A digital-to-analog converter calibration method and apparatus, electronic device, and storage medium

Abstract

The application provides a digital-to-analog converter calibration method and device, electronic equipment and storage medium, and relates to the technical field of digital-to-analog conversion.The method comprises the following steps: acquiring multi-dimensional running time sequence data; extracting time sequence features by using a preset feature extraction model, and screening out preset error feature centers matched with the time sequence features; obtaining a current error source based on the error source corresponding to the preset error feature center, and determining a current error type corresponding to the multi-dimensional running time sequence data; determining an ideal output voltage according to a digital input code word, and determining an actual output voltage according to an output signal, and determining an error index based on the ideal output voltage and the actual output voltage; when the error index meets a preset calibration condition, performing a calibration operation pre-associated with the current error type.The application accurately identifies the error source and the error type, and performs the corresponding calibration operation, which is beneficial to realizing online dynamic calibration of the digital-to-analog converter, and ensures the accuracy of the digital-to-analog converter in long-term operation under complex environments.

Description

Technical Field

[0001] This invention relates to the field of digital-to-analog converter technology, and more specifically, to a digital-to-analog converter calibration method, apparatus, electronic device, and storage medium. Background Technology

[0002] A digital-to-analog converter (DAC) serves as a conversion hub between digital and analog signals. It is widely used in software-defined radio, arbitrary waveform generators, radar systems, broadband communications, and other fields, with output signal frequencies covering radio frequency, microwave, and even millimeter-wave bands.

[0003] Currently, to ensure the accuracy of digital-to-analog converters (DACs), related technologies generally perform a one-time calibration at the factory. However, during actual operation, DACs are highly susceptible to errors caused by various factors, making it difficult to guarantee their long-term operational accuracy and affecting the quality of the output signal. Summary of the Invention

[0004] The problem addressed by this invention is how to ensure the long-term operational accuracy of digital-to-analog converters.

[0005] To address the above problems, this invention provides a digital-to-analog converter calibration method, comprising: Acquire multidimensional runtime timing data of the digital-to-analog converter; wherein, the multidimensional runtime timing data includes digital input codewords and output signals; The temporal features of the multidimensional runtime sequence data are extracted using a preset feature extraction model, and preset error feature centers that match the temporal features are selected from a preset set of error feature centers; wherein, one preset error feature center corresponds to one error source; Based on the error source corresponding to the preset error feature center, the current error source corresponding to the multidimensional runtime sequence data is obtained, and the current error type corresponding to the multidimensional runtime sequence data is determined; wherein, the current error source includes any one of temperature drift, power supply ripple, code pattern dependence, and aging effect; the current error type includes any one of static nonlinear error, dynamic nonlinear error, and clock-related error; The ideal output voltage is determined based on the digital input codeword, and the actual output voltage is determined based on the output signal. An error index is then determined based on the ideal output voltage and the actual output voltage. When the error index meets the preset calibration conditions, the calibration operation pre-associated with the current error type is executed; wherein, the calibration operation includes determining the calibration parameter adjustment value and adjusting the corresponding calibration parameter based on the calibration parameter adjustment value.

[0006] Optionally, before selecting the preset error feature centers that match the time-series features from the preset set of error feature centers, the method further includes: Obtain each multidimensional runtime sequence data sample and the error source label corresponding to each multidimensional runtime sequence data sample, and use the preset feature extraction model to extract the historical time series features corresponding to each multidimensional runtime sequence data sample. Density clustering is performed on each of the historical time-series features to obtain multiple clusters. The error feature center corresponding to each cluster is used as the preset error feature center, and the preset error feature center set is constructed. The error sources of the preset error feature centers are determined based on the error source labels corresponding to the clusters, and the error sources are associated with the corresponding preset error feature centers; The error type is determined based on the multidimensional runtime sequence data samples corresponding to the cluster, and the error type is associated with the error feature center.

[0007] Optionally, the multidimensional runtime timing data further includes kernel temperature and clock jitter; the step of determining the error type based on each of the multidimensional runtime timing data samples corresponding to the cluster includes: Determine the integral nonlinearity error, differential nonlinearity error, spurious-free dynamic range, and inter-symbol interference for each of the multidimensional runtime sequence data samples. When a preset first characteristic condition is met, the error type is the static nonlinear error; wherein, the preset first characteristic condition includes, at the same kernel temperature, the first standard deviation between the integral nonlinear errors corresponding to the same digital input codeword is less than a preset first standard deviation threshold, and / or, the second standard deviation between the differential nonlinear errors corresponding to the same digital input codeword is less than a preset second standard deviation threshold. When the preset second characteristic condition is met, the error type is the dynamic nonlinear error; wherein, the preset second characteristic condition includes that each of the spurious-free dynamic ranges decreases as the jump frequency or jump amplitude corresponding to the digital input codeword increases, and / or, the inter-code interference increases as the jump frequency or jump amplitude corresponding to the digital input codeword increases. When the preset first characteristic condition and the preset second characteristic condition are not met and the clock jitter is greater than the preset jitter threshold, and / or when the preset first characteristic condition and the preset second characteristic condition are not met and duty cycle distortion is detected based on the harmonic components of the output signal, the error type is the clock-related error.

[0008] Optionally, the digital input codeword includes sub-codewords corresponding to multiple sub-intervals, the sub-intervals being obtained by dividing the digital input codeword based on a preset range interval; the error index includes sub-error indices corresponding to the sub-codewords within each sub-interval; after determining the error index based on the ideal output voltage and the actual output voltage, the method further includes: Obtain the error threshold corresponding to each of the sub-intervals pre-associated with the current error source. When the sub-error index corresponding to any sub-interval is greater than the corresponding error threshold, determine that the error index meets the preset calibration condition.

[0009] Optionally, the digital-to-analog converter calibration method further includes: The sub-error index corresponding to each sub-interval is obtained according to a preset monitoring period; wherein, the preset monitoring period is longer than the sampling period of the multi-dimensional runtime sequence data; When the rate of change of the sub-error index corresponding to each sub-interval with the same range is less than the preset rate of change for a preset number of consecutive preset times, the error threshold corresponding to that sub-interval is increased according to the preset gradient. When the rate of change of the sub-error index corresponding to each sub-interval with the same range is greater than the preset rate of change, the error threshold corresponding to the sub-interval is reduced according to the preset gradient, and the preset monitoring period is shortened.

[0010] Optionally, after obtaining the current error source corresponding to the multidimensional runtime sequence data based on the error source corresponding to the preset error feature center, the method further includes: The target probability of the current error source shifting to other error sources is determined based on the acquired historical dataset; wherein, the historical dataset includes historical error sources corresponding to each historical multidimensional runtime sequence data; When the target probability is greater than a preset probability threshold, the error source corresponding to the target probability is taken as the target error source, and the target error type corresponding to the target error source is determined. Obtain the calibration operation pre-associated with the target error type, obtain the target calibration operation, and execute the preparation operation corresponding to the target calibration operation; wherein, the preparation operation includes determining the calibration parameter adjustment value.

[0011] Optionally, determining the calibration parameter adjustment value includes: When the current error type is the static nonlinear error, determine the calibration word of the current source array and obtain the calibration parameter adjustment value; When the current error type is the dynamic nonlinear error, determine the adjustment coefficient and / or predistortion compensation value of the deemphasis filter to obtain the calibration parameter adjustment value; When the current error type is the clock-related error, determine the control word of the phase interpolator and / or determine the bias current of the clock buffer stage to obtain the calibration parameter adjustment value.

[0012] In this invention, acquiring multi-dimensional runtime timing data of the digital-to-analog converter (DAC), such as digital input codewords and output signals, facilitates understanding the DAC's output response under different inputs, providing reliable data support for accurately identifying the current error source and type. Based on this, the invention utilizes a preset feature extraction model to extract the temporal features of the multi-dimensional runtime timing data, which helps to uncover hidden error-related patterns and rules within the data. Different error sources correspond to different error patterns and rules. In this invention, each preset error feature center corresponds to one error source. By selecting preset error feature centers that match the temporal features, the error source corresponding to that feature can be determined, thereby quickly locating the influencing factors causing errors in the current multi-dimensional runtime timing data. Since different error sources may produce different error types, the corresponding calibration operations also differ. By determining the error type reflected in the multi-dimensional runtime timing data, this invention provides a reliable basis for selecting appropriate calibration operations. Furthermore, this invention determines the ideal output voltage based on the digital input codeword, the actual output voltage based on the output signal, and determines the error index based on both. This facilitates the precise quantification of the performance deviation of the digital-to-analog converter (DAC), providing a reliable reference for whether to perform subsequent calibration operations. This avoids unnecessary calibration operations when the DAC's accuracy is already high. When the error index meets the preset calibration conditions, it indicates that the current DAC error is large. At this time, the pre-associated calibration operation for the current error type is performed, the calibration parameter adjustment value is determined, and the corresponding calibration parameters are adjusted based on the adjustment value. This facilitates online dynamic calibration of the DAC, ensuring its accuracy even during long-term operation in complex environments. This effectively reduces the continuous degradation of the DAC's output signal quality caused by error sources such as temperature drift, power supply ripple, code pattern dependence, and aging effects during operation.

[0013] The present invention also provides a digital-to-analog converter calibration device, comprising: The data acquisition module is used to acquire multi-dimensional runtime timing data of the digital-to-analog converter; wherein, the multi-dimensional runtime timing data includes digital input codewords and output signals; The feature extraction module is used to extract the temporal features of the multidimensional runtime sequence data using a preset feature extraction model, and to filter out preset error feature centers that match the temporal features from a preset set of error feature centers; wherein, one preset error feature center corresponds to one error source; The type determination module is used to obtain the current error source corresponding to the multidimensional runtime sequence data based on the error source corresponding to the preset error feature center, and to determine the current error type corresponding to the multidimensional runtime sequence data; wherein, the current error source includes any one of temperature drift, power supply ripple, code pattern dependence, and aging effect; the current error type includes any one of static nonlinear error, dynamic nonlinear error, and clock-related error; An error determination module is used to determine an ideal output voltage based on the digital input codeword, and to determine an actual output voltage based on the output signal, and to determine an error index based on the ideal output voltage and the actual output voltage; The calibration execution module is used to execute a pre-associated calibration operation for the current error type when the error index meets the preset calibration conditions; wherein the calibration operation includes determining a calibration parameter adjustment value and adjusting the corresponding calibration parameter based on the calibration parameter adjustment value.

[0014] The digital-to-analog converter calibration device and the digital-to-analog converter calibration method provided by this invention have essentially the same advantages over the prior art, and will not be repeated here.

[0015] The present invention also provides an electronic device, including a memory and a processor; The memory is used to store computer programs; The processor is used to implement the digital-to-analog converter calibration method as described above when executing the computer program.

[0016] The electronic device provided by this invention has essentially the same advantages as the digital-to-analog converter calibration method compared to the prior art, and will not be repeated here.

[0017] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the digital-to-analog converter calibration method as described above.

[0018] The advantages of the computer-readable storage medium and the digital-to-analog converter calibration method provided by this invention are basically the same as those of the prior art, and will not be repeated here. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the digital-to-analog converter calibration method according to an embodiment of the present invention; Figure 2 This is a timing diagram of the digital-to-analog converter calibration method according to an embodiment of the present invention; Figure 3 This is a timing diagram of a digital-to-analog converter calibration method according to another embodiment of the present invention; Figure 4This is a schematic diagram of the digital-to-analog converter calibration device according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0020] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the accompanying drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0021] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0022] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0023] It should be noted that the terms "a" and "a plurality of" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0024] like Figure 1 As shown in the figure, an embodiment of the present invention provides a digital-to-analog converter calibration method, which includes the following steps: S1: Acquire multi-dimensional runtime timing data of the digital-to-analog converter; wherein, the multi-dimensional runtime timing data includes digital input codewords and output signals.

[0025] Specifically, the multidimensional runtime timing data referred to in this embodiment represents the collection of data in different dimensions that change over time during the operation of the digital-to-analog converter (DAC). It can include digital input codewords and output signals. The digital input codeword is the digital signal received by the DAC for conversion, representing different values ​​in binary code form. The output signal represents the analog signal output by the DAC after converting the digital input codeword. For example, in this embodiment, a high-speed comparator array can be used to sample the analog output of the DAC at twice the DAC sampling clock frequency to obtain the output signal.

[0026] S2: Use a preset feature extraction model to extract the temporal features of multidimensional runtime sequence data, and select preset error feature centers that match the temporal features from the preset error feature center set; where each preset error feature center corresponds to an error source.

[0027] Specifically, the preset feature extraction model referred to in this embodiment is used to extract information reflecting the operating characteristics of the digital-to-analog converter from multi-dimensional runtime sequence data, and can be trained in advance. The time-series features referred to in this embodiment represent the time-varying feature information extracted from the multi-dimensional runtime sequence data by the preset feature extraction model, which may include short-term change feature vectors and long-term drift feature vectors. The preset error feature centers referred to in this embodiment represent typical features corresponding to different error sources. Multi-dimensional runtime sequence data samples corresponding to digital-to-analog converters with different error sources can be obtained in advance, and feature extraction can be performed on each multi-dimensional runtime sequence data sample to obtain multiple time-series features corresponding to each error source. Then, typical time-series features (such as the average value of multiple time-series features) are selected as preset error feature centers. Based on this, the error sources corresponding to each preset error feature center can be associated to construct a preset error feature center set.

[0028] In one embodiment, a Long Short-Term Memory (LSTM) network can be used as the initial feature extraction model. A training dataset is pre-constructed based on multi-dimensional runtime data samples and corresponding temporal feature labels. Based on this, the initial feature extraction model can be trained using the multi-dimensional runtime data samples as input and the temporal feature labels as output, thereby obtaining a preset feature extraction model. In this embodiment, the LSTM network may include an input layer, a first LSTM layer, a second LSTM layer, and an output layer. The input layer receives multi-dimensional runtime data (such as multi-dimensional feature vectors composed of temporal data). The first LSTM layer extracts temporal features from the multi-dimensional runtime data, outputting short-term changing feature vectors. These short-term changing feature vectors are then input into the second LSTM layer to further extract higher-order temporal features and long-term drift feature vectors. Furthermore, the output layer concatenates the short-term changing feature vectors and the long-term drift feature vectors to obtain temporal features. Based on this, preset error feature centers matching the temporal features can be selected from a preset set of error feature centers. For example, the distance between the time series feature and each preset error feature center can be determined, and the closest one can be selected as the preset error feature center that matches the time series feature.

[0029] Optionally, the multidimensional runtime timing data in this embodiment may further include at least one of core temperature, power supply parameters (such as peak voltage ripple, spectral distribution characteristics, etc.), and bit error rate. After acquiring the multidimensional timing data, error indices can be determined based on the digital input codeword and output signal; wherein, the error indices may include at least one of integral nonlinearity error, differential nonlinearity error, spurious-free dynamic range, and inter-symbol interference. Based on this, the digital input codeword, error indices, core temperature, power supply parameters, bit error rate, and output signal can be concatenated into a data vector and used as the input to the LSTM network, which is beneficial to further improve the accuracy of current error source identification.

[0030] S3: Based on the error source corresponding to the preset error feature center, obtain the current error source corresponding to the multi-dimensional runtime sequence data, and determine the current error type corresponding to the multi-dimensional runtime sequence data; wherein, the current error source includes any one of temperature drift, power supply ripple, code pattern dependence and aging effect; the current error type includes any one of static nonlinear error, dynamic nonlinear error and clock-related error.

[0031] Specifically, in this embodiment, the current error source refers to the cause of error in the current multi-dimensional runtime timing data output by the digital-to-analog converter, determined according to the matched preset error feature center. The current error source includes any one of temperature drift, power supply ripple, code pattern dependence, and aging effect. The current error type in this embodiment represents the error classification corresponding to the current error source, and can include any one of static nonlinear error, dynamic nonlinear error, and clock-related error. Static nonlinear error can include integral nonlinear error and differential nonlinear error, mainly caused by mismatch in the DAC's internal current source array, manifesting as the output characteristic curve deviating from an ideal straight line. Dynamic nonlinear error mainly includes inter-symbol interference and glitches, primarily caused by charge injection and timing deviations during switching, and typically intensifies with increasing input signal frequency. Clock-related error can include clock jitter and duty cycle distortion, mainly caused by non-ideal characteristics of the clock path.

[0032] In one embodiment, after selecting preset error feature centers that match the timing characteristics, the error sources corresponding to the preset error feature centers can be used as the current error sources for the multidimensional runtime timing data. Based on this, the multidimensional runtime timing data can be analyzed to determine its corresponding current error type. For example, multiple quality indicators corresponding to the multidimensional runtime timing data can be determined, such as integral nonlinearity error, differential nonlinearity error, inter-symbol interference, glitches, and clock jitter. Each of these quality indicators is compared with its corresponding preset standard threshold, and the quality indicator with the largest deviation from the preset standard threshold is selected, with its corresponding error type being used as the current error type.

[0033] S4: Determine the ideal output voltage based on the digital input codeword, and determine the actual output voltage based on the output signal, and determine the error index based on the ideal output voltage and the actual output voltage.

[0034] Specifically, in this embodiment, the ideal output voltage refers to the theoretically expected output voltage value determined based on the conversion principle of the digital-to-analog converter (DAC) and the digital input codeword. The actual output voltage refers to the actual voltage value obtained by measuring the output signal of the DAC. The error index in this embodiment refers to an indicator used to measure the difference between the actual output and the ideal output of the DAC, such as integral nonlinearity error or differential nonlinearity error.

[0035] In one embodiment, the ideal output voltage corresponding to each digital input codeword can be determined based on the parameters of the digital-to-analog converter and the digital input codewords, according to the conversion relationship. Simultaneously, the actual output voltage corresponding to each digital input codeword can be obtained from the output signal. Based on this, an error index (such as determining the integral nonlinearity error based on the maximum deviation between the actual and ideal output voltages) can be determined according to the ideal and actual output voltages corresponding to each digital output codeword.

[0036] S5: When the error index meets the preset calibration conditions, perform the calibration operation pre-associated with the current error type; wherein, the calibration operation includes determining the calibration parameter adjustment value and adjusting the corresponding calibration parameter based on the calibration parameter adjustment value.

[0037] Specifically, the preset calibration conditions referred to in this embodiment represent pre-set standards for determining whether calibration is required. For example, when the error index is greater than a preset error threshold (e.g., integral nonlinearity error greater than 1 LSB), the current output signal quality accuracy is considered poor, thus meeting the preset calibration conditions; when the error index is less than or equal to the threshold, the current output signal quality accuracy is considered acceptable, but the preset calibration conditions are not met. The calibration operation referred to in this embodiment represents a pre-set operation to adjust the digital-to-analog converter parameters to reduce the error for the current error type. This can be pre-set according to the characteristics corresponding to different error types. The calibration operation includes determining the calibration parameter adjustment value of the digital-to-analog converter and adjusting the corresponding calibration parameters based on the calibration parameter adjustment value. For some error types, the calibration operation may also include a trigger operation. For example, the trigger operation pre-associated with clock-related errors may include triggering a duty cycle correction circuit.

[0038] In one embodiment, a calibration operation library can be pre-built based on the characteristics of different error types, which can record the calibration methods and calibration parameters corresponding to different error types. When the error index meets the preset calibration conditions, the calibration operation pre-associated with the current error type can be retrieved from the calibration operation library, the calibration parameter adjustment value can be determined (e.g., by looking up a pre-built mapping table), and the corresponding calibration parameter can be adjusted using the calibration parameter adjustment value through the internal register of the digital-to-analog converter or the external control circuit (e.g., by replacing the attribute value corresponding to the calibration parameter with the calibration parameter adjustment value), thereby realizing the online dynamic adjustment of the calibration parameter.

[0039] In this embodiment, acquiring multi-dimensional runtime timing data of the digital-to-analog converter (DAC), such as digital input codewords and output signals, helps to understand the DAC's output response under different inputs, providing reliable data support for accurately identifying the current error source and type. Based on this, this embodiment utilizes a preset feature extraction model to extract the temporal features of the multi-dimensional runtime timing data, which helps to uncover hidden error-related patterns and rules within the multi-dimensional runtime timing data. Different error sources correspond to different error patterns and rules. In this embodiment, each preset error feature center corresponds to one error source. By selecting preset error feature centers that match the temporal features, the error source corresponding to that temporal feature can be determined, thereby quickly locating the influencing factors causing errors in the current multi-dimensional runtime timing data. Since different error sources may produce different error types, the corresponding calibration operations also differ. This embodiment, by determining the error type reflected in the multi-dimensional runtime timing data, provides a reliable basis for selecting appropriate calibration operations. Furthermore, this embodiment determines the ideal output voltage based on the digital input codeword, the actual output voltage based on the output signal, and the error index based on both. This facilitates the precise quantification of the performance deviation of the digital-to-analog converter (DAC), providing a reliable reference for whether to perform subsequent calibration operations. This avoids unnecessary calibration operations when the DAC's accuracy is high. When the error index meets the preset calibration conditions, it indicates that the current DAC error is large. At this time, the pre-associated calibration operation for the current error type is performed, the calibration parameter adjustment value is determined, and the corresponding calibration parameters are adjusted based on the adjustment value. This facilitates online dynamic calibration of the DAC, ensuring its accuracy even during long-term operation in complex environments. This effectively reduces the continuous degradation of the DAC's output signal quality caused by error sources such as temperature drift, power supply ripple, code pattern dependence, and aging effects during operation.

[0040] Optionally, before selecting preset error feature centers that match the time-series features from the preset set of error feature centers, the method further includes: Obtain multi-dimensional runtime sequence data samples and error source labels corresponding to each multi-dimensional runtime sequence data sample, and use a preset feature extraction model to extract historical time series features corresponding to each multi-dimensional runtime sequence data sample. Density clustering is performed on each historical time series feature to obtain multiple clusters. The error feature center corresponding to each cluster is used as the preset error feature center, and a preset error feature center set is constructed. The error sources of the preset error feature centers are determined based on the error source labels corresponding to each cluster, and the error sources are associated with the corresponding preset error feature centers; Error types are determined based on the multidimensional runtime sequence data samples corresponding to the clusters, and the error types are associated with the error feature centers.

[0041] Specifically, in this embodiment, the error source label refers to the marker indicating the cause of the error corresponding to each multi-dimensional runtime timing data sample. Examples include temperature drift, power supply ripple, code pattern dependence, and aging effect, used to clarify the error attributes of the sample data. For instance, the multi-dimensional runtime timing data samples corresponding to a DAC exhibiting a single error source of aging effect can be obtained, and the error source label for each of these samples can be set to aging effect. In this embodiment, the historical time-series features refer to the time-varying feature information extracted from each multi-dimensional runtime timing data sample by a preset feature extraction model.

[0042] In one embodiment, after extracting historical time-series features corresponding to each multi-dimensional runtime data sample using a preset feature extraction model, density clustering can be performed on each historical time-series feature to obtain multiple clusters. For example, each historical time-series feature can be classified according to the error source label to obtain multiple historical time-series feature sets. For each historical time-series feature set, the Euclidean distance between each historical time-series feature in the set can be obtained. When the Euclidean distance between two historical time-series features is less than a preset neighborhood radius parameter, they are determined to belong to the same neighborhood, thereby achieving density clustering and obtaining clusters corresponding to different historical time-series feature sets. In this embodiment, the preset neighborhood radius parameter corresponding to different historical time-series feature sets during clustering can be determined according to the error source label. For example, for temperature drift errors, the preset neighborhood radius parameter can be set to 0.15; for power supply ripple errors, the preset neighborhood radius parameter can be set to 0.08; for code pattern dependency errors, the preset neighborhood radius parameter can be set to 0.12; and for aging effect errors, the preset neighborhood radius parameter can be set to 0.1. After obtaining multiple clusters, the error feature centers corresponding to each cluster are used as preset error feature centers (e.g., the error feature centers can be obtained based on the mean of each historical time series feature in the cluster), and a preset error feature center set is constructed. Furthermore, the unified error source label corresponding to the cluster can be directly used as the error source of the preset error feature center, and the error source can be associated with the corresponding preset error feature center.

[0043] Optionally, the multidimensional runtime timing data also includes kernel temperature and clock jitter; the error type is determined based on the multidimensional runtime timing data samples corresponding to each cluster, including: Determine the integral nonlinearity error, differential nonlinearity error, spurious-free dynamic range, and inter-symbol interference for each multidimensional runtime sequence data sample; When the preset first characteristic condition is met, the error type is static nonlinear error; wherein, the preset first characteristic condition includes, at the same core temperature, the first standard deviation between the integral nonlinear errors corresponding to the same digital input codeword is less than the preset first standard deviation threshold, and / or, the second standard deviation between the differential nonlinear errors corresponding to the same digital input codeword is less than the preset second standard deviation threshold. When the preset second characteristic condition is met, the error type is dynamic nonlinear error; wherein, the preset second characteristic condition includes that each spurious-free dynamic range decreases as the corresponding transition frequency or transition amplitude of the digital input codeword increases, and / or, inter-code interference increases as the corresponding transition frequency or transition amplitude of the digital input codeword increases. When the preset first characteristic condition and the preset second characteristic condition are not met and the clock jitter is greater than the preset jitter threshold, and / or when the preset first characteristic condition and the preset second characteristic condition are not met and duty cycle distortion is detected based on the harmonic components of the output signal, the error type is clock-related error.

[0044] Specifically, the core temperature referred to in this embodiment is the operating temperature of the core components inside the digital-to-analog converter (DAC). For example, in this embodiment, temperature monitoring modules (such as thermistors) can be integrated at different locations on the DAC chip, and the average value of the temperatures detected by all temperature monitoring modules can be used as the core temperature. Clock jitter referred to in this embodiment is the instability of the DAC clock signal, manifested as random variations in the clock signal period. For example, in this embodiment, a clock monitoring module (such as a time-to-digital converter) can be set at the output of the clock buffer stage to measure the jitter information (such as histogram statistics) of the clock edges at a preset resolution. Integral nonlinearity error referred to in this embodiment measures the cumulative deviation between the actual output voltage and the ideal output voltage of the DAC, reflecting the linearity deviation of the DAC across the entire input range. Differential nonlinearity error referred to in this embodiment measures the deviation between the outputs corresponding to adjacent digital input codewords of the DAC, reflecting the linearity of the DAC at the interval of a single digital input codeword. Spurious-free dynamic range referred to in this embodiment represents the power ratio of the fundamental component to the strongest spurious component in the DAC output signal, reflecting the purity and anti-interference capability of the DAC output signal. In this embodiment, inter-symbol interference (ISI) refers to the situation where the output of the current digital input codeword is affected by the previous digital input codeword, resulting in distortion of the output signal. It should be understood that the aforementioned integral nonlinearity error, differential nonlinearity error, spurious-free dynamic range, and ISI are all commonly used quality indicators for digital-to-analog converters (DACs). These can be determined using existing algorithms by processing the output signals included in each multi-dimensional runtime timing data sample (i.e., determining the integral nonlinearity error, differential nonlinearity error, spurious-free dynamic range, and ISI corresponding to the output signal in each of the multi-dimensional runtime timing data samples), and will not be elaborated further here.

[0045] In one embodiment, the preset first standard deviation threshold and the preset second standard deviation threshold represent pre-set standards for judging the stability of integral nonlinear error and differential nonlinear error, which can be preset according to accuracy requirements. After determining the above-mentioned quality indicators corresponding to each multi-dimensional runtime sequence data sample, the integral nonlinear error and differential nonlinear error of each digital input codeword corresponding to the same core temperature can be obtained, and their corresponding standard deviations can be determined respectively. Since the standard deviation can reflect the discreteness of each error, at the same core temperature, when the first standard deviation between the integral nonlinear errors corresponding to the same digital input codeword is less than the preset first standard deviation threshold, and / or the second standard deviation between the differential nonlinear errors corresponding to the same digital input codeword is less than the preset second standard deviation threshold, it indicates that at the same temperature, the integral nonlinear error and / or differential nonlinear error of the same codeword remain basically unchanged, which conforms to the characteristics of static nonlinear error.

[0046] In one embodiment, the transition frequency of a digital input codeword refers to the frequency at which the digital input codeword changes per unit time. The transition amplitude of a digital input codeword refers to the magnitude of the change in the level value (or codeword value) during a single transition, typically measured in LSBs. The digital input codewords can be monitored by a digital interface monitoring module, and their distribution can be statistically analyzed to determine the corresponding transition frequency and amplitude. In this embodiment, after obtaining the spurious-free dynamic range and inter-symbol interference (ISI) corresponding to each multi-dimensional runtime sequence data sample, the multi-dimensional runtime sequence data can be arranged from largest to smallest according to transition frequency or transition amplitude to obtain a data sequence. Based on this, the spurious-free dynamic range and ISI corresponding to each multi-dimensional runtime sequence data can be obtained according to the order of the data sequence, and their corresponding rates of change can be fitted. When the spurious-free dynamic range decreases as the transition frequency or amplitude of the corresponding digital input codeword increases (e.g., the rate of change is negative), and / or the inter-symbol interference increases as the transition frequency or amplitude of the corresponding digital input codeword increases (e.g., the rate of change is positive), it indicates that the spurious-free dynamic range and inter-symbol interference are related to the transition mode of the digital input codeword. The error increases as the high-frequency signal component increases, which is consistent with the characteristics of dynamic nonlinear error.

[0047] In one embodiment, a preset jitter threshold represents a pre-set standard for judging whether clock jitter is excessive, which can be set in advance according to the required accuracy. In this embodiment, duty cycle distortion indicates that the ratio of the high-level time of the clock signal to the period has changed, deviating from the ideal duty cycle. Duty cycle distortion can be identified by detecting the harmonic components of the output signal (e.g., the second harmonic amplitude is greater than a preset amplitude). In this embodiment, when the quality indicators corresponding to each multi-dimensional runtime sequence data sample do not meet the preset first characteristic condition and the preset second characteristic condition, it indicates that the error of the current multi-dimensional runtime sequence data is unrelated to the change of the digital input codeword. Under this premise, when the clock jitter is greater than the preset jitter threshold, and / or when duty cycle distortion is detected based on the harmonic components of the output signal, it indicates that the error of the multi-dimensional runtime sequence data sample is directly related to the quality of the sampling clock, which conforms to the characteristics of clock-related errors.

[0048] In this embodiment, by determining various quality indicators (i.e., integral nonlinearity error, differential nonlinearity error, spurious-free dynamic range, and inter-symbol interference, etc.) in each multidimensional runtime sequence data, and based on different preset feature conditions, it is beneficial to accurately identify the error feature performance of the multidimensional runtime sequence data samples corresponding to each historical time-series feature within the cluster, which is beneficial to ensuring the reliability of error type identification.

[0049] Optionally, the digital input codeword includes sub-codewords corresponding to multiple sub-intervals, which are obtained by dividing the digital input codeword based on a preset range interval; the error index includes sub-error indices corresponding to the sub-codewords within each sub-interval; after determining the error index based on the ideal output voltage and the actual output voltage, it also includes: Obtain the error threshold corresponding to each sub-interval pre-associated with the current error source. When the sub-error index corresponding to any sub-interval is greater than the corresponding error threshold, it is determined that the error index meets the preset calibration conditions.

[0050] Specifically, in this embodiment, the sub-interval refers to different intervals obtained by dividing the numeric input codeword based on a preset range interval. For example, for an 8-bit numeric input codeword (range 0-255), assuming a preset range interval of 32, it can be divided into multiple sub-intervals such as 0-31, 32-63, and 64-95. In this embodiment, the sub-codeword refers to the numeric input codeword within each sub-interval. The error index (such as at least one of integral nonlinearity, differential nonlinearity, and spurious-free dynamic range) corresponding to the numeric input codeword within each sub-interval can be determined to obtain each sub-error index.

[0051] In this embodiment, after dividing the digital input codewords based on a preset range interval to obtain sub-codewords corresponding to multiple sub-intervals, the sub-error index corresponding to each sub-codeword can be determined. Based on this, since the digital input codewords in different sub-intervals may correspond to different operating states or characteristics of the digital-to-analog converter, this embodiment can set differentiated error thresholds for different sub-intervals for different error sources (e.g., manually set in advance according to the characteristics of different sub-intervals under different error sources). After obtaining the error thresholds corresponding to each sub-interval pre-associated with the current error source, it can be determined whether the sub-error index corresponding to each sub-interval is greater than the corresponding error threshold. When the sub-error index corresponding to any sub-interval is greater than the corresponding error threshold, it can be determined that the error index meets the preset calibration conditions. This helps to avoid untimely or over-calibrated calibration due to the overall error index masking local errors, further improving the rationality and reliability of calibration.

[0052] Optionally, the digital-to-analog converter calibration method further includes: The error index corresponding to each sub-interval is obtained according to the preset monitoring period; wherein, the preset monitoring period is longer than the sampling period of the multi-dimensional runtime sequence data; When the rate of change of the sub-error index corresponding to each sub-interval with the same range is less than the preset rate of change for a preset number of consecutive preset times, the error threshold corresponding to that sub-interval is increased according to the preset gradient. When the rate of change of the sub-error index corresponding to each sub-interval with the same range is greater than the preset rate of change, the error threshold corresponding to that sub-interval is reduced according to the preset gradient, and the preset monitoring cycle is shortened.

[0053] Specifically, the preset monitoring period referred to in this embodiment represents a pre-set time interval for monitoring the error indicators of each sub-interval of the digital-to-analog converter. The preset monitoring period is longer than the sampling period of the multi-dimensional runtime sequence data, meaning that the same set of multi-dimensional runtime sequence data corresponding to each sub-error indicator will not be repeatedly collected within the monitoring period, avoiding repeated analysis of error performance based on a single set of multi-dimensional runtime sequence data. The preset rate of change referred to in this embodiment represents a pre-set standard value (e.g., 0.1 LSB / ms) used to determine whether the change in the error indicator is significant. When the rate of change of the error indicator is less than this preset rate of change, the change in the error indicator is considered insignificant. In this embodiment, for a certain error indicator, the error thresholds of each sub-interval corresponding to the same error source can be set to the same initial value. In subsequent actual operation, the error thresholds of the sub-intervals can be dynamically adjusted according to the error conditions of different sub-intervals.

[0054] In one embodiment, different sub-intervals correspond to different codeword ranges. Sub-error indices corresponding to sub-intervals with the same range can be arranged according to the acquisition order. The rate of change of the sub-error indices is determined based on the currently acquired sub-error indices and those acquired in the previous preset monitoring period. When the rate of change of the sub-error indices corresponding to sub-intervals with the same range is less than the preset rate of change for a preset number of consecutive preset times (e.g., three times), it indicates that the changes in the sub-error indices acquired within three consecutive preset monitoring periods are not significant, reflecting that the current digital-to-analog converter is operating relatively smoothly and the error is stabilizing. In this case, increasing the error threshold corresponding to the sub-interval according to a preset gradient (e.g., 0.1 to 0.5) helps reduce the sensitivity of the sub-interval to triggering calibration operations and avoids over-calibration. Conversely, when the rate of change of the sub-error indices corresponding to sub-intervals with the same range is greater than the preset rate of change, it indicates that the current digital-to-analog converter is significantly affected by error sources. In this case, decreasing the error threshold corresponding to the sub-interval according to a preset gradient (e.g., 0.05 to 0.2) and shortening the preset monitoring period helps improve the sensitivity of the sub-interval to triggering calibration operations and also facilitates rapid response. It should be understood that the permissible range of each error threshold can be preset in this embodiment. When the adjusted error threshold is less than the minimum value of the permissible range or greater than the maximum value of the permissible range, the corresponding range boundary value is used as the error threshold, which helps to avoid unreasonable adjustment.

[0055] Optionally, after obtaining the current error source corresponding to the multidimensional runtime sequence data based on the error source corresponding to the preset error feature center, the method further includes: The target probability of the current error source shifting to other error sources is determined based on the acquired historical dataset; wherein, the historical dataset includes historical error sources corresponding to each historical multidimensional runtime sequence data; When the target probability is greater than the preset probability threshold, the error source corresponding to the target probability is taken as the target error source, and the target error type corresponding to the target error source is determined. Obtain the calibration operation pre-associated with the target error type, obtain the target calibration operation, and execute the preparation operation corresponding to the target calibration operation; wherein, the preparation operation includes determining the calibration parameter adjustment value.

[0056] Specifically, in this embodiment, the historical dataset includes historical error sources corresponding to each historical multidimensional runtime sequence data. The historical multidimensional runtime sequence data can include multidimensional runtime sequence data collected during the past operation of the digital-to-analog converter, representing the error sources determined based on this data. The target probability of the current error source transitioning to other error sources can be determined based on the historical dataset. For example, a preset prediction model (such as a neural network model) can be trained based on the historical dataset to learn the error source transition patterns during the actual operation of the digital-to-analog converter. In practical use, the currently acquired multidimensional runtime sequence data and the corresponding current error source are input into the preset prediction model to predict the target probability of the current error source transitioning to other error sources.

[0057] In one embodiment, an M×N dimensional state transition matrix can be constructed based on a historical dataset, where M represents the number of error sources (M=4 in this embodiment) and N represents the number of sub-intervals (N=16 in this embodiment). The target probability of the current error source transitioning to other error sources satisfies: ; in, This represents the target probability of transitioning from error source i to error source j. This represents the number of times the error source shifts from error source i to error source j. This represents the total number of occurrences of error source i. It should be understood that the historical dataset in this embodiment may include the historical dataset of digital-to-analog converter samples. As the current digital-to-analog converter operates, the current error source corresponding to the current digital-to-analog converter can be used as supplementary data to the historical dataset, thereby updating the state transition matrix online in real time. This allows the state transition matrix to better reflect the actual operating characteristics of the digital-to-analog converter, improving the accuracy of state transition prediction.

[0058] In one embodiment, when the target probability is greater than a preset probability threshold (e.g., the transition probability of temperature drift to code pattern dependence is greater than 0.75), it indicates that the error source of the digital-to-analog converter may immediately switch to another type. The error source corresponding to the target probability can be taken as the target error source, and the target error type corresponding to the target error source can be determined (e.g., obtaining the error type pre-associated with the target error source to obtain the target error type). Based on this, the calibration operation pre-associated with the target error type can be obtained to obtain the target calibration operation, and the preparation operation corresponding to the target calibration operation can be executed. In this embodiment, the preparation operation refers to the preparations required before performing the calibration operation, which may include determining calibration parameter adjustment values. This facilitates completing the preparation work for the calibration operation in advance, so that after identifying that the current error source has switched to the target error source, the corresponding calibration operation can be quickly executed directly based on the determined calibration parameter adjustment values, ensuring the timeliness and reliability of the calibration.

[0059] Optionally, in this embodiment, after obtaining the target calibration operation, the target calibration operation can also be executed directly (the calibration operation corresponding to the current error type can be executed simultaneously), thereby pre-calibrating the errors that the target error source may cause, and further improving the accuracy of the digital-to-analog converter.

[0060] Optionally, determining calibration parameter adjustment values ​​includes: When the current error type is static nonlinear error, determine the calibration word of the current source array and obtain the calibration parameter adjustment value; When the current error type is dynamic nonlinear error, determine the adjustment coefficient and / or predistortion compensation value of the deemphasis filter to obtain the calibration parameter adjustment value; When the current error type is clock-related error, determine the control word of the phase interpolator and / or determine the bias current of the clock buffer stage to obtain the calibration parameter adjustment value.

[0061] In one embodiment, the current source array represents the current sources within the digital-to-analog converter (DAC) used to generate the analog output current. Different combinations of current sources can produce different output current values, which are then converted into different output voltages. In this embodiment, the calibration word refers to the parameter used to adjust the parameters of each current source in the current source array to correct static nonlinearity errors. By changing the calibration word, the output current of the current sources can be fine-tuned, making the output of the DAC closer to the ideal value. The calibration word of the current source array can be determined using methods such as least squares, thereby correcting static nonlinearity errors.

[0062] In one embodiment, the deemphasis filter represents a filter used to compensate for signal attenuation in the high-frequency band of the digital-to-analog converter (DAC). The dynamic nonlinearity error can be improved by adjusting the adjustment coefficient of the deemphasis filter. The adjustment coefficient of the deemphasis filter can be determined by comparing the actual output voltage (when the transition frequency or amplitude of the digital input codeword is greater than a preset frequency or amplitude, i.e., when a high-frequency component appears) with the ideal output voltage. When the actual output voltage is less than the ideal output voltage, the current coefficient is increased according to a preset gradient to obtain the adjustment coefficient; conversely, the current coefficient is decreased according to a preset gradient to obtain the adjustment coefficient. The predistortion compensation value referred to in this embodiment represents the pre-distortion processing of the input digital signal to offset the dynamic nonlinear distortion generated by the DAC during the conversion process. A predistortion lookup table (LUT) can be dynamically constructed or updated based on the nonlinear characteristics of the DAC. In practical use, the predistortion compensation value can be quickly obtained by looking up the table. Therefore, during subsequent calibration operations, the digital input codeword can be mapped to the pre-distorted codeword based on the pre-distortion compensation value, thereby achieving inverse distortion processing of the input digital signal and making the output signal after digital-to-analog conversion closer to the ideal signal.

[0063] In one embodiment, a phase interpolator represents a circuit module used to adjust the phase of a clock signal. In a digital-to-analog converter (DAC), the phase accuracy of the clock signal has a significant impact on the quality of the output signal. Its input is a differential clock signal from a phase-locked loop (PLL), and its output is the adjusted sampling clock. The adjustment step size of the phase interpolator is from 1 / 64 to 1 / 32 of the sampling clock period, with an adjustment range of ±1 / 4 of the sampling clock period. By adjusting the control word of the phase interpolator, the edge position of the sampling clock can be moved to minimize the aperture error at the sampling time. The control word of the phase interpolator can be determined based on the position of the sampling clock edge relative to the optimal sampling point. For example, when the sampling clock edge is to the left (i.e., leading) relative to the optimal sampling point, the current control word can be increased. When the sampling clock edge is to the right (i.e., lagging) relative to the optimal sampling point, the current control word can be decreased. During calibration, the phase of the clock signal can be adjusted by adjusting the control word of the phase interpolator, thereby correcting errors caused by clock phase problems. In this embodiment, the clock buffer stage refers to a circuit in the clock signal transmission path used to enhance the driving capability of the clock signal. In this embodiment, the current duty cycle can be determined using a duty cycle detector. When the current duty cycle is greater than 50%, the current bias current can be reduced. Conversely, when the current duty cycle is less than 50%, the current bias current can be increased. In this embodiment, the bias current of the clock buffer stage can affect the rising edge, falling edge, and stability of the clock signal. By determining the bias current and adjusting it during subsequent calibration operations, this embodiment helps improve the quality of the clock signal and thus correct clock-related errors.

[0064] In this embodiment, corresponding calibration parameter adjustment values ​​are determined for different error types, which is beneficial to achieve accurate calibration of different errors of the digital-to-analog converter. It can flexibly adapt to various error types that may occur in the digital-to-analog converter under different operating conditions, and improve the stability and reliability of the digital-to-analog converter in long-term operation in complex environments.

[0065] Optionally, the digital-to-analog converter calibration method in this embodiment can adopt two different working modes in actual use, including a background real-time calibration mode and a foreground calibration mode. In the background real-time calibration mode, the DAC outputs a normal signal, and the calibration operation is performed without interrupting the main signal. The timing diagram of the corresponding digital-to-analog converter calibration method is shown below. Figure 3 As shown, in the background real-time calibration mode, the DAC maintains continuous output of the main signal, and the calibration engine (i.e., the execution body of the digital-to-analog converter calibration method in this embodiment) completes calibration with a period of 1ms without interrupting the main signal. Within each 1ms period, the calibration engine identifies and utilizes idle time slots where the DAC digital input codeword has no transitions, executes the calibration method, and then feeds back the updated calibration parameters to the DAC, thereby achieving real-time online calibration. In the foreground calibration mode, calibration of the digital-to-analog converter is triggered when trigger conditions are met. These trigger conditions include: system power-on initialization, detection of control interface triggering, detection of ambient temperature changes exceeding 10°C, and detection of power supply voltage fluctuations exceeding 5%. The timing diagram of the corresponding digital-to-analog converter calibration method in the foreground calibration mode is shown below. Figure 4 As shown, when the trigger condition is met, the system sends a trigger calibration command to the DAC output module, and the DAC immediately pauses normal signal output. The calibration engine sends a scan codeword to the DAC output module, which performs voltage measurement and feeds the result back to the calibration engine. The calibration engine executes the aforementioned digital-to-analog converter calibration method to determine the calibration parameter adjustment value (i.e., calculation parameter), and then writes the calibration word into the DAC output module. After calibration is completed, the system sends a restore output command to the DAC output module, and the DAC resumes normal signal output. The entire calibration process lasts approximately 100μs.

[0066] like Figure 4 As shown, an embodiment of the present invention provides a digital-to-analog converter calibration device 400, comprising: The data acquisition module 410 is used to acquire multi-dimensional runtime timing data of the digital-to-analog converter; wherein, the multi-dimensional runtime timing data includes digital input codewords and output signals; The feature extraction module 420 is used to extract the temporal features of the multidimensional runtime sequence data using a preset feature extraction model, and to select preset error feature centers that match the temporal features from a preset set of error feature centers; wherein, one preset error feature center corresponds to one error source; The type determination module 430 is used to obtain the current error source corresponding to the multidimensional runtime sequence data based on the error source corresponding to the preset error feature center, and to determine the current error type corresponding to the multidimensional runtime sequence data; wherein, the current error source includes any one of temperature drift, power supply ripple, code pattern dependence and aging effect; the current error type includes any one of static nonlinear error, dynamic nonlinear error and clock-related error; Error determination module 440 is used to determine the ideal output voltage based on the digital input codeword, and to determine the actual output voltage based on the output signal, and to determine an error index based on the ideal output voltage and the actual output voltage; The calibration execution module 450 is used to execute a calibration operation pre-associated with the current error type when the error index meets the preset calibration conditions; wherein, the calibration operation includes determining a calibration parameter adjustment value and adjusting the corresponding calibration parameter based on the calibration parameter adjustment value.

[0067] The digital-to-analog converter calibration device and digital-to-analog converter calibration method provided in this embodiment can produce basically the same technical effects, and will not be described again here.

[0068] like Figure 5 As shown, an electronic device 500 provided in this embodiment of the invention includes a memory 510 and a processor 520; the memory 510 is used to store a computer program; the processor 520 is used to implement the digital-to-analog converter calibration method as described above when the computer program is executed.

[0069] Alternatively, an electronic device 500 includes a memory 510 and a processor 520 coupled to the memory 510; the memory 510 is configured to store a computer program; and the processor 520 is configured to perform the following operations when the computer program is executed: Acquire multidimensional runtime timing data of the digital-to-analog converter; wherein, the multidimensional runtime timing data includes digital input codewords and output signals; The temporal features of the multidimensional runtime sequence data are extracted using a preset feature extraction model, and preset error feature centers that match the temporal features are selected from a preset set of error feature centers; wherein, one preset error feature center corresponds to one error source; Based on the error source corresponding to the preset error feature center, the current error source corresponding to the multidimensional runtime sequence data is obtained, and the current error type corresponding to the multidimensional runtime sequence data is determined; wherein, the current error source includes any one of temperature drift, power supply ripple, code pattern dependence, and aging effect; the current error type includes any one of static nonlinear error, dynamic nonlinear error, and clock-related error; The ideal output voltage is determined based on the digital input codeword, and the actual output voltage is determined based on the output signal. An error index is then determined based on the ideal output voltage and the actual output voltage. When the error index meets the preset calibration conditions, the calibration operation pre-associated with the current error type is executed; wherein, the calibration operation includes determining the calibration parameter adjustment value and adjusting the corresponding calibration parameter based on the calibration parameter adjustment value.

[0070] The electronic device and digital-to-analog converter calibration method provided in this embodiment can produce basically the same technical effects, and will not be described again here.

[0071] This invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the digital-to-analog converter calibration method as described above.

[0072] Alternatively, a non-volatile computer-readable storage medium storing a computer program that, when executed by a processor, causes the processor to perform the following operations: Acquire multidimensional runtime timing data of the digital-to-analog converter; wherein, the multidimensional runtime timing data includes digital input codewords and output signals; The temporal features of the multidimensional runtime sequence data are extracted using a preset feature extraction model, and preset error feature centers that match the temporal features are selected from a preset set of error feature centers; wherein, one preset error feature center corresponds to one error source; Based on the error source corresponding to the preset error feature center, the current error source corresponding to the multidimensional runtime sequence data is obtained, and the current error type corresponding to the multidimensional runtime sequence data is determined; wherein, the current error source includes any one of temperature drift, power supply ripple, code pattern dependence, and aging effect; the current error type includes any one of static nonlinear error, dynamic nonlinear error, and clock-related error; The ideal output voltage is determined based on the digital input codeword, and the actual output voltage is determined based on the output signal. An error index is then determined based on the ideal output voltage and the actual output voltage. When the error index meets the preset calibration conditions, the calibration operation pre-associated with the current error type is executed; wherein, the calibration operation includes determining the calibration parameter adjustment value and adjusting the corresponding calibration parameter based on the calibration parameter adjustment value.

[0073] The computer-readable storage medium and the digital-to-analog converter calibration method provided in this embodiment produce essentially the same technical effects, and will not be described again here.

[0074] Electronic device 500, which can serve as a server or client of the present invention, is described below as an example of a hardware device applicable to various aspects of the present invention. Electronic device 500 is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic device 500 can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0075] Electronic device 500 includes a computing unit that can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) or a computer program loaded from a storage unit into random access memory (RAM). The RAM may also store various programs and data required for device operation. The computing unit, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0076] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. In this application, the units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention according to actual needs. Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units can be implemented in hardware or as software functional units.

[0077] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A method for calibrating a digital-to-analog converter, characterized in that, include: Acquire multidimensional runtime timing data of the digital-to-analog converter; wherein, the multidimensional runtime timing data includes digital input codewords and output signals; The temporal features of the multidimensional runtime sequence data are extracted using a preset feature extraction model, and preset error feature centers that match the temporal features are selected from a preset set of error feature centers; wherein, one preset error feature center corresponds to one error source; Based on the error source corresponding to the preset error feature center, the current error source corresponding to the multidimensional runtime sequence data is obtained, and the current error type corresponding to the multidimensional runtime sequence data is determined; wherein, the current error source includes any one of temperature drift, power supply ripple, code pattern dependence, and aging effect; the current error type includes any one of static nonlinear error, dynamic nonlinear error, and clock-related error; The ideal output voltage is determined based on the digital input codeword, and the actual output voltage is determined based on the output signal. An error index is then determined based on the ideal output voltage and the actual output voltage. When the error index meets the preset calibration conditions, the calibration operation pre-associated with the current error type is executed; wherein, the calibration operation includes determining the calibration parameter adjustment value and adjusting the corresponding calibration parameter based on the calibration parameter adjustment value.

2. The digital-to-analog converter calibration method according to claim 1, characterized in that, Before selecting the preset error feature centers that match the time series features from the preset set of error feature centers, the method further includes: Obtain each multidimensional runtime sequence data sample and the error source label corresponding to each multidimensional runtime sequence data sample, and use the preset feature extraction model to extract the historical time series features corresponding to each multidimensional runtime sequence data sample. Density clustering is performed on each of the historical time-series features to obtain multiple clusters. The error feature center corresponding to each cluster is used as the preset error feature center, and the preset error feature center set is constructed. The error sources of the preset error feature centers are determined based on the error source labels corresponding to the clusters, and the error sources are associated with the corresponding preset error feature centers; The error type is determined based on the multidimensional runtime sequence data samples corresponding to the cluster, and the error type is associated with the error feature center.

3. The digital-to-analog converter calibration method according to claim 2, characterized in that, The multidimensional runtime timing data also includes kernel temperature and clock jitter; the determination of error type based on each multidimensional runtime timing data sample corresponding to the cluster includes: Determine the integral nonlinearity error, differential nonlinearity error, spurious-free dynamic range, and inter-symbol interference for each of the multidimensional runtime sequence data samples. When a preset first characteristic condition is met, the error type is the static nonlinear error; wherein, the preset first characteristic condition includes, at the same kernel temperature, the first standard deviation between the integral nonlinear errors corresponding to the same digital input codeword is less than a preset first standard deviation threshold, and / or, the second standard deviation between the differential nonlinear errors corresponding to the same digital input codeword is less than a preset second standard deviation threshold. When the preset second characteristic condition is met, the error type is the dynamic nonlinear error; wherein, the preset second characteristic condition includes that each of the spurious-free dynamic ranges decreases as the jump frequency or jump amplitude corresponding to the digital input codeword increases, and / or, the inter-code interference increases as the jump frequency or jump amplitude corresponding to the digital input codeword increases. When the preset first characteristic condition and the preset second characteristic condition are not met and the clock jitter is greater than the preset jitter threshold, and / or when the preset first characteristic condition and the preset second characteristic condition are not met and duty cycle distortion is detected based on the harmonic components of the output signal, the error type is the clock-related error.

4. The digital-to-analog converter calibration method according to claim 1, characterized in that, The digital input codeword includes multiple sub-codewords corresponding to sub-intervals, and the sub-intervals are obtained by dividing the digital input codeword based on a preset range interval; the error index includes the sub-error index corresponding to the sub-codeword in each of the sub-intervals; After determining the error index based on the ideal output voltage and the actual output voltage, the method further includes: Obtain the error threshold corresponding to each of the sub-intervals pre-associated with the current error source. When the sub-error index corresponding to any sub-interval is greater than the corresponding error threshold, determine that the error index meets the preset calibration condition.

5. The digital-to-analog converter calibration method according to claim 4, characterized in that, The digital-to-analog converter calibration method further includes: The sub-error index corresponding to each sub-interval is obtained according to a preset monitoring period; wherein, the preset monitoring period is longer than the sampling period of the multi-dimensional runtime sequence data; When the rate of change of the sub-error index corresponding to each sub-interval with the same range is less than the preset rate of change for a preset number of consecutive preset times, the error threshold corresponding to that sub-interval is increased according to the preset gradient. When the rate of change of the sub-error index corresponding to each sub-interval with the same range is greater than the preset rate of change, the error threshold corresponding to the sub-interval is reduced according to the preset gradient, and the preset monitoring period is shortened.

6. The digital-to-analog converter calibration method according to claim 1, characterized in that, After obtaining the current error source corresponding to the multidimensional runtime sequence data based on the error source corresponding to the preset error feature center, the method further includes: The target probability of the current error source shifting to other error sources is determined based on the acquired historical dataset; wherein, the historical dataset includes historical error sources corresponding to each historical multidimensional runtime sequence data; When the target probability is greater than a preset probability threshold, the error source corresponding to the target probability is taken as the target error source, and the target error type corresponding to the target error source is determined. Obtain the calibration operation pre-associated with the target error type, obtain the target calibration operation, and execute the preparation operation corresponding to the target calibration operation; wherein, the preparation operation includes determining the calibration parameter adjustment value.

7. The digital-to-analog converter calibration method according to claim 6, characterized in that, The determination of calibration parameter adjustment values ​​includes: When the current error type is the static nonlinear error, determine the calibration word of the current source array and obtain the calibration parameter adjustment value; When the current error type is the dynamic nonlinear error, determine the adjustment coefficient and / or predistortion compensation value of the deemphasis filter to obtain the calibration parameter adjustment value; When the current error type is the clock-related error, determine the control word of the phase interpolator and / or determine the bias current of the clock buffer stage to obtain the calibration parameter adjustment value.

8. A digital-to-analog converter calibration device, characterized in that, include: The data acquisition module is used to acquire multi-dimensional runtime timing data of the digital-to-analog converter; wherein, the multi-dimensional runtime timing data includes digital input codewords and output signals; The feature extraction module is used to extract the temporal features of the multidimensional runtime sequence data using a preset feature extraction model, and to filter out preset error feature centers that match the temporal features from a preset set of error feature centers; wherein, one preset error feature center corresponds to one error source; The type determination module is used to obtain the current error source corresponding to the multidimensional runtime sequence data based on the error source corresponding to the preset error feature center, and to determine the current error type corresponding to the multidimensional runtime sequence data; wherein, the current error source includes any one of temperature drift, power supply ripple, code pattern dependence, and aging effect; the current error type includes any one of static nonlinear error, dynamic nonlinear error, and clock-related error; An error determination module is used to determine an ideal output voltage based on the digital input codeword, and to determine an actual output voltage based on the output signal, and to determine an error index based on the ideal output voltage and the actual output voltage; The calibration execution module is used to execute a pre-associated calibration operation for the current error type when the error index meets the preset calibration conditions; wherein the calibration operation includes determining a calibration parameter adjustment value and adjusting the corresponding calibration parameter based on the calibration parameter adjustment value.

9. An electronic device, characterized in that, Including memory and processor; The memory is used to store computer programs; The processor is configured to implement the digital-to-analog converter calibration method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program that, when executed by a processor, implements the digital-to-analog converter calibration method as described in any one of claims 1 to 7.

Images