Digital compensation-based DAC output error correction method

Abstract

The invention discloses a DAC output error correction method based on digital compensation, and the method comprises the steps: collecting the amplitude deviation, phase deviation and time sequence jitter data of a DAC output signal through a quantum sensing DAC calibration analysis platform, and constructing a multi-dimensional error original data set; calling an operational amplifier imbalance dynamic coupling prediction model to analyze a coupling relationship, and extracting dynamic association features; separating harmonic components and screening error contribution factors based on a broadband DAC harmonic distortion correction model; calculating a gradient change rule and an extreme point position by adopting a differential nonlinear gradient correction model; and integrating the characteristics and the parameters to construct a digital compensation parameter matrix, and dynamically adjusting an output code value through real-time digital modulation to realize error correction. According to the method, dynamic association and distribution characteristics of multi-source errors are accurately captured through multi-model collaborative operation and full-process closed-loop processing, multi-dimensional error synchronous correction is achieved, and DAC output precision and working condition adaptability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital-to-analog error correction, and particularly relates to a DAC output error correction method based on digital compensation. BACKGROUND

[0002] In the case of electronic information, communication engineering, etc., the digital-to-analog converter (DAC) as the core device for converting digital signals to analog signals, its output precision directly affects the performance of the entire electronic system. With the rapid development of wideband communication, quantum sensing and other technologies, higher requirements are put forward for the output bandwidth, conversion rate and linearity of the DAC. However, during the operation of the DAC, the dynamic change of the operational amplifier (op-amp) offset voltage, the cumulative effect of harmonic distortion and the error caused by the differential nonlinearity will cause the output signal to have amplitude deviation, phase shift and timing jitter, etc., which seriously affects the signal transmission quality and system operation stability. The traditional DAC calibration method relies on a single error correction mechanism, which is difficult to consider the coupling effect of multiple error sources, and an integrated multi-dimensional error factor and accurate dynamic compensation technology is needed to meet the stringent requirements of high-precision electronic systems for DAC output precision.

[0003] The existing technology has two significant shortcomings: on the one hand, the existing error correction method does not fully consider the dynamic coupling characteristics of the op-amp offset, and only uses a static calibration mode to compensate for a single error source, which cannot capture the dynamic relationship between the op-amp offset and harmonic distortion, differential nonlinearity error, resulting in a significant decrease in correction effect when multiple error sources are superimposed, making it difficult to adapt to error change patterns under complex working conditions; on the other hand, the existing technology lacks an integrated calibration analysis platform and multi-model collaborative correction mechanism, and does not combine quantum sensing technology with harmonic distortion correction and differential nonlinearity gradient correction, which cannot realize high-precision collection of error data and synchronous correction of multi-dimensional errors, and the construction of digital compensation parameters does not fully integrate error gradient change patterns and extreme point distribution characteristics, resulting in insufficient adaptability of compensation parameters, making it difficult to realize real-time dynamic correction of DAC output errors. SUMMARY

[0004] In order to overcome the shortcomings and deficiencies of the existing technology, the present application provides a DAC output error correction method based on digital compensation.

[0005] The technical scheme adopted by the present application is a DAC output error correction method based on digital compensation, comprising the following steps: S1, collecting amplitude deviation, phase offset and timing jitter data of a DAC output signal through a quantum sensing DAC calibration analysis platform, and establishing a multi-dimensional error original data set; S2, calling an operational amplifier misadjustment dynamic coupling prediction model to analyze the coupling relationship of the error original data set, and extracting the dynamic correlation characteristics of the operational amplifier misadjustment voltage and the output error; S3, separating the harmonic components of the coupling correlation characteristics based on a wideband DAC harmonic distortion correction model, and screening out error contribution factors corresponding to each harmonic; S4, performing gradient operation on the harmonic error contribution factors using a differential nonlinear gradient correction model, and determining the gradient change rule and extreme point position of the error distribution; S5, constructing a digital compensation parameter matrix combining the dynamic correlation characteristics, the error contribution factors and the gradient change rule, wherein the digital compensation parameter matrix comprises amplitude compensation coefficients, phase calibration coefficients and timing adjustment parameters; S6, performing real-time digital modulation on the DAC output signal according to the digital compensation parameter matrix, correcting the error by dynamically adjusting the output code value, and the digital modulation process synchronously responds to the change of the error distribution extreme point position.

[0006] Further, the expression of the operational amplifier misadjustment dynamic coupling prediction model is: is the initial operational amplifier misadjustment voltage, is the coupling coefficient, is the initial operational amplifier misadjustment voltage, is the clock frequency influence factor, is the system clock frequency, is the output voltage deviation sensitivity coefficient, is the DAC output voltage deviation, is the superposition weight coefficient, is the i-th operational amplifier misadjustment voltage, is the i-th phase deviation influence coefficient, is the i-th signal phase deviation.

[0007] Further, the expression of the wideband DAC harmonic distortion correction model is: in which, is the harmonic distortion correction coefficient, is the fundamental amplitude, is the fundamental angular frequency, is the time variable, is the initial phase of the fundamental, is the harmonic order, is the m-th harmonic amplitude correction factor, is the m-th harmonic amplitude, is the initial phase of the m-th harmonic, is the m-th harmonic integral correction factor,​ is an integral variable.

[0008] Further, the expression of the differential nonlinear gradient correction model is: wherein, is a differential nonlinear gradient value, is a gradient operator, is an original differential nonlinear coefficient, is an original differential nonlinear error value, is a first-order derivative weight coefficient, is a DAC input code value variable, is a second-order derivative square weight coefficient.

[0009] Further, the error collection model expression of the quantum sensing DAC calibration analysis platform is: wherein, is a comprehensive collection error value, is a collection accuracy coefficient, is an amplitude deviation, is a phase shift, is a timing jitter, is a covariance weight coefficient, is a three-dimensional covariance of the amplitude deviation, the phase shift, and the timing jitter.

[0010] Further, the construction model expression of the digital compensation parameter matrix is: wherein, is a digital compensation parameter matrix, is an amplitude compensation weight, is a phase compensation weight, is a timing compensation weight, is an amplitude-phase cross compensation coefficient, is a phase-timing cross compensation coefficient, is a timing-amplitude cross compensation coefficient.

[0011] Further, the S3 includes the following steps: S31, inputting the harmonic separation module of the harmonic distortion correction model of the wideband DAC associated with the characteristics of the dynamic coupling of the op-amp mismatch into the harmonic separation module, and decomposing the composite signal into the fundamental component and each harmonic component through frequency domain transformation; S32, extracting the amplitude and phase of each harmonic component after decomposition, and recording the characteristic parameters of the harmonic components at different frequencies; S33, calculating the contribution proportion of each harmonic to the DAC output error based on the characteristic parameters, and eliminating the harmonic components whose contribution proportion is lower than the set threshold; S34, mapping the characteristic parameters corresponding to the retained high-contribution harmonic components and the error data to form a harmonic error contribution factor set.

[0012] Further, the S4 comprises the following steps: S41, introducing the harmonic error contribution factor into the gradient calculation unit of the differential non-linear gradient correction model, setting the input code value step length, and traversing the entire input code value range; S42, calculating the differential non-linear error value corresponding to each input code value, and solving the error gradient value through the error difference value of adjacent code values; S43, performing sliding window filtering processing on the gradient value to eliminate the interference of random noise on the gradient change law; S44, judging the rising and falling intervals of the error distribution through the positive and negative changes of the gradient value, and positioning the error extreme point position where the gradient value is zero.

[0013] Further, the S5 comprises the following steps: S51, collecting the coupling coefficient in the dynamic correlation feature, the harmonic amplitude and phase parameters in the error contribution factor, and the gradient extreme value data in the gradient change law, and establishing an original set of compensation parameters; S52, filtering and standardizing the parameters in the original set according to the parameter constraint conditions of the digital compensation model, and eliminating invalid parameters; S53, classifying and distributing the standardized parameters to the corresponding compensation dimensions according to the functional requirements of amplitude compensation, phase calibration and timing adjustment; S54, integrating the compensation parameters of each dimension through matrix operation to generate a digital compensation parameter matrix matched in dimensions.

[0014] A DAC output error correction method based on digital compensation, which is realized by different units, comprising: a quantum sensing multi-dimensional error acquisition unit for acquiring amplitude deviation, phase offset and timing jitter data of a DAC output signal and generating an error original data set, which is connected to an operational amplifier misadjustment dynamic coupling feature analysis unit; the operational amplifier misadjustment dynamic coupling feature analysis unit is used to call the operational amplifier misadjustment dynamic coupling prediction model to analyze the coupling relationship in the error original data set, and extract dynamic correlation features, which are connected to a wideband DAC harmonic error separation unit; the wideband DAC harmonic error separation unit is used to separate the harmonic component in the coupling feature based on the wideband DAC harmonic distortion correction model, and filter error contribution factors, which are connected to a differential non-linear gradient operation unit; the differential non-linear gradient operation unit is used to calculate the gradient change law of the error contribution factor through the differential non-linear gradient correction model, and locate the extreme point position, which is connected to a digital compensation parameter matrix construction unit; the digital compensation parameter matrix construction unit is used to construct a compensation parameter matrix including amplitude compensation coefficient, phase calibration coefficient and timing adjustment parameter by combining dynamic correlation feature, error contribution factor and gradient change law, which is connected to a DAC output real-time modulation unit; the DAC output real-time modulation unit is used to dynamically adjust the code value of the DAC output signal according to the digital compensation parameter matrix, and correct the error in response to the change of the extreme point position, which is connected to the output end of the digital compensation parameter matrix construction unit, and directly connected to the DAC output link.

[0015] Beneficial effects: The present application proposes a DAC output error correction method based on digital compensation, uses a calibration analysis platform constructed by quantum sensing technology to efficiently collect multi-dimensional error original data, combines an op-amp misadjustment dynamic coupling prediction model to deeply analyze the dynamic correlation characteristics between multi-source errors, completely changes the limitations of the prior art static calibration that cannot cope with error coupling effects, enables error correction to accurately match the error change law under complex working conditions, and realizes harmonic component separation, error contribution factor screening and gradient change law analysis through the collaborative operation of a wideband DAC harmonic distortion correction model and a differential nonlinear gradient correction model, and then integrates multi-dimensional error parameters to construct a digital compensation parameter matrix, forms a multi-model collaborative correction mechanism, and makes up for the lack of an integrated calibration platform and synchronous correction capability in the prior art. At the same time, based on the dynamic adjustment of the digital modulation strategy based on the error gradient extreme value point position, the compensation parameters are ensured to be real-time adapted to the error distribution, the synchronous and accurate correction of multi-source errors is realized, the amplitude, phase and timing accuracy of the DAC output signal are greatly improved, the strict requirements of high-precision electronic systems on the output accuracy of the DAC are effectively met, and the core problems of the prior art, such as limited correction effect and insufficient adaptability, are fundamentally solved. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The method flowchart of the present application is shown in the figure.

[0017] Figure 2 The method implementation unit composition diagram of the present application is shown in the figure. DETAILED DESCRIPTION

[0018] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict, and the present application will be further described in detail below in combination with the drawings and specific embodiments.

[0019] As shown in the figure, a DAC output error correction method based on digital compensation comprises the following steps: Figure 1 S1, collect the amplitude deviation, phase offset and timing jitter data of the DAC output signal through the quantum sensing DAC calibration analysis platform, and establish a multi-dimensional error original data set;

[0020]

[0021] ​Specifically, the implementation process of step S1 is: starting the quantum sensing DAC calibration analysis platform, accurately setting the sampling frequency to 5 million data points per second, setting the collection time to 30 minutes, fully covering the full working voltage range of 0 to full range of DAC input voltage and the frequency interval of 1 kHz to 100 MHz, focusing on collecting three types of core error data of amplitude deviation, phase offset and timing jitter of output signal, wherein the amplitude deviation collection accuracy is strictly controlled within ±5 microvolts, the phase offset collection resolution is set to 0.1 milliradian, and the timing jitter collection minimum time interval is 10 nanoseconds. Through the built-in 8-channel synchronous acquisition module of the platform, the output signal is systematically collected under the conditions of full range of DAC input code value from 0000H to FFFFH and load resistance from 10Ω to 1kΩ, and 2000 groups of sample data are collected under each working condition to ensure that all typical working conditions such as light load, heavy load, low frequency and high frequency are covered. During the collection process, relying on the electromagnetic shielding structure and adaptive noise suppression algorithm built in the quantum sensing unit, the external electromagnetic interference in the frequency range of 20MHz to 1GHz and the environmental noise with amplitude lower than 2 microvolts are effectively filtered, and finally all the collected multi-dimensional data are classified and integrated according to the time stamp, input code value and load condition, forming a multi-dimensional error original data set including 1.2 million effective records, which includes the error dynamic change information of three dimensions of amplitude, phase and timing. The data set provides high-fidelity, full-scene basic data support for subsequent error coupling relationship analysis and compensation parameter construction, and the completeness and accuracy of data collection directly determine the targeting and final effect of subsequent error correction process.

[0022] S2, calling the op-amp mismatch dynamic coupling prediction model to analyze the coupling relationship of the error original data set, and extracting the dynamic correlation characteristics of the op-amp mismatch voltage and the output error;

[0023] Specifically, the implementation process of step S2 is as follows: the op-amp misadjustment dynamic coupling prediction model trained in advance based on a large amount of measured data is called, and the multi-dimensional error original data set obtained in step S1 is input into the model after being segmented according to time sequence. The model analyzes the dynamic correlation between the op-amp misadjustment voltage and the output error in the data through a 3-layer convolution feature extraction network and a 2-layer fully connected analysis network. When implemented, first, the coupling relationship analysis threshold of the model is set, and the abnormal data whose amplitude deviation exceeds 5 microvolts and phase shift exceeds 0.1 milliradian in the error data are marked as key analysis objects. At the same time, a fixed time window length of 10 milliseconds is set, and the segmented analysis is performed on the continuously collected data to accurately capture the dynamic characteristics of the error changing with time. The model calculates the Pearson correlation coefficient of the op-amp misadjustment voltage and the amplitude deviation, phase shift, and timing jitter in each time window, selects the strong correlation combination with an absolute value of the correlation coefficient higher than 0.8, and then extracts the quadratic function variation law of the op-amp misadjustment voltage changing with time, input code value, and load condition, as well as the mapping relationship between the law and various output errors through linear fitting and nonlinear regression algorithms, to finally form a dynamic correlation feature set including 20 core feature parameters. In this process, the model adjusts the convolution kernel size and the fully connected layer weight adaptively by monitoring the data distribution characteristics in real time, so as to ensure that the core correlation features can be accurately extracted under complex working conditions such as high frequency and heavy load, and provide analysis objects with strong pertinence and high recognition for subsequent harmonic component separation and gradient operation. The accuracy of feature extraction directly affects the accuracy and efficiency of the entire error correction process.

[0024] S3, separating harmonic components from the coupled correlation features based on a wideband DAC harmonic distortion correction model, and screening error contribution factors corresponding to each harmonic;

[0025] Specifically, the implementation process of step S3 is: based on the wideband DAC harmonic distortion correction model, the dynamic correlation features extracted in step S2 are subjected to harmonic component separation processing. When implementing, the frequency analysis range of the model is first set to 1 kHz to 100 MHz, which is completely matched with the DAC working bandwidth, and the upper limit of harmonic order analysis is set to 15 orders, which ensures that the main harmonic components affecting the output accuracy are included. The model decomposes the complex signal corresponding to the dynamic correlation features into fundamental component and 1st to 15th harmonic components through fast Fourier transform algorithm, and then calculates the peak amplitude, initial phase and center frequency parameters of each harmonic component through amplitude detection module and phase detection module respectively. Through the amplitude ratio and phase difference calculation of the fundamental component, the influence degree of each harmonic on the output error is quantitatively determined. On this basis, an error contribution factor screening threshold of 0.05 is set, and the amplitude ratio and phase offset influence degree of each harmonic component are weighted calculated with the weight of 0.6 and 0.4, and the harmonic components with weighted score exceeding the threshold are selected as the key error sources to form the error contribution factor set including 3 to 8 core harmonic parameters. In the implementation process, the model dynamically adjusts the frequency domain resolution to 10 Hz to ensure the accuracy of harmonic component separation, avoid missing high contribution harmonic components due to insufficient resolution, and eliminate false harmonic signals caused by high frequency noise with frequency higher than 100 MHz through wavelet threshold denoising algorithm to ensure the reliability and purity of the error contribution factor, providing accurate and focused error analysis basis for subsequent differential nonlinear gradient operation, directly affecting the efficiency of gradient calculation and the accuracy of extreme point positioning.

[0026] S4, using a differential nonlinear gradient correction model to perform gradient operation on the harmonic error contribution factor to determine the gradient variation law and extreme point position of the error distribution;

[0027] Specifically, the implementation process of step S4 is as follows: the harmonic error contribution factor screened out in step S3 is subjected to gradient operation by using a differential nonlinear gradient correction model. In the implementation, the change step of the input code value is first set to 1 / 10 of 1 LSB, i.e., 0.1 LSB, to ensure the refinement and accuracy of the operation. The model takes the DAC input code value as the independent variable and the integrated error value corresponding to the error contribution factor as the dependent variable. The error difference between adjacent code values in the full range of input code values from 0000H to FFFFH is calculated point by point through an adjacent code value error difference calculation algorithm, the gradient value of the error distribution is obtained, and a continuous gradient change curve is constructed through a data fitting algorithm. Subsequently, a gradient change rate threshold of 0.02 is set, the slope change trend of the gradient change curve is analyzed by traversing the entire input code value range, the intervals of positive, negative and near-zero gradient values are determined, the rising section, falling section and stable section of the error distribution are accurately divided, and the area where the error extreme point may exist is determined. In the implementation process, a sliding window with a length of 50 code value units is used to perform moving average filtering on the gradient value, which effectively eliminates the interference of random noise with an amplitude less than 0.01 on the gradient calculation. At the same time, the extreme point determination condition is set. When the absolute values of the gradient values corresponding to five consecutive adjacent code values are all less than 0.005 and the gradient signs of the adjacent intervals before and after change reversely, it is determined that it is an error extreme point. Finally, the input code value position and the corresponding error amplitude of all extreme points are output, which provides key feature information of the error distribution for the construction of the subsequent digital compensation parameter matrix, ensures that the compensation strategy can respond to the error extreme area, and improves the overall error correction effect.

[0028] S5, constructing a digital compensation parameter matrix combining the dynamic correlation characteristics, error contribution factors and gradient change law, wherein the digital compensation parameter matrix includes amplitude compensation coefficients, phase calibration coefficients and timing adjustment parameters;

[0029] Specifically, the implementation process of step S5 is: comprehensively constructing the digital compensation parameter matrix by integrating the dynamic correlation features extracted in step S2, the error contribution factors screened in step S3 and the gradient variation law determined in step S4. In implementation, firstly, the maximum and minimum normalization algorithm is used to pre-process the three types of features and parameters, so as to uniformly map the data of different dimensions and different magnitudes to the numerical range of 0 to 1. Then, the weight coefficients of each parameter are set through the analytic hierarchy process, wherein the weight coefficient of the dynamic correlation feature is set to 0.35, the weight coefficient of the error contribution factor is set to 0.4, and the weight coefficient of the gradient variation law is set to 0.25. The weight coefficients of the subdivision parameters corresponding to the high contribution error are increased in proportion. Based on the set weight coefficients, the coupling coefficient in the dynamic correlation feature, the harmonic amplitude and phase parameters in the error contribution factor, and the gradient extreme value and extreme point position information in the gradient variation law are respectively classified and integrated into the amplitude compensation, phase calibration and time sequence adjustment three functional dimensions. In each dimension, the distribution law of each parameter is determined by statistical histogram analysis, and the value range of the compensation coefficient is determined, wherein the amplitude compensation coefficient covers the error deviation interval of -10 microvolts to +10 microvolts, the phase calibration coefficient corresponds to the phase shift range of -0.5 milliradians to +0.5 milliradians, and the time sequence adjustment parameter matches the time sequence jitter amplitude of -50 nanoseconds to +50 nanoseconds. Finally, the parameters in the three dimensions are integrated through matrix transposition and weighted summation operation to form a digital compensation parameter matrix with 65536 rows (corresponding to the input code value range) and 3 columns (corresponding to the three compensation dimensions). The integrity and accuracy of the matrix directly determine the error correction effect of the subsequent digital modulation process, and the parameter update frequency and the error data acquisition frequency are kept synchronous to ensure dynamic adaptation to error changes.

[0030] S6, real-time digital modulation of the DAC output signal according to the digital compensation parameter matrix, error correction through dynamic adjustment of the output code value, and the digital modulation process synchronously responding to the change of the error distribution extreme point position.

[0031] Specifically, the implementation process of step S6 is: strictly according to the digital compensation parameter matrix constructed in step S5, the DAC output signal is real-time digitally modulated, and when implemented, firstly, the one-to-one correspondence between the compensation parameter matrix and the DAC output code value is established through the lookup table mapping algorithm, the amplitude compensation coefficient, the phase calibration coefficient and the timing adjustment parameter in the matrix are respectively converted into the corresponding DAC output code value adjustment amount, wherein the amplitude compensation corresponds to the lowest 4-bit adjustment of the code value, the phase calibration corresponds to the middle 4-bit adjustment of the code value, and the timing adjustment corresponds to the highest 4-bit adjustment of the code value. The modulation update frequency is set to 1MHz, which is higher than the maximum output rate 500kHz of the DAC, to ensure real-time response to error changes, and the step precision of code value adjustment is set to 0.1LSB, which is consistent with the input code value step of step S4, to avoid overcompensation or undercompensation problems caused by mismatching of adjustment precision. In the modulation process, the extreme value point monitoring module is used to track the extreme value point position determined in step S4 in real time, and when it is monitored that the extreme value point position deviates more than 10 code value units, the update parameters in the corresponding region of the compensation parameter matrix are synchronously called to dynamically adjust the amplitude compensation amount, the phase calibration amount and the timing offset amount of the output code value. In the implementation process, the modulated DAC output signal is fed back to the quantum sensing DAC calibration analysis platform in real time through the closed-loop feedback link, the adjustment effect is indirectly verified, and if the feedback data shows that the output error still exceeds the set threshold, the corresponding parameters in the compensation parameter matrix are fine-tuned through the incremental adjustment algorithm to optimize the modulation strategy, so that in various working conditions such as different input code values, different loads and different working frequencies, the error can be accurately corrected by dynamically adjusting the output code value, and the amplitude stability, phase accuracy and timing consistency of the DAC output signal are significantly improved.

[0032] Preferably, the expression of the op-amp misadjustment dynamic coupling prediction model is: is the op-amp misadjustment dynamic coupling quantity, is the coupling coefficient, is the initial op-amp misadjustment voltage, is the clock frequency influence factor, is the system clock frequency, is the output voltage deviation sensitivity coefficient, is the DAC output voltage deviation, is the superposition weight coefficient, is the i-th op-amp misadjustment voltage, is the i-th phase deviation influence coefficient, is the i-th signal phase deviation.

[0033] ​Specifically, the op-amp misadjustment dynamic coupling prediction model is used to analyze the dynamic correlation between the op-amp misadjustment voltage and the DAC output error. In implementation, the coupling coefficient value range is set to 0.1 to 0.9, the measurement accuracy of the initial op-amp misadjustment voltage is controlled to be in the microvolt level, the clock frequency influence factor is set to 0.01 to 0.1 according to the actual range of the system clock frequency, the output voltage deviation sensitivity coefficient is calibrated according to 1% to 5% of the output voltage range, and the superposition weight coefficient is set to 0.3 to 0.7 through multiple actual measurement verifications. In specific application, the model first acquires initial op-amp misadjustment voltage data, combines the system clock frequency and the DAC output voltage deviation, quantifies the comprehensive influence of the clock frequency and the voltage deviation on the op-amp misadjustment through a sine function operation, integrates the correlation between each op-amp misadjustment voltage and the corresponding phase deviation through a logarithmic operation, and finally obtains the op-amp misadjustment dynamic coupling quantity. In the implementation process, for 1 to 8 op-amp channels, the phase deviation influence coefficient is set to 0.2 to 0.8 respectively to ensure that the misadjustment influence of each op-amp can be accurately quantified. The model dynamically integrates voltage, frequency and phase multi-dimensional parameters to realize accurate prediction of the op-amp misadjustment coupling effect and provide core parameter support for subsequent error correction. The prediction accuracy directly determines the accuracy of error analysis.

[0034] Preferably, the expression of the wideband DAC harmonic distortion correction model is: wherein, is a harmonic distortion correction coefficient, is a fundamental amplitude, is a fundamental angular frequency, is a time variable, is a fundamental initial phase, is a harmonic order, is an mth harmonic amplitude correction factor, is an mth harmonic amplitude, is an mth harmonic initial phase, is an mth harmonic integral correction factor, is an integral variable.

[0035] Specifically, the wideband DAC harmonic distortion correction model takes harmonic distortion correction as the core target, and sets the acquisition accuracy of the fundamental amplitude to be millivolt level, the measurement resolution of the initial phase of the fundamental wave to be 0.1 milliradian, the upper limit of the harmonic order to be 15 orders, covering the main distortion harmonic components, the harmonic amplitude correction factor to be 0.8 to 1.2, and the integral correction factor to be 0.05 to 0.2. In application, the model first extracts the amplitude and phase parameters of the fundamental wave and each harmonic through frequency domain analysis, then constructs the time domain signal expression of the fundamental wave and the harmonic through the cosine function, performs integral operation on each harmonic signal to compensate for the distortion caused by phase lag, and finally obtains the harmonic distortion correction coefficient through the ratio operation of the fundamental wave signal and the integrated harmonic signal. In the implementation process, the time variable is sampled at nanosecond level intervals, and the integral variable covers the complete period of the signal, ensuring comprehensive compensation for harmonic distortion. The model integrates amplitude correction and integral correction mechanisms to accurately correct each harmonic distortion in the wideband range, effectively improving the spectral purity of the DAC output signal and providing high-quality error data for subsequent gradient operation.

[0036] Preferably, the expression of the differential nonlinearity gradient correction model is: wherein, is the differential nonlinearity gradient value, is the gradient operator, is the original differential nonlinearity coefficient, is the original differential nonlinearity error value, is the first derivative weight coefficient, is the DAC input code value variable, is the second derivative square weight coefficient.

[0037] Specifically, the differential nonlinearity gradient correction model performs gradient analysis on the differential nonlinearity error, and sets the measurement accuracy of the original differential nonlinearity coefficient to be 0.01 LSB, the first derivative weight coefficient to be 0.3 to 0.6, the second derivative square weight coefficient to be 0.1 to 0.3, and the input code value variable to be adjusted in steps of one-tenth of the minimum code value. In application, the model first obtains the original differential nonlinearity error value, quantifies the error rate with respect to the input code value through first derivative operation, then strengthens the nonlinearity characteristics of the error rate through second derivative square operation, and finally obtains the differential nonlinearity gradient value through the gradient operator integrating the above parameters. In the implementation process, the input code value is operated in the full range to ensure the comprehensiveness of the gradient value calculation, and the gradient data is processed through sliding window filtering with a window length of 50 code value units to eliminate random noise interference. The model accurately captures the gradient change law of the differential nonlinearity error, provides a core basis for error extreme point positioning, and greatly improves the targeting and accuracy of error correction.

[0038] Preferably, the quantum sensing DAC calibration analysis platform error collection model expression is: wherein, is the integrated collection error value, is the collection accuracy coefficient, is the amplitude deviation, is the phase shift, is the timing jitter, is the covariance weight coefficient, is the three-dimensional covariance of the amplitude deviation, the phase shift and the timing jitter.

[0039] Specifically, the quantum sensing DAC calibration analysis platform error collection model is used to integrate multi-dimensional collection errors. When implemented, the collection accuracy coefficient is set to a value range of 0.9 to 0.99 to ensure the reliability of the collected data, and the covariance weight coefficient is set to 0.2 to 0.4 to balance the coupling effects of amplitude, phase and timing errors. In application, the model first collects amplitude deviation, phase shift and timing jitter data, wherein the amplitude deviation collection accuracy is microvolt level, the phase shift is milliradian level, and the timing jitter is nanosecond level. Then, the absolute deviation of the three types of errors is integrated through square sum and square root operation, and the coupling degree of the three types of errors is quantified through three-dimensional covariance calculation. Finally, the integrated collection error value is obtained. In the implementation process, the data collection frequency is set to 5 million data points per second, and the collection time length covers 30 minutes to ensure the representativeness of the data. This model provides comprehensive error evaluation basis for subsequent compensation parameter construction through multi-dimensional error integration and coupling analysis, effectively improving the integrity of error correction.

[0040] Preferably, the digital compensation parameter matrix construction model expression is: wherein, is the digital compensation parameter matrix, is the amplitude compensation weight, is the phase compensation weight, is the timing compensation weight, is the amplitude-phase cross compensation coefficient, is the phase-timing cross compensation coefficient, is the timing-amplitude cross compensation coefficient.

[0041] Specifically, the construction model of the digital compensation parameter matrix takes the integration of multi-dimensional compensation parameters as the goal, and sets the amplitude compensation weight value range as 0.3 to 0.5, the phase compensation weight as 0.2 to 0.4, the time sequence compensation weight as 0.1 to 0.3, and the cross compensation coefficient as 0.05 to 0.15 in implementation to balance the influence of main compensation and cross compensation. In application, the model first acquires three types of core parameters of operational amplifier dynamic coupling, harmonic distortion correction coefficient and differential nonlinear gradient value, then calculates the amplitude, phase and time sequence single-dimensional compensation parameters according to the set weights, integrates the coupling influence of different dimensional parameters through the cross compensation coefficient, and finally constructs a three-dimensional compensation parameter matrix. In the implementation process, the number of rows of the matrix corresponds to the full range of input code values (65536 rows), and the number of columns corresponds to three types of compensation dimensions (3 columns), which ensures the accurate matching of the parameter matrix and the working state of the DAC. The model provides comprehensive and accurate parameter support for real-time digital modulation by systematically integrating multi-source error compensation parameters, and realizes the synchronous correction of multi-dimensional errors.

[0042] Preferably, the S3 comprises the following steps: S31, inputting the operational amplifier dynamic coupling associated characteristics into the harmonic separation module of the wideband DAC harmonic distortion correction model, decomposing the composite signal into the fundamental component and each harmonic component through frequency domain transformation; S32, extracting the amplitude and phase of each harmonic component after decomposition, and recording the characteristic parameters of the harmonic component at different frequencies; S33, calculating the contribution proportion of each harmonic to the DAC output error based on the characteristic parameters, and eliminating the harmonic components with a contribution proportion lower than a set threshold; S34, mapping and associating the characteristic parameters corresponding to the retained high-contribution harmonic components with the error data to form a harmonic error contribution factor set.

[0043] Specifically, step S3 includes sub-steps S31-S34: S31 first inputs the op-amp dynamic coupling associated characteristics into the harmonic separation module of the harmonic distortion correction model of the wideband DAC, starts the frequency domain transformation algorithm, sets the transformation sample point number to 1024 points, ensures the resolution of signal decomposition, accurately decomposes the composite signal into the fundamental component and each harmonic component through the algorithm, and maintains the integrity of the amplitude and phase information of the signal during the decomposition process; S32 starts the amplitude and phase extraction process for each harmonic component after the decomposition, sets the amplitude extraction precision to the microvolt level and the phase extraction resolution to the milliradian level, records the amplitude, phase, frequency and other core characteristic parameters of the harmonic component at different frequencies one by one through a special detection module, and the parameter recording interval is consistent with the signal sampling period; S33 determines the contribution proportion of each harmonic to the DAC output error based on the extracted characteristic parameters, sets the contribution proportion calculation weight, wherein the amplitude proportion weight is 0.6 and the phase influence weight is 0.4, sets the contribution proportion threshold to 5%, and eliminates the harmonic components below the threshold to avoid invalid parameters occupying the operation resources; S34 maps and associates the characteristic parameters corresponding to the retained high-contribution harmonic components with the error original data, establishes a one-to-one correspondence between the parameters and the error, forms a structured harmonic error contribution factor set including the harmonic frequency, error contribution value, phase offset and other key information, and provides focused and accurate input data for the gradient operation of subsequent step S4. The whole sub-step implementation process ensures the efficiency and accuracy of harmonic separation and contribution factor screening through the collaborative linkage between modules.

[0044] Preferably, S4 includes the following sub-steps: S41, the harmonic error contribution factor is introduced into the gradient calculation unit of the differential nonlinear gradient correction model, the input code value change step is set, and all input code value ranges are traversed; S42, the differential nonlinear error value corresponding to each input code value is calculated, and the error gradient value is solved through the error difference value of adjacent code values; S43, the gradient value is subjected to sliding window filtering processing to eliminate the interference of random noise on the gradient change law; S44, the rising and falling intervals of the error distribution are judged through the positive and negative changes of the gradient value, and the error extreme point position where the gradient value is zero is located.

[0045] Specifically, step S4 includes sub-steps S41-S44: S41 first introduces the harmonic error contribution factor into the gradient calculation unit of the differential nonlinear gradient correction model, sets the input code value change step to one-tenth of the minimum code value unit, clearly defines the input code value traversal range as the full range, sequentially advances from the minimum input code value to the maximum input code value to ensure coverage of all working states; S42, for each input code value, calculates the corresponding differential nonlinear error value through the error detection unit, uses the adjacent code value error difference algorithm to perform difference operation on the current code value error and the previous code value error to obtain the error gradient value corresponding to each code value, and maintains the calculation precision to 0.01 units during the operation; S43 starts sliding window filtering processing on the calculated gradient value, sets the window length to 50 code value units, and uses the moving average algorithm to smooth the gradient values in the window to effectively eliminate the interference of random noise on the gradient change law, and the fluctuation amplitude of the filtered gradient value is controlled within the set range; S44 determines the error distribution trend through the gradient value analysis module, marks the rising and falling interval demarcation points of the error distribution when the gradient value changes from positive to negative or from negative to positive, and determines the error extreme point position when the gradient value tends to zero, and records the input code value and error amplitude corresponding to the extreme point, which provides key error distribution characteristic information for subsequent compensation parameter construction, and each sub-step is closely connected to ensure the accuracy of gradient operation and extreme point positioning.

[0046] Preferably, S5 includes the following sub-steps: S51, collect the coupling coefficient in the dynamic correlation feature, the harmonic amplitude and phase parameters in the error contribution factor, and the gradient extreme value data in the gradient change law, and establish an original set of compensation parameters; S52, according to the parameter constraint conditions of the digital compensation model, screen and standardize the parameters in the original set, and eliminate invalid parameters; S53, according to the functional requirements of amplitude compensation, phase calibration and time sequence adjustment, classify and distribute the standardized parameters to the corresponding compensation dimensions; S54, integrate the compensation parameters in each dimension through matrix operation to generate a dimension-matched digital compensation parameter matrix.

[0047] Specifically, step S5 includes sub-steps S51-S54: S51 first starts a parameter collection process, comprehensively collects the coupling coefficient in the dynamic correlation feature, the harmonic amplitude and phase parameters in the error contribution factor, and the gradient extreme data in the gradient variation law, sets the data collection accuracy consistent with the previous collection accuracy, ensures the data integrity, integrates all the collected data to form a compensation parameter original set, and stores the set according to the data type; S52 filters the parameters in the original set according to the parameter constraint conditions of the digital compensation model, eliminates abnormal parameters beyond the reasonable value range, and then starts a standardization process, which uniformly maps parameters of different dimensions and different orders of magnitude to the value interval of 0 to 1, and keeps the relative relationship between the parameters unchanged in the standardization process; S53 classifies and distributes the standardized parameters to the corresponding compensation dimensions according to the functional requirements of amplitude compensation, phase calibration and timing adjustment, sets a dedicated parameter storage area for each dimension to ensure the convenience of parameter calling, and records the weight proportion of each parameter in the corresponding dimension; S54 integrates the compensation parameters of each dimension through matrix operation, sets the number of matrix rows as the full range of input code values, and the number of columns as three compensation dimensions, adopts a weighted summation algorithm in the operation process, and generates a dimension-matched digital compensation parameter matrix combined with the weight proportion of each parameter, which includes specific compensation parameters corresponding to each input code value, and provides a direct basis for digital modulation of step S6. The implementation of each sub-step ensures the comprehensiveness, accuracy and adaptability of the compensation parameter matrix.

[0048] The operational amplifier misadjustment dynamic coupling prediction model is a technical model specially analyzing the dynamic correlation between the operational amplifier misadjustment voltage and the DAC output error, focusing on the coupling relationship between the operational amplifier misadjustment and the amplitude deviation, phase shift and timing jitter under different working conditions, realizing correlation feature extraction through multi-dimensional parameter integration, and solving the problem that the traditional static model cannot capture the dynamic correlation of errors. The implementation process of the model is as follows: taking the multi-dimensional error original data set collected by the quantum sensing DAC calibration analysis platform as input, setting the coupling relationship analysis threshold as amplitude deviation 5 microvolts and phase shift 0.1 milliradians, using a 10 millisecond fixed time window length to segment and analyze the data, through a 3-layer convolution feature extraction network and a 2-layer fully connected analysis network, calculating the Pearson correlation coefficient of the operational amplifier misadjustment voltage and various output errors, screening out the strong correlation combination with an absolute value of the correlation coefficient higher than 0.8, and then through linear fitting and nonlinear regression algorithm, extracting the quadratic function variation law of the operational amplifier misadjustment with time, input code value and load condition and the mapping relationship with the error, forming a dynamic correlation feature set including 20 core feature parameters, and simultaneously adaptively adjusting the convolution kernel size and the fully connected layer weight to adapt to complex working conditions. Precisely capturing the dynamic coupling effect of the operational amplifier misadjustment, breaking the limitation of static calibration, providing a targeted analysis object for subsequent harmonic separation, avoiding the analysis deviation caused by the superposition of multiple error sources, and ensuring the pertinence and accuracy of error analysis. The model fills the gap of the existing technology in ignoring the dynamic coupling effect of the operational amplifier misadjustment, improves the error analysis from a single dimension to a multi-dimensional dynamic correlation level, makes the error source positioning more accurate, provides high-quality pre-stage data support for the entire error correction process, helps to improve the overall effect and working condition adaptability of the DAC output error correction, and meets the stringent requirements of high-precision electronic systems for error analysis.

[0049] The wideband DAC harmonic distortion correction model is a technical model for separating the fundamental wave and harmonic components in the DAC output signal, screening key error sources, and specifically addressing the output error problem caused by harmonic distortion accumulation in the wideband scenario. The model achieves accurate quantification of harmonic error through frequency domain analysis. The implementation process of the model is as follows: taking the dynamic coupling correlation characteristics of operational amplifier misadjustment as input, setting the frequency analysis range to 1 kHz to 100 MHz, which completely matches the DAC working bandwidth, setting the upper limit of harmonic order analysis to 15 orders to cover the main harmonic components, decomposing the composite signal into fundamental wave and each harmonic component through fast Fourier transform algorithm, setting the amplitude extraction accuracy to microvolt level and the phase extraction resolution to 0.1 milliradian, collecting the peak amplitude, initial phase and center frequency parameters of each harmonic one by one, calculating the amplitude ratio and phase difference of each harmonic and the fundamental wave, setting the weighted calculation rule according to the amplitude ratio of 0.6 and the phase influence degree of 0.4, setting the error contribution factor screening threshold of 0.05, eliminating the harmonic components with weighted score below the threshold, retaining 3 to 8 high contribution harmonic parameters, forming a structured error contribution factor set, and eliminating false harmonic signals through 10 Hz frequency domain resolution adjustment and wavelet threshold denoising algorithm. Accurately separate the harmonic components in the wideband range, quantify the contribution degree of each harmonic to the output error, screen out the key error sources, provide focused and pure error data for subsequent gradient operation, and avoid invalid harmonic component interference in the subsequent processing flow. The model breaks through the limitations of traditional harmonic correction models in the wideband scenario, such as insufficient resolution and low screening accuracy, realizes accurate separation of harmonic components and efficient screening of high-contribution error sources, ensures that error correction can specifically address the main harmonic distortion problem, and improves the spectral purity of the DAC output signal in the wideband scenario. It provides key technical support for high-precision communication, quantum sensing and other fields of signal conversion.

[0050] The differential nonlinear gradient correction model is a technical model for analyzing the gradient variation law of harmonic error contribution factor and locating the error extreme point. The error distribution characteristics are deeply mined by accurately calculating the error gradient, and the core basis is provided for the compensation parameter construction. The implementation process of the model is as follows: taking the harmonic error contribution factor as the input, setting the input code value change step as 0.1 LSB, covering the full input code value range from minimum to maximum, taking the input code value as the independent variable and the comprehensive error value as the dependent variable, the error gradient value is calculated point by point through the adjacent code value error difference algorithm, the continuous gradient change curve is constructed, the gradient change rate threshold of 0.02 is set, the gradient curve slope change is analyzed by traversing the full input code value range, the moving average filtering is carried out by using the sliding window with a length of 50 code value units, the random noise interference with amplitude less than 0.01 is eliminated, the extreme point judgment condition is set as the absolute value of the gradient value of the continuous 5 adjacent code values less than 0.005 and the gradient sign of the front and back opposite, and finally the input code value position and error amplitude corresponding to the extreme point are output. The gradient variation law of the differential nonlinear error is accurately captured, the rising section, the falling section and the extreme point position of the error distribution are determined, the key feature information of the error distribution is provided for the construction of the digital compensation parameter matrix, the compensation strategy can be targeted to cope with the error extreme value area, and the blindness of the compensation is avoided. The model solves the problem that the prior art cannot accurately locate the error extreme point, realizes the quantitative analysis of the error distribution characteristics through the gradient operation, makes the compensation parameter construction can accurately adapt based on the dynamic distribution of the error, greatly improves the targeting and efficiency of the error correction, lays a core technical foundation for realizing the real-time dynamic correction of the DAC output error, and helps the high-precision electronic system to obtain more stable signal output performance.

[0051] The quantum sensing DAC calibration analysis platform is an integrated device for collecting multi-dimensional error data of DAC output and providing high-fidelity raw data support. It combines quantum sensing technology and multi-channel synchronous acquisition capability to achieve high-precision and full-scene acquisition of error data. The implementation process of the platform is as follows: after starting, set the sampling frequency to 5 million data points per second, collect for 30 minutes, cover the full working range of DAC input voltage 0 to full scale and frequency 1 kHz to 100 MHz, through 8-channel synchronous acquisition module, for different working conditions of input code value 0000H to FFFFH and load resistance 10Ω to 1kΩ, collect 2000 groups of sample data for each state, focus on collecting three core error parameters of amplitude deviation, phase offset and timing jitter, among which the amplitude deviation collection accuracy is controlled within ±5 microvolts, the phase offset resolution is 0.1 milliradian, and the timing jitter collection minimum interval is 10 nanoseconds. Relying on the built-in electromagnetic shielding structure and adaptive noise suppression algorithm, filter external electromagnetic interference of 20MHz to 1GHz and environmental noise with amplitude below 2 microvolts, classify and integrate the collected data according to time stamp, input code value and load condition, form a multi-dimensional error raw data set including 1.2 million effective records. Provide comprehensive, accurate and high-fidelity basic data for subsequent error analysis and correction process, ensure that the operational amplifier misadjustment dynamic coupling prediction and harmonic distortion correction model can be based on the real error distribution for analysis, and avoid the correction deviation caused by data distortion. The model breaks through the limitations of traditional calibration equipment with single collection dimension, insufficient precision and weak anti-interference ability. Through the high sensitivity characteristics of quantum sensing technology and multi-channel synchronous acquisition design, it realizes full-scene coverage and high-precision capture of error data, provides reliable data guarantee for multi-model collaborative correction, makes the entire error correction process have solid raw data support, and is the core basic equipment to improve the DAC output error correction effect.

[0052] As Figure 2As shown, a digital compensation-based DAC output error correction method is implemented by different units, including: a quantum sensing multi-dimensional error acquisition unit, configured to acquire amplitude deviation, phase offset and timing jitter data of a DAC output signal and generate error raw data sets, and connected to an operational amplifier (op-amp) misadjustment dynamic coupling feature analysis unit; the op-amp misadjustment dynamic coupling feature analysis unit is configured to call an op-amp misadjustment dynamic coupling prediction model to analyze the coupling relationship in the error raw data set, extract dynamic correlation features, and connect the output end to a wideband DAC harmonic error separation unit; the wideband DAC harmonic error separation unit is configured to separate harmonic components in the coupling features based on a wideband DAC harmonic distortion correction model, and filter error contribution factors, and connect the output end to a differential non-linear gradient operation unit; the differential non-linear gradient operation unit is configured to calculate the gradient change rule of the error contribution factor by a differential non-linear gradient correction model, and locate the extreme point position, and connect the output end to a digital compensation parameter matrix construction unit; the digital compensation parameter matrix construction unit is configured to combine the dynamic correlation features, the error contribution factors and the gradient change rule to construct a compensation parameter matrix including amplitude compensation coefficients, phase calibration coefficients and timing adjustment parameters, and connect the output end to a DAC output real-time modulation unit; the DAC output real-time modulation unit is configured to dynamically adjust the code value of the DAC output signal according to the digital compensation parameter matrix, and correct the error in response to the change of the extreme point position, and the input end is connected to the output end of the digital compensation parameter matrix construction unit, and the output end is directly connected to the DAC output link.

[0053] A digital compensation-based DAC output error correction method integrates op-amp misadjustment dynamic coupling prediction, wideband harmonic distortion correction, and differential non-linear gradient correction models to form a full-process processing link covering error correlation analysis, harmonic separation, and gradient operation, changes the limitations of single error source calibration in the prior art, can fully capture the dynamic correlation between op-amp misadjustment and harmonic distortion, differential non-linearity, realize synchronous response and collaborative correction of multi-source errors, and completely solve the problem of poor correction effect caused by multi-error superposition under complex working conditions. At the same time, relying on quantum sensing technology to build a dedicated calibration analysis platform, greatly improves the dimension and accuracy of error data acquisition, provides high-reliability data support for subsequent compensation parameter construction, and overcomes the defects of one-sided data acquisition and difficulty in reflecting real error distribution in traditional methods.

[0054] The application breaks the limitation of poor adaptability of traditional static compensation parameters by analyzing the error gradient variation law and positioning the extreme value point position, and constructing a dynamically updated compensation parameter matrix, so that the digital modulation process can respond to the dynamic change of error distribution in real time. This real-time adjustment mechanism based on the dynamic characteristics of the error ensures that the compensation strategy is always accurately matched with the error change trend, significantly improving the correction stability and adaptability under different working conditions. At the same time, the combination of multi-model cooperation and real-time modulation realizes the whole process closed-loop processing from error collection, analysis to correction, solves the problem that the prior art lacks integrated mechanism and cannot realize synchronous correction of multi-dimensional error, greatly improves the amplitude, phase and timing accuracy of the DAC output signal, and fully meets the stringent requirements of high-precision electronic systems.

[0055] In the description of the application, it should be noted that, unless otherwise specified and limited, the terms "set", "install", "connect", "connect", "fix" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the communication inside two elements. For those skilled in the art, the specific meaning of the above terms in the application can be understood according to the specific circumstances.

[0056] Although the embodiments of the application have been shown and described, it can be understood by those skilled in the art that various equivalent changes, modifications, replacements and variations of the embodiments can be made without departing from the principles and spirits of the application, and the scope of the application is defined by the appended claims and their equivalent scope.

Claims

1. A method for correcting DAC output error based on digital compensation, characterized in that, Includes the following steps: S1. Collect amplitude deviation, phase shift, and timing jitter data of the DAC output signal using a quantum sensing DAC calibration and analysis platform to establish a multi-dimensional original error dataset. S2. Use the op-amp offset dynamic coupling prediction model to analyze the coupling relationship of the original error dataset and extract the dynamic correlation characteristics between the op-amp offset voltage and the output error. S3. Based on a broadband DAC harmonic distortion correction model, separate the harmonic components of the coupling correlation characteristics and select the error contribution factors corresponding to each harmonic. S4. Use a differential nonlinear gradient correction model to perform gradient calculations on the harmonic error contribution factors to determine the gradient change law and extreme point location of the error distribution. S5. Combine the dynamic correlation characteristics, error contribution factors, and gradient change law to construct a digital compensation parameter matrix, which includes amplitude compensation coefficients, phase calibration coefficients, and timing adjustment parameters. S6. Perform real-time digital modulation of the DAC output signal based on the digital compensation parameter matrix, and correct the error by dynamically adjusting the output code value. The digital modulation process synchronously responds to changes in the extreme point location of the error distribution.

2. The DAC output error correction method based on digital compensation according to claim 1, characterized in that, The expression for the operational amplifier offset dynamic coupling prediction model is as follows: This refers to the dynamic coupling quantity of the operational amplifier offset. The coupling coefficient is... This is the initial op-amp offset voltage. Clock frequency influence factor, For system clock frequency, The output voltage deviation sensitivity coefficient. For DAC output voltage deviation, To add weighting coefficients, The offset voltage of the i-th operational amplifier. Let i be the phase deviation influence coefficient of the i-th path. For the first Road signal phase deviation.

3. The DAC output error correction method based on digital compensation according to claim 1, characterized in that, The expression for the broadband DAC harmonic distortion correction model is as follows: middle, This is the harmonic distortion correction factor. The fundamental amplitude, The fundamental angular frequency, For time variables, The initial phase of the fundamental wave. The harmonic order is... The amplitude correction factor for the m-th harmonic is... Let m be the amplitude of the m-th harmonic. The initial phase of the m-th harmonic. The integral correction factor for the m-th harmonic is given. It is the integral variable.

4. The DAC output error correction method based on digital compensation according to claim 1, characterized in that, The expression for the differential nonlinear gradient correction model is: ,in, The differential nonlinear gradient value, For gradient operators, These are the original differential nonlinear coefficients. This represents the original differential nonlinear error value. The first derivative weighting coefficients are... For DAC input code value variables, The weighting coefficient is the squared second derivative.

5. The DAC output error correction method based on digital compensation according to claim 1, characterized in that, The error acquisition model expression of the quantum sensing DAC calibration and analysis platform is as follows: ,in, To comprehensively collect error values, This is the data acquisition accuracy coefficient. For amplitude deviation, For phase shift, For timing jitter, The covariance weighting coefficients are... The three-dimensional covariance is the sum of amplitude deviation, phase shift, and timing jitter.

6. The DAC output error correction method based on digital compensation according to claim 1, characterized in that, The model expression for constructing the digital compensation parameter matrix is ​​as follows: in, For digital compensation parameter matrix, For amplitude compensation weight, For phase compensation weights, For time-series compensation weights, The amplitude-phase crossover compensation coefficient, For phase-timing crossover compensation coefficients, This is the timing-amplitude cross-compensation coefficient.

7. The DAC output error correction method based on digital compensation according to claim 1, characterized in that, S3 includes the following sub-steps: S31, inputting the operational amplifier offset dynamic coupling correlation characteristics into the harmonic separation module of the broadband DAC harmonic distortion correction model, and decomposing the composite signal into the fundamental component and each harmonic component through frequency domain transformation; S32, extracting the amplitude and phase of each harmonic component after decomposition, and recording the characteristic parameters of the harmonic components at different frequencies; S33, calculating the contribution ratio of each harmonic to the DAC output error based on the characteristic parameters, and removing harmonic components with a contribution ratio lower than a set threshold; S34, mapping and associating the characteristic parameters corresponding to the retained high-contribution harmonic components with the error data to form a set of harmonic error contribution factors.

8. The DAC output error correction method based on digital compensation according to claim 1, characterized in that, The S4 includes the following sub-steps: S41, importing the harmonic error contribution factor into the gradient calculation unit of the differential nonlinear gradient correction model, setting the input code value change step size, and traversing the entire range of input code values. S42, calculate the differential nonlinear error value corresponding to each input code value, and solve the error gradient value by the error difference between adjacent code values; S43, perform sliding window filtering on the gradient value to eliminate the interference of random noise on the gradient change law; S44, determine the rising and falling intervals of the error distribution by the positive and negative changes of the gradient value, and locate the error extreme point where the gradient value is zero.

9. The DAC output error correction method based on digital compensation according to claim 1, characterized in that, S5 includes the following sub-steps: S51, collecting coupling coefficients from dynamic correlation features, harmonic amplitude and phase parameters from error contribution factors, and gradient extreme value data from gradient change patterns to establish an original set of compensation parameters; S52, filtering and standardizing the parameters in the original set according to the parameter constraints of the digital compensation model, and removing invalid parameters; S53, classifying and assigning the standardized parameters to the corresponding compensation dimensions according to the functional requirements of amplitude compensation, phase calibration, and timing adjustment; S54, integrating the compensation parameters of each dimension through matrix operations to generate a dimension-matched digital compensation parameter matrix.

10. A DAC output error correction method based on digital compensation according to any one of claims 1-9, characterized in that, This method is implemented through different units, including: a quantum sensing multi-dimensional error acquisition unit, used to acquire amplitude deviation, phase shift, and timing jitter data of the DAC output signal and generate a raw error dataset, whose output is connected to an op-amp offset dynamic coupling feature analysis unit; an op-amp offset dynamic coupling feature analysis unit, used to call the op-amp offset dynamic coupling prediction model to analyze the coupling relationship in the raw error dataset and extract dynamic correlation features, whose output is connected to a broadband DAC harmonic error separation unit; a broadband DAC harmonic error separation unit, used to separate harmonic components in the coupling features based on a broadband DAC harmonic distortion correction model and screen error contribution factors, whose output is connected to a differential nonlinear gradient operation unit; and a differential nonlinear gradient operation unit. The calculation unit is used to calculate the gradient change law of the error contribution factor through the differential nonlinear gradient correction model and locate the extreme point. Its output is connected to the digital compensation parameter matrix construction unit. The digital compensation parameter matrix construction unit is used to construct a compensation parameter matrix including amplitude compensation coefficient, phase calibration coefficient and timing adjustment parameter by combining dynamic correlation characteristics, error contribution factor and gradient change law. Its output is connected to the DAC output real-time modulation unit. The DAC output real-time modulation unit is used to dynamically adjust the code value of the DAC output signal according to the digital compensation parameter matrix and to correct the error in response to the change of extreme point position. Its input is connected to the output of the digital compensation parameter matrix construction unit and its output is directly connected to the DAC output link.

Images