, the deviation anomaly value of the silicon carbide chip during the electromagnetic interference process is collected through controlled electromagnetic interference testing.

[0032] In this embodiment, the multi-condition pulse excitation of the silicon carbide chip based on the deviation anomaly value to obtain the multi-dimensional compensation matrix specifically includes:

[0033] Based on the deviation from the outlier, multiple sets of pulse excitation condition tests were performed on the silicon carbide chip to obtain the coupling relationship between the response parameters and the abnormal characteristics of the silicon carbide chip.

[0034] A multidimensional compensation matrix is ​​constructed based on the aforementioned coupling relationship.

[0035] Secondly, this application provides a performance evaluation system for silicon carbide chips, used to execute a performance evaluation method for silicon carbide chips, the performance evaluation system comprising:

[0036] The data acquisition module is used to acquire dynamic response data of silicon carbide chips under stable operating conditions and separate nonlinear disturbance components from the dynamic response data.

[0037] The link compensation module is used to construct a link transmission compensation function based on the nonlinear disturbance component, and to determine dynamic characteristic parameters through the link transmission compensation function and the dynamic response data.

[0038] The spectrum analysis module is used to perform multi-dimensional spectrum analysis on the dynamic characteristic parameters to obtain high-frequency performance parameters, and then to deduce the thermal efficiency of the silicon carbide chip based on the high-frequency performance parameters to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels.

[0039] The operating condition excitation module is used to collect the deviation anomaly values ​​of the silicon carbide chip during the electromagnetic interference process, and to perform multi-operating condition pulse excitation on the silicon carbide chip based on the deviation anomaly values ​​to obtain a multi-dimensional compensation matrix.

[0040] The coupling compensation module is used to perform adaptive coupling compensation on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels based on the multi-dimensional compensation matrix, thereby obtaining the performance evaluation dataset of the silicon carbide chip.

[0041] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0042] Dynamic response data of a silicon carbide chip under stable operating conditions is collected, and nonlinear disturbance components are separated from the dynamic response data. A link transmission compensation function is constructed based on the nonlinear disturbance components, and dynamic characteristic parameters are determined by the link transmission compensation function and the dynamic response data. Multidimensional spectrum analysis is performed on the dynamic characteristic parameters to obtain high-frequency performance parameters. Then, the thermal efficiency of the silicon carbide chip is derived based on the high-frequency performance parameters to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels. Deviation anomalies of the silicon carbide chip during electromagnetic interference are collected, and multi-condition pulse excitation of the silicon carbide chip is performed based on the deviation anomalies to obtain a multidimensional compensation matrix. Adaptive coupling compensation is performed on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels based on the multidimensional compensation matrix to obtain a performance evaluation dataset of the silicon carbide chip.

[0043] Therefore, this application can improve the reliability of silicon carbide chip performance evaluation. First, by collecting dynamic response data of silicon carbide chips under stable operating conditions under electromagnetic shielding, and separating the nonlinear disturbance component, the test signal can accurately reflect the intrinsic characteristics of the chip, effectively avoiding the loss of operating condition-related information caused by traditional filtering methods, thereby improving the accuracy of dynamic characteristic extraction. Furthermore, a link transmission compensation function is constructed based on the nonlinear disturbance component, and this function is used to correct the dynamic response data, ensuring that the dynamic characteristic parameters can eliminate the distortion effects of the test link, thereby improving the consistency and reliability of cross-frequency band performance indicators. Second, multi-dimensional spectrum analysis is performed on the dynamic characteristic parameters to extract high-frequency harmonic components. This study identifies parasitic oscillation components and transient jitter characteristics, and combines these with a thermal model to derive the energy efficiency trend and energy efficiency-thermal coupling characteristics of the chip at different power levels. This comprehensive modeling of electrothermal characteristics helps to overcome the shortcomings of performance evaluation in a single domain. Then, under electromagnetic interference conditions, outlier values ​​of the silicon carbide chip are collected and used as inputs for multi-condition pulse excitation to construct a multi-dimensional compensation matrix. This effectively preserves the operating condition information contained in the outliers, avoiding the problem of missing evaluation information caused by directly removing outliers in traditional methods. Finally, based on the multi-dimensional compensation matrix, adaptive coupling compensation is performed on the energy efficiency trend and energy efficiency-thermal coupling characteristics at different power levels, enabling unified evaluation results across power, temperature, and interference conditions.

[0044] In summary, the technical solution adopted in this application can realize multi-dimensional adaptive coupling compensation based on outlier-driven methods to improve the reliability of silicon carbide chip performance evaluation. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only for this embodiment of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 is a flowchart of a performance evaluation method for a silicon carbide chip according to this application;

[0047] Figure 2 is an exemplary flowchart of determining dynamic characteristic parameters according to the present application;

[0048] Figure 3 is an exemplary flowchart for determining high-frequency performance parameters according to this application;

[0049] Figure 4 is a module structure diagram of a silicon carbide chip performance evaluation system provided in this application. Detailed Implementation

[0050] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0051] This application provides a performance evaluation method and system for silicon carbide chips. The core of the method involves collecting dynamic response data of the silicon carbide chip under stable operating conditions and separating nonlinear disturbance components from the dynamic response data. A link transmission compensation function is constructed based on the nonlinear disturbance components, and dynamic characteristic parameters are determined using the link transmission compensation function and the dynamic response data. Multidimensional spectrum analysis is performed on the dynamic characteristic parameters to obtain high-frequency performance parameters. Then, thermal efficiency is derived for the silicon carbide chip based on the high-frequency performance parameters to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels. Deviation anomalies of the silicon carbide chip during electromagnetic interference are collected, and multi-condition pulse excitation is applied to the silicon carbide chip based on the deviation anomalies to obtain a multidimensional compensation matrix. Adaptive coupling compensation is performed on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels based on the multidimensional compensation matrix, thereby obtaining a performance evaluation dataset for the silicon carbide chip.

[0052] Example 1: To better understand the above technical solution, the following will describe the above technical solution in detail with reference to the accompanying drawings and specific embodiments. Referring to Figure 1, this figure is an exemplary flowchart of a performance evaluation method for a silicon carbide chip according to this embodiment of the present application. The performance evaluation method includes the following steps:

[0053] In step S1, dynamic response data of the silicon carbide chip under stable operating conditions is collected, and nonlinear disturbance components are separated from the dynamic response data.

[0054] In this embodiment, electromagnetic shielding testing is used to collect dynamic response data of the silicon carbide chip under stable operating conditions. In practice, electromagnetic shielding testing involves arranging a silicon carbide chip test circuit within a metal shielding cavity and suppressing interference from external electromagnetic fields on the test signal to ensure that the collected signal primarily reflects the operating characteristics of the silicon carbide chip itself. The electromagnetic shielding test includes: placing the silicon carbide chip in the shielding cavity; providing excitation voltage and current through an external signal source; collecting transient signals at key nodes of the chip using high-speed voltage and current probes, and recording them in real time using a data acquisition card. Thus, dynamic response data of the silicon carbide chip under stable operating conditions can be collected through electromagnetic shielding testing.

[0055] It should be noted that stable operating conditions refer to the continuous and balanced operation of the silicon carbide chip under rated voltage, rated current, and constant ambient temperature. Dynamic response data refers to the transient information of the output voltage and current of the silicon carbide chip under different input excitations, which is used to characterize the electrical behavior of the silicon carbide chip during dynamic switching, conduction, and turn-off processes. The dynamic response data includes: switching transient waveforms, conduction region fluctuation curves, and turn-off signals. Among them, the switching transient waveforms are used to describe the rise and fall characteristics of current and voltage; the conduction region fluctuation curves are used to characterize the response offset caused by parasitic parameters; and the turn-off signals are used to characterize electromagnetic interference and nonlinear disturbance characteristics.

[0056] In this embodiment, the nonlinear disturbance component can be separated from the dynamic response data in the following manner:

[0057] The dynamic response data is baseline-fitted to obtain the residual sequence;

[0058] The residual sequence is subjected to time-frequency joint analysis to obtain the nonlinear perturbation component.

[0059] In practice, the dynamic response data is first de-biased (i.e., by subtracting the time average of the signal) and impulsive noise is removed using a median filter to obtain a preprocessed response signal. The curve obtained by Savitzky-Golay smoothing filter fitting is used as the baseline signal. The time series obtained by subtracting the preprocessed dynamic response data and the baseline signal point by point is used as the residual sequence. The residual sequence is subjected to short-time Fourier transform and continuous wavelet transform to obtain a short-time spectrum and a wavelet energy map. The mean of the median of the short-time spectrum and the median of the wavelet energy map is used as the standard deviation to select high-energy time-frequency regions. Morphological processing (i.e., removing isolated points and small connected regions) is then performed on the high-energy time-frequency regions to obtain time-frequency high-energy components, which are then used as nonlinear perturbation components.

[0060] It should be noted that the nonlinear disturbance component in this application refers to the abnormal transient component in the dynamic response process of the silicon carbide chip. The nonlinear disturbance component has repeatability and operating condition dependence, and can represent the dynamic stability of the silicon carbide chip under different electrical conditions. It can reveal the transient behavior of the silicon carbide chip under high-frequency driving and electromagnetic interference conditions, and can be obtained by screening and reconstructing through time-frequency joint analysis.

[0061] In step S2, a link transmission compensation function is constructed based on the nonlinear disturbance component, and dynamic characteristic parameters are determined through the link transmission compensation function and the dynamic response data.

[0062] In this embodiment, the link transmission compensation function constructed based on the nonlinear disturbance component can be implemented in the following manner:

[0063] Construct the initial transfer function model for the silicon carbide chip test link;

[0064] The nonlinear disturbance component is mapped to the initial transmission function model for iterative correction, thereby obtaining the link transmission compensation function.

[0065] In practice, firstly, the input voltage and output current signals in the silicon carbide chip test circuit are acquired using an oscilloscope. Then, the amplitude-frequency curve and phase-frequency curve in the frequency domain are established using the fast Fourier transform method, and these curves are used as the initial transfer function model. Next, the nonlinear disturbance component is superimposed on the initial transfer function model, and the error between the amplitude-frequency curve and the phase-frequency curve is continuously corrected using the least squares iterative algorithm until the error converges to a smooth horizontal line, thus obtaining the corrected amplitude-frequency curve and phase-frequency curve. Finally, both the corrected amplitude-frequency curve and phase-frequency curve are used as the link transmission compensation function.

[0066] It should be noted that the link transmission compensation function in this application is a mathematical mapping relationship, which can be used to characterize and compensate for the frequency domain and time domain distortions in amplitude and phase of the signal under test caused by the test link. It includes amplitude frequency correction term and phase frequency correction term. The amplitude frequency correction term is used to correct the amplitude attenuation or gain error caused by the link, and the phase frequency correction term is used to correct the phase delay or phase shift caused by the link.

[0067] Preferably, referring to Figure 2, which is an exemplary flowchart for determining dynamic characteristic parameters according to the present application, in this embodiment, the determination of dynamic characteristic parameters through the link transmission compensation function and the dynamic response data can be achieved by the following steps:

[0068] First, in step S21, the dynamic response data is corrected for amplitude and phase frequency using the link transmission compensation function to obtain a compensated response signal;

[0069] Then, in step S22, the compensation response signal is subjected to joint time-domain and frequency-domain analysis, and the switching transient characteristics, parasitic response characteristics and power loss characteristics of the silicon carbide chip are obtained by separation and extraction.

[0070] Finally, in step S23, the dynamic change trends of the switching transient characteristics, the parasitic response characteristics, and the power loss characteristics are analyzed to obtain a set of dynamic characteristic parameters.

[0071] In specific implementation, firstly, the dynamic response data is imported into a signal analysis environment (e.g., Python). Each waveform undergoes DC bias removal and is aligned according to the trigger point. Then, a Fast Fourier Transform (FFT) is performed on the processed time-domain waveform to obtain its spectral representation. The spectrum is then subjected to amplitude and phase correction according to its frequency points and the link transmission compensation function. Finally, an Inverse Fast Fourier Transform (IFT) is performed on the corrected spectrum to obtain the time-domain compensation response signal. Then, in the time domain, threshold detection and first derivative are used to identify the start and end times of each switching event. Switching transient characteristics such as rise time, fall time, delay time, amplitude overshoot, and steady-state drift are calculated. Finally, a short-time test is performed on the compensation response signal in the frequency domain. Fourier transform and continuous wavelet transform are used to obtain the time-frequency joint spectrum, and high-frequency harmonic components and parasitic resonance peaks are identified from the time-frequency joint spectrum. Parasitic response features (including resonance frequency, amplitude and damping characteristics) are extracted using a peak detection algorithm. The power loss characteristics of the silicon carbide chip are obtained by statistical means. Finally, the dynamic change trends of the switching transient features, parasitic response features and power loss features are statistically analyzed using a sliding window method. The statistics and trend curves of each switching transient feature, parasitic response feature and power loss feature are obtained. The statistics and trend curves of the switching transient features, parasitic response features and power loss features are then used as a set of dynamic characteristic parameters.

[0072] It should be noted that the dynamic characteristic parameter set in this application is a structured parameter set composed of characteristic statistics and their time series trends. It can be used as input characterization for modules such as spectrum analysis, thermal efficiency derivation, and coupling compensation in this application, facilitating consistent evaluation across operating conditions. The compensation response signal refers to the time-domain voltage and current waveforms after correction by the link transmission compensation function. The switching transient characteristics refer to several time-domain indicators characterizing the dynamic behavior of the silicon carbide chip during the switching process, including: rise time, fall time, switching delay, overshoot amplitude, and steady-state recovery time. The parasitic response characteristics are characteristic quantities of frequency-domain or time-domain resonance phenomena caused by parasitic inductance and capacitance in the connection traces and test links of the silicon carbide chip, including: resonance frequency, resonance amplitude, and damping ratio. The power loss characteristics are indicators characterizing the energy dissipation of the silicon carbide chip under a given operating condition, including: single switching energy, switching power loss, and conduction power loss.

[0073] In step S3, multidimensional spectrum analysis is performed on the dynamic characteristic parameters to obtain high-frequency performance parameters. Then, based on the high-frequency performance parameters, the thermal efficiency of the silicon carbide chip is derived to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels.

[0074] Preferably, referring to Figure 3, which is an exemplary flowchart for determining high-frequency performance parameters according to the present application, in this embodiment, the high-frequency performance parameters are obtained by performing multi-dimensional spectrum analysis on the dynamic characteristic parameters, which can be achieved by the following steps:

[0075] First, in step S31, Fourier transform and wavelet decomposition are performed on the dynamic characteristic parameters to obtain the time-frequency joint spectrum;

[0076] Then, in step S32, high-frequency harmonic components, parasitic oscillation components, and transient jitter characteristics are extracted from the time-frequency joint spectrum;

[0077] Finally, in step S33, the high-frequency performance parameter set of the silicon carbide chip under high-frequency drive is determined by the high-frequency harmonic components, the parasitic oscillation components, and the transient jitter characteristics.

[0078] In specific implementation, firstly, the dynamic characteristic parameters are frequency-domain transformed using short-time Fourier transform (SFT) and then processed by Mohrley wavelet filtering. The time-frequency spectrum of the SFT and the scaling graph of the wavelet transform are then fused by frequency band weighting (e.g., wavelet graph weights are used for narrow-band transients, and SFT weights are used for stationary harmonics) to obtain the joint time-frequency spectrum. Then, spectral peak detection is performed in the joint time-frequency spectrum, and the amplitude and frequency of each detected harmonic are tracked over time to obtain high-frequency harmonic components. Finally, spectral lines with narrow bandwidths and decreasing envelopes over time are retrieved from the joint time-frequency graph to obtain parasitic oscillation components. Specifically, the envelopes of spectral lines with narrow bandwidths and decreasing envelopes over time are extracted using Hilbert transform, and then the envelopes are processed... Exponential fitting yields the parasitic oscillation component. In the time domain, edge detection is performed on each switching event using first-order derivative peak or threshold zero-crossing detection. The time deviation of each edge relative to the reference trigger is statistically analyzed to calculate the root mean square and peak values ​​of the edge time jitter. Simultaneously, Hilbert transform is used to obtain the instantaneous phase for phase extraction of the compensation response signal. The variance of the instantaneous phase is calculated within a short time window to quantify the phase noise, thus obtaining the transient jitter characteristics. Finally, the frequency and amplitude of the high-frequency harmonic components, the resonant frequency and attenuation rate of the parasitic oscillation component, and the time jitter of the transient jitter characteristics (such as RMS, peak value, and short-time phase variance) are structured and stored. The set of stored high-frequency harmonic components, parasitic oscillation components, and transient jitter characteristics is used as the high-frequency performance parameter set.

[0079] It should be noted that the high-frequency performance parameter set in this application is a structured parameter set composed of high-frequency indicators, which can be used as a direct input characterization for thermal efficiency derivation, coupling compensation, and performance evaluation; the time-frequency joint spectrum refers to the time-frequency representation that takes into account both short-time Fourier transform and continuous wavelet transform results, which is convenient for distinguishing between persistent harmonics and transient components; the high-frequency harmonic components refer to the steady-state or quasi-steady-state spectral components existing in the high-frequency band, which can be used to evaluate the high-frequency linear or nonlinear behavior of silicon carbide chips; the parasitic oscillation components are represented by narrowband resonances formed by silicon carbide chip packaging, inductor-capacitor coupling, or test traces; the transient jitter characteristics are indicators for quantifying switching timing stability and phase noise levels.

[0080] In this embodiment, the thermal efficiency of the silicon carbide chip is derived based on the high-frequency performance parameters to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels. Specifically, this can be achieved in the following manner:

[0081] The power loss of the silicon carbide chip at different power levels is determined by the aforementioned high-frequency performance parameters.

[0082] A structural thermal model of a silicon carbide chip is obtained, and then all power losses are input into the structural thermal model to obtain junction temperature change curves and power loss curves.

[0083] Based on the junction temperature change curve and the power loss curve, the energy efficiency trend under different power levels is derived.

[0084] Junction temperature fluctuations and energy efficiency degradation are extracted from all energy efficiency trends, and then the energy efficiency thermal coupling characteristics are determined through the junction temperature fluctuations and energy efficiency degradation.

[0085] In practical implementation, firstly, high-frequency performance parameters are input into the anti-aliasing filter, and the instantaneous voltage and current data output by the anti-aliasing filter are multiplied point by point to obtain the power curve. The power curve is then piecewise integrated at different stages, where switching losses are obtained during the switching transient stage and conduction losses are obtained during the conduction stage, i.e., power losses under different power levels. Secondly, initial structural parameters can be obtained by consulting the silicon carbide chip's packaging material handbook, and a structural thermal model of the silicon carbide chip is obtained by combining the temperature curves of the silicon carbide chip during heating and cooling processes. The power losses under different operating conditions are then substituted into the thermal model as input quantities. The method involves iteratively solving the problem using numerical simulation software (such as Python numerical calculation tools) to obtain the junction temperature change curve and power loss curve over time. Then, the numerical simulation software is used to output the energy efficiency trend under different power conditions by calculating the trend of the ratio of the junction temperature change curve and the power loss curve. Finally, the continuous high peak interval in the energy efficiency trend is taken as the junction temperature fluctuation interval, and the continuous low peak interval in the energy efficiency trend is taken as the energy efficiency decay interval. Then, the junction temperature fluctuation interval and the energy efficiency decay interval are converted into a matrix in a row-ordered form through one-hot encoding, and this matrix is ​​used as the energy efficiency thermal coupling feature.

[0086] It should be noted that the energy efficiency thermal coupling characteristic in this application refers to a comprehensive characterization quantity composed of junction temperature fluctuation and energy efficiency degradation, used to reveal the coupling relationship between the thermal effect and energy efficiency decline of silicon carbide chips; the junction temperature change curve refers to the curve of the junction temperature of silicon carbide chips evolving over time, derived from the thermal model, used to reflect the thermal response of silicon carbide chips under different operating conditions; the power loss curve refers to the curve of the power loss of silicon carbide chips changing over time under different operating conditions, used as the input of the thermal model and the basis for energy efficiency derivation; the energy efficiency trend refers to the change law of energy conversion efficiency of silicon carbide chips with power level or time evolution under different power operating conditions, used to represent the overall energy efficiency performance of silicon carbide chips; in addition, the junction temperature fluctuation in this embodiment refers to the temperature fluctuation range of the junction within a specific time period, used to measure the thermal stability of silicon carbide chips; the energy efficiency degradation refers to the trend of gradual decrease in energy efficiency of silicon carbide chips after continuous operation or multiple cycles, used to characterize the performance degradation of silicon carbide chips. The junction temperature fluctuation and energy efficiency degradation can both be obtained through visualization libraries and peak extraction algorithms.

[0087] In step S4, the deviation anomaly value of the silicon carbide chip during the electromagnetic interference process is collected, and the silicon carbide chip is subjected to multi-condition pulse excitation based on the deviation anomaly value to obtain a multi-dimensional compensation matrix.

[0088] In this embodiment, controlled electromagnetic interference (EMI) testing is used to collect deviation anomalies of the silicon carbide chip during EMI. In actual implementation, anomaly detection methods (such as time-frequency energy anomaly determination based on short-time spectrum and wavelet energy) can be used to identify reproducible deviation anomalies that occur during the injected interference. It should be noted that controlled EMI testing refers to applying electromagnetic interference (including radiated and conducted interference) with known parameters to the silicon carbide chip under controllable and repeatable conditions. Its function is to reveal the sensitive response of the silicon carbide chip at specific frequencies, amplitudes, and modulation characteristics by artificially and reproducibly changing the external electromagnetic environment, so as to quantitatively evaluate the electromagnetic compatibility and immunity of the silicon carbide chip. Deviation anomalies are the deviations in the comparison with the baseline response and can be used as the basis for multi-condition pulse excitation design to retain and utilize the condition-sensitive information carried by the anomalies.

[0089] In this embodiment, the multi-condition pulse excitation of the silicon carbide chip based on the deviation anomaly value to obtain the multi-dimensional compensation matrix can be specifically achieved in the following manner:

[0090] Based on the deviation from the outlier, multiple sets of pulse excitation condition tests were performed on the silicon carbide chip to obtain the coupling relationship between the response parameters and the abnormal characteristics of the silicon carbide chip.

[0091] A multidimensional compensation matrix is ​​constructed based on the aforementioned coupling relationship.

[0092] In practical implementation, firstly, a set of pulse excitation condition templates is designed based on the frequency, amplitude, and timing of outliers. These templates include common parameters such as pulse amplitude, pulse width, rise / fall rate, pulse repetition rate, and duty cycle. Then, a programmable pulse generator sequentially outputs pulse signals according to the pulse excitation condition templates, and the pulses are applied to the ports of the silicon carbide chip via a power amplifier. Under each excitation condition, a high-speed oscilloscope is used to record transient voltage and current waveforms and environmental quantities (such as temperature and power supply voltage). The recorded data undergoes preprocessing such as DC bias removal, baseline subtraction, and trigger alignment before being processed. Statistical techniques are used to extract peak amplitude, rise / fall time, overshoot amplitude, ring frequency, decay time, instantaneous energy, and edge time deviation from the preprocessed data to obtain an experimental data table. The experimental data table is normalized, and a mapping relationship matrix is ​​established using a multiple linear regression algorithm. This mapping relationship matrix is ​​then used as the coupling relationship between the response parameters and abnormal features of the silicon carbide chip. Next, the outliers and the coupling relationship are weighted and fused, and the matrix obtained by weighted fusion is used as a multidimensional compensation matrix. The weight values ​​of the weighted fusion can be determined by the reciprocal of the number of outliers occurring in the above-mentioned multiple linear regression algorithm.

[0093] It should be noted that the multidimensional compensation matrix in this application is a two-dimensional numerical table with excitation parameters as rows and response deviations as columns. Its function is to solidify the excitation-response coupling relationship under abnormal triggering in a lookupable and calculable form, which is conducive to realizing targeted compensation based on abnormality. The coupling relationship between the response parameters and abnormal characteristics of silicon carbide chips refers to the correlation and dependence pattern between response parameters as the excitation changes under the same or different excitation conditions. It can reveal how pulse stimulation systematically affects multiple response dimensions such as switching transients, parasitic oscillations and energy dissipation, thereby providing a quantitative basis for targeted compensation.

[0094] In step S5, adaptive coupling compensation is performed on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels based on the multidimensional compensation matrix, thereby obtaining the performance evaluation dataset of the silicon carbide chip.

[0095] In this embodiment, adaptive coupling compensation is performed on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels based on the multi-dimensional compensation matrix to obtain the performance evaluation dataset of the silicon carbide chip. Specifically, this can be achieved in the following manner:

[0096] The multidimensional compensation matrix is ​​used to perform abnormal compensation on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels, and the initial compensation result and feedback factor are obtained.

[0097] Based on the feedback factor, the initial compensation results are optimized, updated, and their consistency is checked to obtain a performance evaluation dataset for silicon carbide chips.

[0098] In practice, firstly, raw measurement data of the energy efficiency trend and energy efficiency thermal coupling characteristics of silicon carbide chips at different power levels can be collected and stored in matrix form. Then, matrix operations are used to calculate the difference between this matrix and the multidimensional compensation matrix to obtain abnormal compensation values. The abnormal compensation values ​​are then used to replace the raw measurement data in the matrix to obtain the initial compensation result. The deviation between the initial compensation result and the raw data (i.e., the reciprocal of the matrix correlation coefficient) is calculated and used as a feedback factor. Finally, the feedback factor is used as a weight value to perform weighted updates and consistency checks on the initial compensation result to obtain the performance evaluation dataset of the silicon carbide chip. The consistency check includes mean correction and variance check.

[0099] It should be noted that the feedback factor in this application is a vector calculated from the deviation between the initial compensation result and the measurement data, used to guide subsequent compensation optimization and make the compensation process adaptive.

[0100] In summary, the technical solution adopted in this application can realize multi-dimensional adaptive coupling compensation based on outlier-driven methods to improve the reliability of silicon carbide chip performance evaluation.

[0101] Example 2: This application provides a performance evaluation system for silicon carbide chips. Referring to Figure 4, which is a module structure diagram of a performance evaluation system for silicon carbide chips provided in this application, the performance evaluation system includes:

[0102] The data acquisition module 100 is used to acquire dynamic response data of silicon carbide chip under stable operating conditions, and to separate nonlinear disturbance components from the dynamic response data.

[0103] Link compensation module 200 is used to construct a link transmission compensation function based on the nonlinear disturbance component, and to determine dynamic characteristic parameters through the link transmission compensation function and the dynamic response data;

[0104] The spectrum analysis module 300 is used to perform multi-dimensional spectrum analysis on the dynamic characteristic parameters to obtain high-frequency performance parameters, and then to deduce the thermal efficiency of the silicon carbide chip based on the high-frequency performance parameters to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power.

[0105] The operating condition excitation module 400 is used to collect the deviation anomaly value of the silicon carbide chip during the electromagnetic interference process, and to perform multi-operating condition pulse excitation on the silicon carbide chip according to the deviation anomaly value to obtain a multi-dimensional compensation matrix.

[0106] The coupling compensation module 500 is used to perform adaptive coupling compensation on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels based on the multi-dimensional compensation matrix, thereby obtaining the performance evaluation dataset of the silicon carbide chip.

[0107] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0108] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compactdisc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0109] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

Claims

1. A method for evaluating the performance of a silicon carbide chip, characterized in that, The performance evaluation method includes the following steps: collecting dynamic response data of a silicon carbide chip under stable operating conditions, and separating nonlinear disturbance components from the dynamic response data; constructing a link transmission compensation function based on the nonlinear disturbance components, and determining dynamic characteristic parameters through the link transmission compensation function and the dynamic response data; performing multidimensional spectrum analysis on the dynamic characteristic parameters to obtain high-frequency performance parameters, and then deriving the thermal efficiency of the silicon carbide chip based on the high-frequency performance parameters to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels; The deviation anomaly value of the silicon carbide chip during the electromagnetic interference process is collected, and the silicon carbide chip is subjected to multi-condition pulse excitation based on the deviation anomaly value to obtain a multi-dimensional compensation matrix; Based on the multidimensional compensation matrix, adaptive coupling compensation is performed on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels, thereby obtaining the performance evaluation dataset of silicon carbide chips.

2. The performance evaluation method for a silicon carbide chip as described in claim 1, characterized in that, Dynamic response data of silicon carbide chips under stable operating conditions were collected through electromagnetic shielding testing.

3. The performance evaluation method for a silicon carbide chip as described in claim 1, characterized in that, Separating the nonlinear disturbance component from the dynamic response data specifically includes: performing baseline fitting on the dynamic response data to obtain a residual sequence; performing time-frequency joint analysis on the residual sequence to obtain the nonlinear disturbance component.

4. The performance evaluation method for a silicon carbide chip as described in claim 1, characterized in that, The link transmission compensation function based on the nonlinear disturbance component specifically includes: constructing an initial transmission function model for the silicon carbide chip test link; mapping the nonlinear disturbance component to the initial transmission function model for iterative correction, thereby obtaining the link transmission compensation function.

5. The performance evaluation method for a silicon carbide chip as described in claim 1, characterized in that, Determining dynamic characteristic parameters using the link transmission compensation function and the dynamic response data specifically includes: correcting the amplitude and phase frequencies of the dynamic response data using the link transmission compensation function to obtain a compensated response signal; performing joint time-domain and frequency-domain analysis on the compensated response signal, and then extracting the switching transient characteristics, parasitic response characteristics, and power loss characteristics of the silicon carbide chip; and analyzing the dynamic trends of the switching transient characteristics, the parasitic response characteristics, and the power loss characteristics to obtain a set of dynamic characteristic parameters.

6. The performance evaluation method for a silicon carbide chip as described in claim 1, characterized in that, The process of performing multidimensional spectrum analysis on the dynamic characteristic parameters to obtain high-frequency performance parameters specifically includes: performing Fourier transform and wavelet decomposition on the dynamic characteristic parameters to obtain a time-frequency joint spectrum; extracting high-frequency harmonic components, parasitic oscillation components, and transient jitter features from the time-frequency joint spectrum; and determining the high-frequency performance parameter set of the silicon carbide chip under high-frequency drive through the high-frequency harmonic components, the parasitic oscillation components, and the transient jitter features.

7. The performance evaluation method for a silicon carbide chip as described in claim 1, characterized in that, The thermal efficiency derivation of the silicon carbide chip based on the high-frequency performance parameters, and the resulting energy efficiency trends and energy efficiency-thermal coupling characteristics at different power levels, specifically include: determining the power loss of the silicon carbide chip at different power levels using the high-frequency performance parameters; obtaining the structural thermal model of the silicon carbide chip, and then inputting all power losses into the structural thermal model to obtain junction temperature variation curves and power loss curves; deriving the energy efficiency trends at different power levels based on the junction temperature variation curves and the power loss curves; extracting junction temperature fluctuations and energy efficiency degradation from all energy efficiency trends, and then determining the energy efficiency-thermal coupling characteristics using the junction temperature fluctuations and the energy efficiency degradation.

8. The performance evaluation method for a silicon carbide chip as described in claim 1, characterized in that, Abnormal deviation values ​​of silicon carbide chips during electromagnetic interference processes are collected through controlled electromagnetic interference testing.

9. The performance evaluation method for a silicon carbide chip as described in claim 1, characterized in that, The process of performing multi-condition pulse excitation on the silicon carbide chip based on the deviation anomaly values ​​to obtain a multi-dimensional compensation matrix specifically includes: performing multiple sets of pulse excitation condition tests on the silicon carbide chip based on the deviation anomaly values ​​to obtain the coupling relationship between the response parameters and abnormal characteristics of the silicon carbide chip; and constructing a multi-dimensional compensation matrix based on the coupling relationship.

10. A performance evaluation system for a silicon carbide chip, used to execute a performance evaluation method for a silicon carbide chip as described in any one of claims 1 to 9, characterized in that, The performance evaluation system includes: a data acquisition module for acquiring dynamic response data of a silicon carbide chip under stable operating conditions and separating nonlinear disturbance components from the dynamic response data; a link compensation module for constructing a link transmission compensation function based on the nonlinear disturbance components and determining dynamic characteristic parameters through the link transmission compensation function and the dynamic response data; a spectrum analysis module for performing multidimensional spectrum analysis on the dynamic characteristic parameters to obtain high-frequency performance parameters, and then deriving the thermal efficiency of the silicon carbide chip based on the high-frequency performance parameters to obtain the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels; an operating condition excitation module for acquiring deviation anomalies of the silicon carbide chip during electromagnetic interference, performing multi-operating condition pulse excitation on the silicon carbide chip based on the deviation anomalies, and obtaining a multidimensional compensation matrix; and a coupling compensation module for performing adaptive coupling compensation on the energy efficiency trend and energy efficiency thermal coupling characteristics under different power levels based on the multidimensional compensation matrix, thereby obtaining a performance evaluation dataset of the silicon carbide chip.

Images