Method for checking a power component

Abstract

The invention relates to a method (100) for checking a power component (1), comprising: - providing (101) an electrical signal, wherein the electrical signal is characterized by at least one first-order parameter, - determining (102) a carrier frequency of the electrical signal, - undersampling (103) the electrical signal using the determined carrier frequency in order to reduce a frequency content of the electrical signal, - applying (104) a frequency analysis to the electrical signal in order to determine the at least one first-order parameter, - checking (105) the power component (1) on the basis of an analysis of the determined at least one first-order parameter. The invention also relates to a computer program, to a device, and to a storage medium for this purpose.

Description

[0001] R.415320

[0002] - 1 -

[0003] Description

[0004] title

[0005] Method for checking a power component

[0006] The invention relates to a method for testing a power component. Furthermore, the invention relates to a computer program, a device, and a storage medium for this purpose.

[0007] State of the art

[0008] Modern semiconductor devices such as SiC or GaN power FETs have enabled the development of power electronics applications characterized by higher power density and efficiency. However, these improvements also bring disadvantages, such as increased parasitic oscillations, electromagnetic interference, and additional power losses resulting in higher temperatures, to name just a few. All these undesirable effects increase the likelihood of degradation and eventual failure of the interconnect technology (AIT) of power semiconductor devices, drivers, and / or assemblies. Therefore, continuous assessment of the condition of the various components is crucial for ensuring reliable system functionality.

[0009] High-speed conversion systems are expensive and rely on very fast digital electronics, low-noise circuits such as oversampling ADCs, and higher-order digital filters. Such complex circuits are often necessary, according to current technology, to recover useful signal information hidden within the noise.

[0010] Disclosure of the invention R.415320

[0011] - 2 -

[0012] The invention relates to a method with the features of claim 1, a computer program with the features of claim 10, a device with the features of claim 11, and a computer-readable storage medium with the features of claim 12. Further features and details of the invention will become apparent from the respective dependent claims, the description, and the drawings. Features and details described in connection with the method according to the invention naturally also apply in connection with the computer program, the device, and the computer-readable storage medium according to the invention, and vice versa, so that mutual reference is always possible with regard to the disclosure of the invention.

[0013] The invention relates in particular to a method for testing a power component, comprising:

[0014] Providing an electrical signal, wherein the electrical signal is characterized by at least one first-order parameter, wherein the at least one first-order parameter is, for example, a change in the amplitude and / or phase of the electrical signal, in particular a change in the amplitude and / or phase within a defined time period or a change across the entire electrical signal, wherein the electrical signal is in particular superimposed with noise.

[0015] Determining a carrier frequency of the electrical signal, wherein the carrier frequency can be determined, for example, by automated or manual localization of maxima in an amplitude spectrum of the electrical signal,

[0016] Undersampling the electrical signal using the specified carrier frequency to reduce a frequency content of the electrical signal, i.e., in particular sampling below twice the specified carrier frequency.

[0017] Applying a frequency analysis, in particular an (inverse) Fourier transform, to the electrical signal to determine at least one first-order parameter, R.415320

[0018] - 3 -

[0019] Verification of the power component based on an analysis of at least one determined first-order parameter.

[0020] By testing the power component according to the inventive method, a defect and / or degradation, i.e., an age-related impairment, in the power component can advantageously be detected, particularly even if the electrical signal is superimposed with noise. Changes in amplitude can, for example, indicate resistive defects in the power component, while changes in phase can indicate inductive and / or capacitive defects in the power component.

[0021] It may optionally be possible that providing the electrical signal includes:

[0022] Capturing a step response, e.g. an internal gate voltage, of the power component to a step or pulse excitation applied, for example, to a gate terminal of the power component, and providing the captured step response as the electrical signal.

[0023] It is optionally provided that the procedure also includes:

[0024] Modulating the electrical signal, especially the undersampled signal, to shift the electrical signal into a lower frequency range.

[0025] This makes it easier to separate the electrical signal from noise. Furthermore, it is conceivable that the electrical signal is multiplied by an exponential function during modulation, preferably by the following exponential function:

[0026]

[0027] It is possible that the procedure may also include:

[0028] Decimating the electrical signal, especially the modulated signal, to reduce a frequency content and / or remove redundant information.

[0029] Decimation is, or includes, in particular, a low-pass filtering of the electrical signal. This allows the electrical signal to be analyzed more effectively. (R.415320)

[0030] - 4 -

[0031] This is because, for example, less signal information needs to be processed and noise in the electrical signal can be reduced.

[0032] Furthermore, it may be provided that the electrical signal is further characterized by at least one second-order parameter and that the method further comprises:

[0033] Determining at least one second-order parameter based on an analysis of at least one first-order parameter, wherein the at least one second-order parameter can, for example, be a change in the slope of the electrical signal, in particular a change in the slope of the electrical signal within a defined time period or a change across the entire electrical signal.

[0034] The verification (of the power component) is furthermore carried out based on an analysis of at least one second-order parameter. This allows for a more differentiated analysis of the electrical signal and thus also a more precise verification of the power component.

[0035] According to an advantageous embodiment of the invention, the method may further comprise:

[0036] Applying a Kalman filter to the electrical signal,

[0037] Determining at least one first-order parameter, and preferably also at least one second-order parameter, based on an analysis of the result obtained by applying the Kalman filter, wherein the verification (of the power component) is further carried out based on an analysis of the at least one first-order parameter, and preferably also at least one second-order parameter, determined by applying the Kalman filter. Applying the Kalman filter provides an additional analysis of the electrical signal, thereby enabling more effective verification of the power component.

[0038] The Kalman filter is, in particular, a mathematical algorithm that can be used to estimate the state of a dynamic system in real time, even if the measurements are noisy or incomplete. R.415320

[0039] - 5 -

[0040] The Kalman filter is based primarily on a recursive estimation method that continuously generates predictions about the system state using a mathematical model and combines these predictions with current measurements. For example, the Kalman filter integrates stochastic models of system dynamics and measurement uncertainties to deliver optimal estimates. The filtering process occurs in two steps: First, a prediction of the state (prediction step) is generated based on the model, and then this prediction is compared with the actual measurements (correction step) to improve the estimate.

[0041] The Kalman filter can be used to smooth noisy signals or interpolate missing data. For example, it can be used for noise reduction in sensors to reconstruct the actual signal from noisy measurements. Its ability to account for both temporal and spatial correlations in signals makes the Kalman filter a particularly powerful tool for improving signal quality and the reliability of estimates in dynamic systems.

[0042] The invention also relates to a computer program, in particular a computer program product, comprising instructions which, when executed by a computer, cause the computer to execute the method according to the invention. Thus, the computer program according to the invention offers the same advantages as those described in detail with reference to a method according to the invention.

[0043] Also part of the invention is a device for data processing, which is configured to carry out the method according to the invention. The device can, for example, be a computer which executes the computer program according to the invention. The computer can have at least one processor for executing the computer program. A non-volatile data storage device can also be provided in which the computer program is stored and from which the computer program can be read by the processor for execution. R.415320

[0044] - 6 -

[0045] The invention may also relate to a computer-readable storage medium which contains the computer program according to the invention and / or includes instructions which, when executed by a computer, cause the computer to execute the method according to the invention. The storage medium is, for example, designed as a data storage device such as a hard drive and / or non-volatile memory and / or a memory card. The storage medium can, for example, be integrated into the computer.

[0046] Furthermore, the method according to the invention can also be implemented as a computer-implemented method. Alternatively or additionally, at least one of the disclosed method steps can be computer-implemented and / or carried out automatically.

[0047] Further advantages, features, and details of the invention will become apparent from the following description, in which exemplary embodiments of the invention are described in detail with reference to the drawings. The features mentioned in the claims and in the description can each be essential to the invention individually or in any combination. The drawings show:

[0048] Fig. 1 shows a schematic visualization of a method, a device, a storage medium and a computer program according to exemplary embodiments of the invention.

[0049] Fig. 2 shows a schematic representation of a power component and an analog-to-digital converter according to exemplary embodiments of the invention.

[0050] Fig. 1 schematically shows a method 100, a device 10, a storage medium 15 and a computer program 20 according to exemplary embodiments of the invention.

[0051] Fig. 1 shows in particular an embodiment of a method 100 for checking a power component 1. In a first step 101, an electrical signal is provided, wherein the electrical signal is characterized by at least one first-order parameter, wherein the electrical signal R.415320

[0052] - 7 -

[0053] in particular, it is superimposed with noise. In a second step 102, a carrier frequency of the electrical signal is determined. In a third step 103, the electrical signal is undersampled using the determined carrier frequency to reduce a frequency content of the electrical signal. In a fourth step 104, a frequency analysis, in particular an inverse Fourier transform, is applied to the electrical signal to determine the at least one first-order parameter. In a fifth step 105, the power component 1 is checked based on an analysis of the determined at least one first-order parameter.

[0054] Fig. 2 shows a power component 1 and an analog-to-digital converter 2 according to embodiments of the invention, wherein the analog-to-digital converter 2 can be used for subsampling 103 of the electrical signal.

[0055] According to exemplary embodiments of the invention, signal processing methods are provided to obtain useful signal information from a noisy electrical signal, e.g.

[0056] Phase changes, changes in amplitude, and exceeding / falling below an over- / under-voltage threshold. Within the scope of the present invention, these parameters are also referred to as first-order parameters, while changes in signal slope are referred to as second-order parameters.

[0057] The first- and second-order parameters can change according to the state of the power component 1 and are, in particular, (natural) system responses to an input step or pulse excitation. However, due to high transients and the coupling of various unwanted noise sources during operation in a specific application, special signal processing methods are required, not only to filter out a useful signal from the noise, but also to evaluate the first- and / or second-order parameters, which otherwise cannot be accurately measured or observed, or only with increased implementation effort. R.415320

[0058] - 8 -

[0059] According to exemplary embodiments of the invention, two types of signal processing methods can be used, for example, to obtain small signal changes from a noisy electrical signal. The first signal processing method is based, in particular, on undersampling, and can therefore be cost-effective and successfully preserve changes in the amplitude and / or phase of a system step response. Amplitude and phase changes are, in particular, first-order parameters that are sensitive to the state of the power component 1 and its connections. Second-order parameters from the system step response, such as changes in signal slope, are, in particular, more difficult to measure or observe accurately.The second signal processing method, which is based in particular on the use of a Kalman filter, demonstrates a potential use for estimating and predicting first- and second-order parameters and can be applied to any type of signal.

[0060] The first signal processing method is preferably based on undersampling of a band-limited electrical signal, which is a typical signal type in power electronics.

[0061] For example, a phase node in a half-bridge application with a transition regime characterized by decaying oscillations until the electrical signal eventually settles at either ground or VDD is a band-limited electrical signal. In the frequency domain, a band-limited electrical signal is characterized, in particular, by having significant frequency components within a defined frequency range [fL, ..., fH]. Below a low frequency (fL) of this defined frequency range and above a high frequency (fH) of this defined frequency range, the frequency components are particularly smaller.

[0062] A band-limited electrical signal is, for example, a response (e.g., internal gate voltage) of a power device 1 to a step or pulse excitation applied to its gate. In the time domain, an initial oscillation behavior can change depending on the switching state. By undersampling such a transient response with a sampling rate lower than R.415320

[0063] - 9 -

[0064] As a Nyquist rate of 2xfH, i.e., less than twice the carrier frequency of the electrical signal, amplitude and phase changes due to system degradation hidden in the noise can be preserved by modulation to shift its spectrum into the low-frequency range and / or decimation to reduce the number of samples. The first and last steps of signal processing, i.e., subsampling and decimation, can favor cost-effective application: For subsampling, an analog-to-digital converter with a low equivalent number of bits can be used, while decimation may advantageously require less memory for system acquisition.

[0065] The response of the power component 1 to a stage applied to its gate is, in particular, a band-limited electrical signal, which, according to the invention, is acquired by undersampling using an analog-to-digital converter 2 with a low effective bit count. The resulting undersampled signal is, in particular, a digital representation of the response of the power component 1 and can be further processed in the frequency domain by the detection and decimation block as follows: The carrier frequency can be determined by locating maxima in the amplitude spectrum of the undersampled electrical signal. The spectrum of the undersampled signal can be shifted into the low-frequency range, where it can be recovered via the decimation block, which acts as a low-pass filter.The shift of the spectrum into the low-frequency range, which corresponds to the modulation of the undersampled signal, is preferably achieved by multiplying each spectral component with an exponential function, in particular the following exponential function:.

[0066]

[0067] where f c the carrier frequency is f s the subsampling rate is and <p eine Phase ist, die erforderlich ist, um eine Zentrierung der Signalspektren um die Frequenz 0 anzupassen.R.415320

[0068] - 10 -

[0069] The subsampling method according to exemplary embodiments of the invention can be used to determine changes in the amplitude or phase of the electrical signal, or the step response, and / or the natural frequency of the damped oscillations. The amplitude of the electrical signal, or the step response, is particularly sensitive to system changes associated with changes in the system resistance components due to degradation or defect, e.g., a degradation of the on-resistance (Rdson) of the power component 1.

[0070] The phase change is dominated in particular by changes in, for example, inductances and capacitances associated with the connections of the power component 1 to a technical system.

[0071] The method according to exemplary embodiments of the invention can also be used to detect changes in a signal envelope and offers a cost-effective alternative to more complex circuits. The signal envelope is particularly sensitive to changes in the signal amplitude and / or phase. The signal envelope is preferably detected by setting the sampling frequency to the natural frequency of the electrical signal.

[0072] The second method according to the invention, which is based on the estimation and prediction of an electrical signal superimposed with noise, is not limited to band-based responses and can be applied to any type of electrical signal. In particular, a Kalman filter is used to detect first- and second-order parameter changes in the electrical signal.

[0073] It has been shown in particular that changes in the amplitude of the step response of power component 1 indicate resistive defects, while changes in the phase shift may indicate inductive and / or capacitive defects. However, the changes may be small and hidden within the noise. Therefore, an additional filter stage may be required to correctly isolate the signal changes from the overall system noise. R.415320

[0074] - 11 -

[0075] They differ. The Kalman filter can be applied to a system response.

[0076] When using the Kalman filter, an integrated power switch can be described as power component 1 at low and medium frequencies, for example, by an RC system, while a discrete switch as power component 1 with its connections can be modeled by an RLC system.

[0077] The preceding explanation of the embodiments describes the present invention exclusively by way of examples.

[0078] Of course, individual features of the embodiments can be freely combined with one another, provided this is technically feasible, without leaving the scope of the present invention.

Images