Method and system for calibrating time sequence deviation of double-pulse test voltage and current waveforms

Abstract

The invention discloses a method and a system for calibrating time sequence deviation of double-pulse test voltage and current waveforms. The method comprises the steps of signal interception, time domain coarse alignment, periodic oscillation section interception, spectrum decomposition and oscillation frequency phase analysis, and frequency domain fine alignment and waveform reconstruction. According to the method, the inherent phase relation of voltage and current oscillation waveforms after a switching device is turned off is extracted, and signal preprocessing and frequency domain calculation are combined, so that rapid and high-precision automatic time sequence alignment is realized. Compared with a hardware method, the method provided by the invention is simple to operate, does not need to modify the circuit in advance, does not influence the reliability of a test result, does not depend on a specific probe or a test fixture, has very high universality, and can be realized through software.

Description

Technical Field

[0001] The present invention belongs to the technical field of semiconductor device testing, and in particular relates to a method and system for calibrating timing deviation of voltage and current waveforms in a double-pulse test. Background Art

[0002] In the early design stages of power electronic converters, it is crucial to accurately evaluate the losses of semiconductor switching devices (such as MOSFETs). Double Pulse Test (DPT) is currently the most widely accepted and adopted switching loss measurement method. In the double pulse test, engineers can accurately measure the switching losses by measuring the switching waveforms of the device under different conditions. Specifically, taking MOSFET as an example, it is only necessary to measure the gate-source voltage v of the MOSFET under predetermined conditions (off voltage, current at the switching moment, drive speed, etc.). gs (t), drain-source voltage v ds (t) and the drain current i d (t) waveform, and v ds (t) and i d The product of (t), that is, instantaneous loss p(t) = v ds (t)i d (t) is integrated to obtain the corresponding switch loss energy parameters (such as turn-on energy E on and the turn-off energy E off ).

[0003] Accurate measurement of switching loss requires not only that the voltage and current probes used have sufficiently high bandwidth, but also that the measured voltage ds (t) and i d(t) requires consistent conversion time, that is, the timing of the two appearing on the oscilloscope is consistent with the actual situation. Therefore, before the DPT test, the voltage probe and current probe used to measure the two need to be calibrated for timing deviation (deskew). Currently, the timing deviation calibration fixture provided by the probe manufacturer is generally used to calibrate the two. However, this method is not universal and is only suitable for timing deviation calibration of the manufacturer's specific probes. If the voltage and current probes are products of different brands, this solution is often not suitable for calibration. In addition, this method can only ensure that the two signals achieve good timing deviation calibration under small signals, while the actual DPT waveforms are all large signal waveforms. If the timing deviation calibration results are completed under small signal conditions, it may still cause timing offsets of tens to hundreds of ps. Another common approach is to use a single pulse test (SPT). This involves replacing the inductor in the DPT with a non-inductive resistor, then simultaneously measuring the voltage and current waveforms for alignment and calibration. This method is versatile and performs calibration under large signals, but it requires significant modifications to the DPT test platform. Furthermore, any changes to the motherboard or components used in the DPT test require re-runs. Each implementation typically takes several hours, resulting in high time costs and the need for highly specialized personnel. Consequently, accurate hardware timing offset calibration prior to DPT testing through hardware-based means presents numerous difficulties and challenges.

[0004] Therefore, another method is to calibrate the timing deviation of the DPT voltage and current waveforms through data post-processing. This method does not require timing offset calibration before testing. Instead, it processes the test data, uses the specific physical relationship between the voltage and current signals to calculate the timing offset, and performs software compensation. For example, due to the equivalent inductance L of the power circuit, eff The existence of the measured tube is turned on at the moment i d (t) will be in L eff A voltage drop occurs across the ds (t) A drop occurs, which can be achieved by aligning i d (t) rising segment and v ds (t) drop segment to achieve timing offset calibration, however, this method requires manual intervention and cannot guarantee the reliability of the calibration results. Some oscilloscope manufacturers, based on this method, use a program to repeatedly try and correct L eff The value of L is calculated by combining the circuit model to reconstruct the current and voltage waveforms and compare them with the measured data. eff The method is relatively effective, but the implementation complexity is very high, and the iterative process still takes more than ten minutes of computing time, which is still not fast enough. Summary of the Invention

[0005] In order to solve the problems existing in the prior art, the present invention proposes a method and system for calibrating timing deviation of voltage and current waveforms in a double-pulse test.

[0006] The technical solutions of the present invention are as follows:

[0007] A method for calibrating timing deviation of voltage and current waveforms in a double-pulse test is performed after the first pulse ends and the original voltage signal and original current signal of the device under test are obtained after the device under test is turned off. The method includes:

[0008] intercepting a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal;

[0009] Using the time difference between the peak moment of the original voltage signal in the preprocessed signal and the first zero crossing moment of the original current signal as the time domain coarse alignment offset, the original current signal is subjected to time domain coarse alignment compensation to obtain a coarsely aligned current signal;

[0010] Eliminating the switching transient edge segments of the original voltage signal and the coarse alignment current signal, retaining the periodic oscillation segment, and forming a fine alignment input signal;

[0011] Obtain the voltage spectrum and current spectrum corresponding to the precisely aligned input signal, extract the frequency with the largest amplitude as the actual oscillation frequency, and calculate the original phase difference at the actual oscillation frequency;

[0012] Based on the actual oscillation frequency and the original phase difference, the timing offset calibration time is calculated and used as the frequency domain fine alignment offset, and the original current signal in the preprocessed signal is subjected to frequency domain fine alignment compensation to complete the dual-pulse test voltage and current waveform timing deviation calibration.

[0013] Furthermore, the specific method of intercepting a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal includes:

[0014] The turn-off moment of the measured period is identified according to the rising edge of the original voltage signal, and a preset time window is shifted to intercept a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment having the original voltage signal and the original current signal before and after the turn-off moment as a preprocessing signal.

[0015] Furthermore, the device under test includes MOSFET / HEMT, IGBT / BJT.

[0016] Furthermore, the method for calculating the time domain coarse alignment offset is as follows:

[0017] Δt coarse =t vds_pk -t id_zcd

[0018] Where, Δt coarse is the time domain coarse alignment offset; t vds_pk is the peak moment of the original voltage signal; t id_zcd is the first zero-crossing moment of the original current signal;

[0019] Furthermore, the specific steps of obtaining the voltage spectrum and current spectrum corresponding to the precisely aligned input signal, extracting the frequency with the largest amplitude as the actual oscillation frequency, and calculating the original phase difference at the actual oscillation frequency include:

[0020] The original voltage signal v in the precise alignment input signal ds_dsk (t) and the coarse alignment current signal i d_dsk (t), perform fast Fourier transform to obtain the corresponding voltage spectrum V ds_dsk (jω) and current spectrum I d_dsk (jω), from the voltage spectrum V ds_dsk (jω) and current spectrum I d_dsk (jω) extract the amplitude and phase of each frequency respectively;

[0021] Get the voltage spectrum V ds_dsk The frequency f with the largest amplitude in (jω) osc_vds and the current spectrum I d_dsk The frequency f with the largest amplitude in (jω) osc_id , as the actual oscillation frequency f osc , where the actual oscillation frequency f osc Satisfies the following formula:

[0022] f osc_vds =f ose_ip =f osc

[0023] Based on the actual oscillation frequency f osc The original phase difference at the actual oscillation frequency is calculated using the following formula and adjusted to the range [-π,π] by modulo 2π operation:

[0024] φ align =∠I d__dsk (jω2πf osc )-∠V ds_dsk (jω2πf ose )

[0025] Where, φ align is the original phase difference; ∠I d_dsk (jω2πf osc) is the phase corresponding to the actual oscillation frequency in the current signal; ∠V ds_dsk (jω2πf osc ) is the phase corresponding to the actual oscillation frequency in the voltage signal.

[0026] Furthermore, the calculation method of the timing offset calibration time is as follows:

[0027] Based on the actual oscillation frequency and the original phase difference, the timing offset calibration time t is calculated as follows: dsk :

[0028]

[0029] Furthermore, the specific method of performing frequency domain precise alignment compensation on the original current signal in the preprocessed signal to complete the timing deviation calibration of the double-pulse test voltage and current waveforms is as follows:

[0030] The original current signal in the preprocessed signal is precisely aligned and compensated in the frequency domain by the following formula to complete the timing deviation calibration of the double-pulse test voltage and current waveforms:

[0031] i′ d (t) = i d (tt dsk )

[0032] Where i d ′(t) is the original current signal after frequency domain precise alignment compensation; i d (tt dsk ) is the process of performing frequency domain precise alignment compensation on the original current signal in the preprocessed signal.

[0033] A system for calibrating timing deviations of voltage and current waveforms in a double-pulse test, for executing any of the above-mentioned methods for calibrating timing deviations of voltage and current waveforms in a double-pulse test, the system comprising a signal interception module, a time domain coarse alignment module, a periodic oscillation segment interception module, a spectrum decomposition and oscillation frequency phase analysis module, and an offset calculation and waveform reconstruction module;

[0034] The signal interception module is used to intercept a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal;

[0035] The time domain coarse alignment module is used to use the time difference between the peak moment of the original voltage signal in the preprocessed signal and the first zero crossing moment of the original current signal as the time domain coarse alignment offset, perform time domain coarse alignment compensation on the original current signal, and obtain a coarse aligned current signal;

[0036] The periodic oscillation segment interception module is used to remove the switching transient edge segments of the original voltage signal and the coarse alignment current signal, retain the periodic oscillation segment, and form a fine alignment input signal;

[0037] The spectrum decomposition and oscillation frequency phase analysis module is used to obtain the voltage spectrum and current spectrum corresponding to the precisely aligned input signal, extract the frequency with the largest amplitude as the actual oscillation frequency, and calculate the original phase difference at the actual oscillation frequency;

[0038] The frequency domain fine alignment and waveform reconstruction module is used to calculate the timing offset calibration time based on the actual oscillation frequency and the original phase difference and use it as the frequency domain fine alignment offset, perform frequency domain fine alignment compensation on the original current signal in the preprocessed signal, and complete the dual pulse test voltage and current waveform timing deviation calibration.

[0039] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor is configured to call and run the computer program stored in the memory to execute any of the above methods.

[0040] A computer-readable storage medium stores a computer program, wherein the computer program implements the steps of any of the above methods when executed by a processor.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] This paper proposes a method and system for calibrating the timing deviation of voltage and current waveforms during dual-pulse testing. This method extracts the inherent phase relationship between the voltage and current oscillation waveforms after a switching device is turned off, combining signal preprocessing with frequency-domain calculations to achieve fast, high-precision automatic timing alignment. Compared to hardware-based methods, the present method is simple to operate, requires no pre-configured circuit modifications, does not affect the reliability of test results, and does not rely on specific probes or test fixtures. It is highly versatile and can be implemented via software.

[0043] The method of the present invention has clear physical principles and is easy to implement through programming. Through a two-stage calibration mechanism of coarse alignment in the time domain and fine alignment in the frequency domain, the calibration of the DPT voltage and current signals can be completed in seconds without human intervention. The results are reliable, trustworthy, and repeatable. In addition, the algorithm proposed in the present invention does not require repeated trial and error. Compared with existing algorithms based on iterative implementation, it can significantly reduce calculation time and has the advantages of high efficiency and accuracy.

[0044] The method of the present invention uses frequency domain phase difference analysis to replace traditional time domain feature point matching, which has strong anti-interference ability. Even when the test data has strong noise and DC offset error, it can still stably achieve accurate timing offset calibration.

[0045] The method of the present invention can be integrated into the oscilloscope post-processing software or used as an independent analysis tool. The entire process does not require interference from factors such as human experience and technical level, which significantly reduces the professional threshold for operators. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 A flow chart of a method for calibrating timing deviation of voltage and current waveforms for double-pulse testing;

[0047] Figure 2 This is the schematic diagram of the double pulse test (DPT) circuit;

[0048] Figure 3 Schematic diagram of voltage and current oscillation waveforms at the DPT shutdown moment;

[0049] Figure 4 This is a schematic diagram of the steps for capturing the waveform segments before and after the shutdown moment;

[0050] Figure 5 Schematic diagram of time domain coarse alignment

[0051] Figure 6 Schematic diagram for intercepting periodic oscillation segments

[0052] Figure 7 Schematic diagram of frequency domain fine alignment

[0053] Figure 8(a) is one of the test result graphs;

[0054] Figure 8(b) is the second test result diagram. DETAILED DESCRIPTION

[0055] The present invention is further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent modifications of the present invention by those skilled in the art fall within the scope defined by the claims attached to this application.

[0056] Example 1:

[0057] A method for calibrating timing deviation of voltage and current waveforms in a double-pulse test according to the present invention is performed after the first pulse ends and the original voltage signal and the original current signal after the device under test is turned off are obtained. The method includes:

[0058] S1. Signal interception: intercepting a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal; the switching transient edge segment includes the rising edge and the falling edge of the signal;

[0059] S2. Time domain coarse alignment: Using the time difference between the peak moment of the original voltage signal in the preprocessed signal and the first zero crossing moment of the original current signal as the time domain coarse alignment offset, the original current signal is subjected to time domain coarse alignment compensation to eliminate large-scale timing deviations and obtain a coarsely aligned current signal.

[0060] S3, periodic oscillation segment interception: remove the switching transient edge segments of the original voltage signal and the coarse alignment current signal, retain the periodic oscillation segment, and form the fine alignment input signal;

[0061] S4. Spectral decomposition and oscillation frequency phase analysis: Obtain the voltage and current spectra corresponding to the precisely aligned input signal, extract the frequency component with the largest amplitude as the actual oscillation frequency, and calculate the original phase difference at the actual oscillation frequency.

[0062] S5. Frequency domain fine alignment and waveform reconstruction: Based on the actual oscillation frequency and the original phase difference, the timing offset calibration time is calculated and used as the frequency domain fine alignment offset. The original current signal in the preprocessed signal is compensated for the frequency domain fine alignment to complete the dual-pulse test voltage and current waveform timing deviation calibration.

[0063] Furthermore, the method of the present invention can also be used for multi-pulse test voltage and current waveform timing deviation calibration, and the steps of the method of the present invention only need to be performed after any pulse of the multi-pulse ends and the original voltage signal and original current signal after the device under test is turned off are obtained.

[0064] Example 2:

[0065] This embodiment is further designed based on the first embodiment. In this embodiment, a method for intercepting a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal includes:

[0066] The turn-off moment of the measured period is identified according to the rising edge of the original voltage signal, and the preset time window is shifted to intercept the combined waveform segment consisting of the switching transient edge segment and the periodic oscillation segment with the original voltage signal and the original current signal before and after the turn-off moment as the preprocessing signal.

[0067] Furthermore, the time window should be able to include the rising edge of the voltage signal, the falling edge of the current signal, and the oscillation waveforms of the two. Generally, 300ns-1us can cover them. A time window that is too long will not affect the processing results of this algorithm.

[0068] Example 3:

[0069] This embodiment is further designed on the basis of the first embodiment in that the device under test in this embodiment includes MOSFET / HEMT, IGBT / BJT or other switching devices. When the device under test is MOSFET or HEMT, the original voltage signal is the drain-source voltage v ds , the original current signal is the drain current i d When the device under test is an IGBT or BJT, the original voltage signal is the collector-emitter voltage v ce , and the original current signal is the collector current i c .

[0070] Example 4:

[0071] This embodiment is further designed based on the first embodiment in that the method for calculating the time domain coarse alignment offset in this embodiment is as follows:

[0072] Δt coarse =t vds_pk -t id_zcd

[0073] Where, Δt coarse is the time domain coarse alignment offset; t vds_pk is the peak moment of the original voltage signal; t id_zcd is the first zero-crossing moment of the original current signal;

[0074] Embodiment 5:

[0075] This embodiment is further designed based on the first embodiment. In this embodiment, the specific steps of obtaining the voltage spectrum and current spectrum corresponding to the precisely aligned input signal, extracting the frequency component with the largest amplitude as the actual oscillation frequency, and calculating the original phase difference at the actual oscillation frequency include:

[0076] The original voltage signal v in the precise alignment input signal ds_dsk (t) and the coarse alignment current signal i d_dsk (t), perform fast Fourier transform to obtain the corresponding voltage spectrum V ds_dsk (jω) and current spectrum I d_dsk (jω), from the voltage spectrum V ds_dsk (jω) and current spectrum I d_dsk (jω) extract the amplitude and phase of each frequency respectively;

[0077] Get the voltage spectrum V ds_dsk The frequency f with the largest amplitude in (jω) osc_vds and the current spectrum I d_dsk The frequency f with the largest amplitude in (jω) osc_id , as the actual oscillation frequency f osc , where the actual oscillation frequency f oscSatisfies the following formula:

[0078] f osc_vds =f osc_ip =f asc

[0079] Based on the actual oscillation frequency f osc The original phase difference at the actual oscillation frequency is calculated using the following formula and adjusted to the range [-π,π] using a modulo 2π operation:

[0080] φ align =∠I d_dsk (jω2πf osc )-∠V ds_dsk (jω2πf osc )

[0081] Where, φ align is the original phase difference; ∠I d_dsk (jω2πf osc ) is the phase corresponding to the actual oscillation frequency in the current signal; ∠V ds_dsk (jω2πf osc ) is the phase corresponding to the actual oscillation frequency in the voltage signal.

[0082] Example 6:

[0083] This embodiment is further designed based on the fifth embodiment in that the calculation method of the timing offset calibration time in this embodiment is as follows:

[0084] Based on the actual oscillation frequency phase difference and the theoretical 90° phase relationship, the timing offset calibration time t is calculated according to the following formula based on the actual oscillation frequency and the original phase difference. dsk :

[0085]

[0086] Embodiment seven:

[0087] This embodiment is further designed on the basis of the fifth embodiment in that, in this embodiment, the original current signal in the preprocessed signal is subjected to frequency domain precise alignment compensation, and the specific method for completing the timing deviation calibration of the voltage and current waveforms in the double-pulse test is as follows:

[0088] The original current signal in the preprocessed signal is precisely aligned and compensated in the frequency domain using the following formula to complete the timing deviation calibration of the voltage and current waveforms in the double-pulse test:

[0089] i′ d (t) = i d (tt dsk )

[0090] Where i′ d(t) is the original current signal after frequency domain precise alignment compensation; i d (tt dsk ) is the process of performing frequency domain precise alignment compensation on the original current signal in the preprocessed signal.

[0091] Embodiment 8:

[0092] A system for calibrating timing deviations of voltage and current waveforms for double-pulse testing according to the present invention is used to execute the method for calibrating timing deviations of voltage and current waveforms for double-pulse testing according to any of the above embodiments. The system comprises: a signal interception module, a time domain coarse alignment module, a periodic oscillation segment interception module, a spectrum decomposition and oscillation frequency and phase analysis module, and an offset calculation and waveform reconstruction module.

[0093] A signal interception module is used to intercept a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal;

[0094] The time domain coarse alignment module is used to use the time difference between the peak moment of the original voltage signal in the preprocessed signal and the first zero crossing moment of the original current signal as the time domain coarse alignment offset, perform time domain coarse alignment compensation on the original current signal to eliminate large-scale timing deviations and obtain a coarsely aligned current signal;

[0095] The periodic oscillation segment interception module is used to remove the switching transient edge segments of the original voltage signal and the coarse alignment current signal, retain the periodic oscillation segment, and form the fine alignment input signal;

[0096] The spectrum decomposition and oscillation frequency phase analysis module is used to obtain the voltage spectrum and current spectrum corresponding to the precisely aligned input signal, extract the frequency with the largest amplitude as the actual oscillation frequency, and calculate the original phase difference at the actual oscillation frequency;

[0097] The frequency domain fine alignment and waveform reconstruction module is used to calculate the timing offset calibration time based on the actual oscillation frequency and the original phase difference and use it as the frequency domain fine alignment offset. It performs frequency domain fine alignment compensation on the original current signal in the preprocessed signal and completes the timing deviation calibration of the double-pulse test voltage and current waveforms.

[0098] Embodiment 9:

[0099] An electronic device of the present invention includes a memory and a processor. The memory stores a computer program. The processor is used to call and run the computer program stored in the memory to execute the method of any of the above embodiments.

[0100] A computer-readable storage medium of the present invention stores a computer program, and when the computer program is executed by a processor, the steps of any of the above embodiments are implemented.

[0101] Application Example 1:

[0102] This example is based on the DPT test circuit and takes MOSFET as an example to further illustrate the principle of the method of the present invention. Figure 2 This is the circuit schematic diagram of DPT test. Generally, DPT test circuit consists of the tested tube Q L , load inductor L and accompanying test diode D (can also be changed to a switch tube), through two pulses to achieve "current pre-charge" and "device test", and finally measure the switching dynamic characteristics of the tested tube under specific voltage and current conditions. During the test, Q L Two pulses are applied to the gate. Q L During the first turn-on period, the current in the inductor L increases linearly, V bus The voltage will charge the inductor current to the target value, and then Q L Shut down, intercept the shutdown instantaneous measured tube Q L The gate-source voltage v gs (t), drain-source voltage v ds (t) and the drain current i d (t) waveform, based on which the turn-off characteristics of the tested tube under the target voltage and current conditions can be analyzed. Before the second pulse arrives, the inductor current continues to flow through the accompanying test diode. When the second pulse is turned on, the inductor current flows from the accompanying test diode to the tested tube Q L Switch flow, intercept the v at the opening time gs (t), v ds (t) and i d (t) waveform, which can be used to analyze the turn-on characteristics of the tested tube under the target voltage and current conditions.

[0103] After the switch is turned off, the equivalent inductance of the power circuit L eff It will resonate with the parasitic capacitance including the junction capacitance, so v ds (t) and i d (t) There will be a significant oscillation process in the waveform. According to circuit knowledge, v ds (t) and i d The oscillation of (t) should have a 90° phase difference, and v ds (t) lags behind i d (t). In other words, after the measured tube is turned off, theoretically i d (t) zero-crossing time, v ds (t) will reach its peak value. This process is as follows Figure 3 shown.

[0104] The method of the present invention relies on this 90° phase difference relationship. Specifically, by identifying and extracting the original phase difference at the oscillation frequency of the measured voltage and current signals, it calculates the time offset that forces them to achieve 90° phase alignment and shifts the waveforms accordingly, achieving accurate timing offset calibration between the two waveforms. The specific steps of the method of the present invention are as follows:

[0105] Algorithm input: i d (t), v ds (t) Original waveform sequence data, i.e., the original voltage signal and original current signal after the device under test is turned off;

[0106] Step 1) Intercept the waveform segments before and after the shutdown moment, from the original v ds (t) and i d In (t), a short period (combined waveform segment) containing the switching transient edge segment and the periodic oscillation segment is intercepted and used as the input signal v for the next step (coarse alignment) ds_dsk_pre1 (t) and i d_dsk_pre1 (t), also known as preprocessing signal; can be calculated based on v ds (t) The rising edge of the signal roughly identifies the shutdown moment of the tested tube, and captures the waveform within hundreds of ns before and after the identified shutdown transient. This process is as follows Figure 4 As shown;

[0107] Step 2) Time domain coarse alignment, identify v ds_dsk_pre1 (t) The time when the peak occurs vds_pk and i d_dsk_pre1 (t) The time of the first zero crossing vds_zcd , use the difference between the two to force a rough alignment, that is:

[0108] Δt coarse =t vds_pk -t id_zcd

[0109] The signal v obtained in this step ds_dsk_pre2 (t) and i d_dsk_pre2 (t) is roughly consistent with the aforementioned 90° phase difference relationship, that is, the timing offset calibration has been preliminarily achieved and is used as the input signal of step 3, such as Figure 5 ;

[0110] Step 3) Intercept the periodic oscillation segment and only intercept and retain the coarse alignment signal v ds_dsk_pre2 (t) and i d_dsk_pre2 (t) at t vds_pk The waveform after the moment, that is, the fast voltage rising edge and current falling edge of the signal are not retained, only the oscillation part is retained; the retained signal is recorded as v ds_dsk (t) and i d_dsk (t) serves as the input signal for subsequent fine alignment, such as Figure 6 ;

[0111] Step 4) For the input signal v ds_dsk (t) and i d_dsk (t) Perform fast Fourier transform (FFT) to obtain the corresponding spectrum V ds_dsk (jω) and I d_dsk (jω), extract the amplitude and phase information of each frequency component from the two spectra.

[0112] Step 5) At V ds_dsk (jω) and I d_dsk In the spectrum of (jω), find the frequency components with the largest amplitude and record the corresponding frequencies as f osc_vds and f osc_id . At this time, the following conditions should be met:

[0113] f ose_vds =f osc_ip =f osc

[0114] Among them, f osc is the actual parasitic oscillation frequency detected from the voltage and current signals.

[0115] Step 6) Calculate the frequency f using the following formula osc The original phase difference φ at align , and adjust it to the range [-π,π] by modulo 2π operation:

[0116] φ align =∠I dsk (jω2πf osc )-∠V ds_dsk (jω2πf osc )

[0117] Step 7) Calculate the timing offset calibration time t by the following formula: dsk :

[0118]

[0119] Step 8) The original current waveform i is calculated by the following formula d (t) Perform time shift to complete the timing offset calibration of voltage and current waveforms:

[0120] i′ d (t) = i d (tt dsk )

[0121] After adjustment, v ds (t) and i d ′(t) is a set of correctly aligned signals that can be used for further switching dynamic analysis such as switching loss calculation.dsk The meaning of each component in Δt coarse It is the time offset obtained by coarse alignment, which can roughly align the voltage and current signals, and then the coefficient is -φ align The term with a coefficient of 1 / 4 corresponds to a quarter cycle (90° phase shift), forcing i d Advance v ds To satisfy the theoretical phase relationship. This step is done by Figure 7 Note: Due to the rough alignment step above, the phase difference between the voltage and current waveforms should be close to 90°. Figure 7 The phase difference between the two waveforms is magnified only to illustrate the principle.

[0122] Here we need to supplement the necessity of the above steps 2 and 3. First, step 2 performs a coarse alignment, and the signal obtained under the coarse alignment has preliminarily realized the timing offset calibration. However, since the two signals may have significant noise, DC bias, DC drop and other interferences in the actual test, these factors will seriously affect the accuracy of the timing offset calibration. The coarse alignment cannot identify and eliminate these interferences, so the subsequent steps proposed by the present invention are required for further calibration. Step 3 is based on step 2 and cuts off the fast edge waveform of the switching transient. This is because the subsequent fine alignment is based on the fast Fourier transform to convert the time domain signal to the frequency domain, and the switching edge contains rich high-order harmonic components. If retained, it cannot be guaranteed that the subsequent steps can accurately identify the oscillation frequency component. Therefore, the correct implementation of steps 2 and 3 is an important guarantee for the subsequent steps of this algorithm to achieve accurate timing offset calibration.

[0123] Application Example 2:

[0124] To verify the method of the present invention, this example was tested on a DPT experimental platform. The device under test was a Navitas G3R40MT12J silicon carbide MOSFET. The test equipment included a Tektronix MDO3054 oscilloscope (500MHz bandwidth, 2.5GS / s sampling rate) with two TPP0500B passive probes (500MHz bandwidth) to measure the gate-source voltage vgs and drain-source voltage vds respectively; the drain current id was measured using a T&M SSDN-015 coaxial shunt resistor. The calibration process was as follows:

[0125] Original timing deviation detection: As shown in Figure 8(a), without calibration, the drain-source voltage v is manually read. ds Peak and drain current i d The time difference of the first zero crossing is measured to be about 4.6ns for the initial timing deviation.

[0126] Algorithm implementation: The method of the present invention is written as a MATLAB program, and the original v is inputds (t) and i d (t) waveform data. The program automatically performs steps such as intercepting the oscillation segment, rough alignment, and spectrum analysis, and finally calculates the timing offset calibration value t dsk =4.678ns.

[0127] Waveform compensation and verification: According to formula (4), the original i d (t) After time shift compensation, the calibration effect is shown in Figure 8(a), v ds Peak and i d It is worth noting that the dotted line v in Figure 8(a) ds Peak and i d There is a slight offset at the first zero crossing, which is caused by i d This interference is caused by the DC offset error of the signal, but this algorithm effectively suppresses this type of interference through frequency domain phase analysis.

[0128] Side verification: Further observation of the turn-on transient process shown in Figure 8(b) shows that after calibration, v ds The voltage drop section and i d The current rise rate (di / dt) periods are completely aligned, indicating that the timing calibration results are accurate and reliable.

[0129] Results: The proposed method accurately calculates timing offsets in a real-world noisy environment, and the results are validated in multiple ways. The entire software processing takes less than one second, requiring no manual intervention or hardware circuit adjustments, demonstrating the efficiency and robustness of the proposed method.

[0130] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a technician familiar with this technical field within the technical scope disclosed by the present invention should be covered by the scope of protection of the present invention.

Claims

1. A method for calibrating the timing deviation of voltage and current waveforms in a double-pulse test, which is performed after the first pulse ends and the original voltage signal and original current signal of the device under test are obtained after the device under test is turned off, characterized in that: The method comprises: intercepting a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal; Using the time difference between the peak moment of the original voltage signal in the preprocessed signal and the first zero crossing moment of the original current signal as the time domain coarse alignment offset, the original current signal is subjected to time domain coarse alignment compensation to obtain a coarsely aligned current signal; Eliminating the switching transient edge segments of the original voltage signal and the coarse alignment current signal, retaining the periodic oscillation segment, and forming a fine alignment input signal; Obtain the voltage spectrum and current spectrum corresponding to the precisely aligned input signal, extract the frequency with the largest amplitude as the actual oscillation frequency, and calculate the original phase difference at the actual oscillation frequency; Based on the actual oscillation frequency and the original phase difference, the timing offset calibration time is calculated and used as the frequency domain fine alignment offset, and the original current signal in the preprocessed signal is subjected to frequency domain fine alignment compensation to complete the dual-pulse test voltage and current waveform timing deviation calibration.

2. The method for calibrating timing deviation of voltage and current waveforms in a double pulse test according to claim 1, characterized in that: The specific method of intercepting a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal includes: The turn-off moment of the measured period is identified according to the rising edge of the original voltage signal, and a preset time window is shifted to intercept a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment having the original voltage signal and the original current signal before and after the turn-off moment as a preprocessing signal.

3. The method for calibrating timing deviation of voltage and current waveforms in a double pulse test according to claim 1, wherein: The devices under test include MOSFET / HEMT, IGBT / BJT.

4. The method for calibrating timing deviation of voltage and current waveforms in a double-pulse test according to claim 1, wherein: The calculation method of the time domain coarse alignment offset is as follows: Δt coarse =t vds_pk -t id_zcd Where, Δt coarse is the time domain coarse alignment offset; t vds_pk is the peak moment of the original voltage signal; t id_zcd is the first zero-crossing moment of the original current signal.

5. The method for calibrating timing deviation of voltage and current waveforms in double-pulse testing according to claim 1, characterized in that: The specific steps of obtaining the voltage spectrum and current spectrum corresponding to the precisely aligned input signal, extracting the frequency with the largest amplitude as the actual oscillation frequency, and calculating the original phase difference at the actual oscillation frequency include: The original voltage signal v in the precise alignment input signal ds_dsk (t) and the coarse alignment current signal i d_dsk (t), perform fast Fourier transform to obtain the corresponding voltage spectrum V ds_dsk (jω) and current spectrum I d_dsk (jω), from the voltage spectrum V ds_dsk (jω) and current spectrum I d_dsk (jω) extract the amplitude and phase of each frequency respectively; Get the voltage spectrum V ds_dsk The frequency f with the largest amplitude in (jω) osc_vds and the current spectrum I d_dsk The frequency f with the largest amplitude in (jω) osc_id , as the actual oscillation frequency f osc , where the actual oscillation frequency f osc Satisfies the following formula: f osc_vds =f osc_ip =f osc Based on the actual oscillation frequency f osc The original phase difference at the actual oscillation frequency is calculated using the following formula and adjusted to the range [-π,π] by modulo 2π operation: f align =∠I d_dsk (jω2πf osc )-∠V ds_dsk (jω2πf osc ) Where, φ align is the original phase difference; ∠I d_dsk (jω2πf osc ) is the phase corresponding to the actual oscillation frequency in the current signal; ∠V ds_dsk (jω2πf osc ) is the phase corresponding to the actual oscillation frequency in the voltage signal.

6. The method for calibrating timing deviation of voltage and current waveforms in a double-pulse test according to claim 5, characterized in that: The calculation method of the timing offset calibration time is as follows: Based on the actual oscillation frequency and the original phase difference, the timing offset calibration time t is calculated as follows: dsk :

7. The method for calibrating timing deviation of voltage and current waveforms in a double-pulse test according to claim 5, characterized in that: The specific method of performing frequency domain precise alignment compensation on the original current signal in the preprocessed signal to complete the timing deviation calibration of the double-pulse test voltage and current waveforms is as follows: The original current signal in the preprocessed signal is precisely aligned and compensated in the frequency domain by the following formula to complete the timing deviation calibration of the double-pulse test voltage and current waveforms: i′ d (t)=i d (t-t dsk ) Where i′ d (t) is the original current signal after frequency domain precise alignment compensation; i d (tt dsk ) is the process of performing frequency domain precise alignment compensation on the original current signal in the preprocessed signal.

8. A system for calibrating timing deviation of voltage and current waveforms in a double-pulse test, for executing the method for calibrating timing deviation of voltage and current waveforms in a double-pulse test described in any one of steps 1 to 7, characterized in that: The system includes a signal interception module, a time domain coarse alignment module, a periodic oscillation segment interception module, a spectrum decomposition and oscillation frequency phase analysis module, and an offset calculation and waveform reconstruction module; The signal interception module is used to intercept a combined waveform segment consisting of a switching transient edge segment and a periodic oscillation segment of the original voltage signal and the original current signal of the device under test as a preprocessing signal; The time domain coarse alignment module is used to use the time difference between the peak moment of the original voltage signal in the preprocessed signal and the first zero crossing moment of the original current signal as the time domain coarse alignment offset, perform time domain coarse alignment compensation on the original current signal, and obtain a coarse aligned current signal; The periodic oscillation segment interception module is used to remove the switching transient edge segments of the original voltage signal and the coarse alignment current signal, retain the periodic oscillation segment, and form a fine alignment input signal; The spectrum decomposition and oscillation frequency phase analysis module is used to obtain the voltage spectrum and current spectrum corresponding to the precisely aligned input signal, extract the frequency with the largest amplitude as the actual oscillation frequency, and calculate the original phase difference at the actual oscillation frequency; The frequency domain fine alignment and waveform reconstruction module is used to calculate the timing offset calibration time based on the actual oscillation frequency and the original phase difference and use it as the frequency domain fine alignment offset, perform frequency domain fine alignment compensation on the original current signal in the preprocessed signal, and complete the dual pulse test voltage and current waveform timing deviation calibration.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor is configured to call and run the computer program stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images