Real-time jitter detection method and system for eye diagram of high-speed oscilloscope

Abstract

The invention relates to the technical field of jitter analysis, in particular to a high-speed oscilloscope eye diagram real-time jitter detection method and system, and the method comprises the following steps: carrying out high-speed collection waveform zero-cross detection to obtain a timestamp, calculating offset grouping to obtain a cross-period edge time sequence, merging a jump cluster according to a sequence hierarchy cluster, calculating offset, obtaining a drift trend vector, and calculating the jitter of the drift trend vector; and based on the trend vector distribution compensation weight, weighting and correcting the time coordinate, calling a corrected coordinate detection period boundary to calculate a difference value exceeding a threshold value, triggering correction to obtain a period correction rear edge coordinate, and extracting a jitter frequency amplitude according to corrected coordinate Fourier analysis to generate an eye pattern jitter detection result. Phase correlation is established based on a cross-period sequence, an analyzable offset mode is formed through a clustering identification distribution structure, accumulated deviation is suppressed through a drift trend and weight correction, an abnormal period is calibrated through period dynamic discrimination, and the relation between a jitter structure and amplitude is separated through frequency analysis, so that the overall stability and integrity are enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of jitter analysis, in particular to a high-speed oscilloscope eye diagram real-time jitter detection method and system. BACKGROUND

[0002] The technical field of jitter analysis involves measuring and analyzing the time offset of high-speed digital signals generated during transmission and measurement. The core tasks include separating the components of clock jitter and data jitter, quantifying the statistical characteristics of different jitter components, extracting the timing deviations of rising edges, falling edges, and zero-crossing points in the eye diagram, and using high-speed sampling instruments to continuously track the edge positions to form a systematic understanding of the overall jitter behavior. This field relies on the sampling architecture, trigger mechanism, and eye diagram reconstruction process of high-speed oscilloscopes to identify and analyze jitter sources.

[0003] Among them, the traditional high-speed oscilloscope eye diagram real-time jitter detection method refers to obtaining the data eye diagram based on the high-speed oscilloscope, marking the time of continuous sampling points, determining the zero-crossing time of each signal jump as the edge time position by interpolation, obtaining the offset of adjacent bit period edges by point-by-point comparison, and dividing periodic jitter and random jitter by performing histogram statistics on a large number of edge time points. It generally uses equal interval time collection under fixed sampling rate, edge capture based on zero-crossing level judgment, linear interpolation to obtain accurate edge time, and accumulates edge time difference in sequence to form jitter distribution to complete the detection and characterization of jitter behavior in the eye diagram.

[0004] Traditional eye diagram jitter detection relies on equal interval sampling to obtain edge time points, and the random scattering of zero-crossing position is affected by noise disturbance and sampling error. The edge time determined by interpolation has a sequence accumulation offset, the lack of structural correlation between cross-period time points makes the overall drift between different periods difficult to identify, the edge difference method only reflects the local jump difference and is difficult to describe long-term trends, the period length is judged by single time interval and is easily affected by occasional abnormalities, making the period deviation continuously amplified in the measurement sequence, and the jitter characteristics rely on simple time distribution statistics to reveal the internal change pattern, resulting in a decrease in edge positioning stability, insufficient period judgment reliability, and incomplete frequency domain characteristics. SUMMARY

[0005] In order to solve the technical problems that the traditional eye diagram jitter detection relies on equal interval sampling to obtain edge time points, the zero-crossing position is affected by noise disturbance and sampling error, the edge time determined by interpolation has a cumulative offset with the sequence, the overall drift between different periods cannot be identified due to the lack of structural correlation of the cross-period time points, the edge difference method only reflects the local jump difference and is difficult to describe the long-term change trend, the period length is easily affected by occasional abnormalities, so that the period deviation is continuously amplified in the measurement sequence, the jitter characteristics are difficult to reveal the internal change mode by simple time distribution statistics, the edge positioning stability is reduced, the period judgment reliability is insufficient, and the frequency domain characteristics are incomplete, etc., the embodiment of the present application provides a high-speed oscilloscope eye diagram real-time jitter detection method and system.

[0006] In order to achieve the above purpose, the present application provides a high-speed oscilloscope eye diagram real-time jitter detection method and system, the high-speed oscilloscope eye diagram real-time jitter detection method comprises the following steps: S1: acquiring the digitized waveform of the measured signal by a high-speed analog-to-digital converter, detecting the time stamp corresponding to the zero-crossing point position based on a zero-crossing detection algorithm, calculating the time offset, grouping according to the bit period, and obtaining the cross-period edge time sequence; S2: according to the cross-period edge time sequence, clustering analysis is performed on the zero-crossing points of adjacent bit periods at the same phase, the edge time points are merged into a jump point cluster, the deviation is calculated, and a cluster drift trend vector is obtained; S3: based on the cluster drift trend vector, the time points in the jump point cluster are assigned an error compensation weight, the time coordinate correction amount is calculated by weighted calculation, and is superimposed to the corresponding time coordinate to obtain the drift compensation edge positioning coordinate; S4: calling the drift compensation edge positioning coordinate, monitoring the time interval of the first and last zero-crossing points of the bit period by a period boundary detection circuit, calculating the difference value with the standard period length, triggering the correction instruction when the correction threshold is exceeded, and obtaining the period correction edge coordinate; S5: according to the period correction edge coordinate, the Fourier transform algorithm is used to analyze the spectral characteristics of the edge position sequence, the jitter frequency component and amplitude are extracted, the jitter amplitude parameter is calculated, and the eye diagram jitter detection result is generated.

[0007] As a further scheme of the present application, the cross-period edge time sequence includes time stamp, bit period grouping and cross-period jump point, the jump point cluster includes cluster center, offset and dispersion, the drift compensation edge positioning coordinate includes correction amount, compensation weight and positioning accuracy, the period correction edge coordinate includes period length deviation, period consistency and correction deviation, and the eye diagram jitter detection result includes jitter frequency component, amplitude component and spectral characteristics.

[0008] As a further scheme of the present application, the specific steps of S1 are: S101: Obtain a digitized waveform through a high-speed analog-to-digital converter, perform comparison on amplitude signs of adjacent sampling points and locate sign change points based on a zero-crossing detection algorithm, perform linear proportional interpolation on the amplitude difference of two sampling points at the sign change points and determine the cross-time position to generate a zero-crossing timestamp sequence; S102: Call the zero-crossing timestamp sequence, perform difference processing on adjacent time position difference values and form time displacement correlation values, then perform proportional mapping on the correlation values according to the sampling time interval to obtain a time displacement amount set, and perform segmented aggregation according to the sampling time sequence to obtain a period time offset matrix; S103: Based on the period time offset matrix, perform sorting on the index positions within multiple segments, aggregate the items at the same index position according to the segment sequence, and obtain a cross-period edge time sequence.

[0009] As a further scheme of the present application, the specific steps of S2 are: S201: Based on the cross-period edge time sequence, perform retrieval on the same phase zero-crossing time values of adjacent bit periods, and perform difference operation on the time values and phase reference time values, then arrange the difference values according to the phase index sequence to generate a zero-crossing difference matrix; S202: Call the zero-crossing difference matrix, perform aggregation on the difference value vectors according to the difference value amplitude and phase index, extract the offset sequence from the aggregated vectors, and arrange the offset values according to the group order to obtain an offset distribution parameter set; S203: Call the offset distribution parameter set, perform sequence arrangement on the offset parameters according to the index order, perform difference operation on adjacent offset parameters to form a difference sequence, extract the direction change items according to the sequence position and accumulate them to obtain a cluster drift trend vector.

[0010] As a further scheme of the present application, the specific steps of S3 are: S301: Based on the cluster drift trend vector, retrieve the time coordinates within the jump point cluster and call the offset parameter items in the offset distribution parameter set, perform comparison on the offset parameter items and the error compensation threshold value, and aggregate the items greater than the threshold value to obtain a weighted coefficient matrix value; The error compensation threshold value is obtained by performing statistical quantity extraction on all offset parameter items in the offset distribution parameter set, and is a quantitative reference value obtained by performing linear combination on the mean value of all offset parameter items in the offset distribution parameter set and the standard deviation of all offset parameter items; S302: calling the weighting coefficient matrix value, performing proportional conversion on the time coordinates in the jump point cluster and the weighting coefficient, and aggregating the conversion results, calculating the offset based on the aggregated offset parameter, performing difference processing on the offset and the time coordinates, and obtaining the time coordinate correction value; S303: calling the time coordinate correction value, performing superposition processing on the time coordinates in the jump point cluster and the correction value, performing time sorting and maintaining index continuity, and obtaining the edge positioning coordinates after drift compensation.

[0011] As a further scheme of the present application, the specific steps of S4 are: S401: based on the edge positioning coordinates after drift compensation, detecting the positioning coordinate time sequence, retrieving the first and last zero-crossing point time positions of each bit period, calculating the first and last zero-crossing point time interval, and performing consistency judgment on the time interval sequence to generate the zero-crossing point interval value; S402: calling the zero-crossing point interval value, performing difference calculation on the zero-crossing point interval value according to the standard period length reference value, comparing the difference value with the correction threshold, and when the difference value exceeds the correction threshold, triggering the correction instruction and recording the trigger identifier, and establishing the period difference identifier value; S403: calling the period difference identifier value, performing time coordinate adjustment on the edge positioning coordinates after drift compensation according to the difference direction and difference amplitude recorded in the identifier value, and converging the adjusted edge time sequence to obtain the edge coordinates after period correction.

[0012] As a further scheme of the present application, the correction threshold is a fixed quantization threshold determined according to the statistical dispersion degree of the zero-crossing point interval value recorded in the original bit period stable segment, and the statistical dispersion degree is obtained by performing mean and deviation analysis on the zero-crossing point interval value in the original bit period; The standard period length reference value is a reference period value obtained according to the nominal period parameter of the rated bit period, and the nominal period parameter is obtained by converting the rated frequency of the clock reference signal.

[0013] As a further scheme of the present application, the specific steps of S5 are: S501: based on the edge coordinate sequence after period correction, performing frequency domain decomposition based on Fourier transform on the discrete amplitude in the sequence and performing comparison between the sample value and the frequency index, performing amplitude positioning and index aggregation after comparison and performing serialization processing, and generating a frequency domain amplitude array; S502: calling the frequency domain amplitude array, performing amplitude difference retrieval on the frequency sample value and the adjacent sample value in the array, and comparing the difference value with the amplitude difference threshold, performing aggregation encoding and sequence arrangement for the indexes greater than the amplitude difference threshold, and obtaining the jitter frequency index set; S503: Call the jitter frequency index set, execute amplitude conversion on the index frequency corresponding amplitude parameter and edge position sequence time interval parameter, and execute weighted superposition on the converted amplitude value, and regularize the amplitude sequence to obtain an eye diagram jitter detection result.

[0014] As a further scheme of the present application, the amplitude difference threshold is based on the statistics of all discrete amplitude sample values in the frequency domain amplitude array, and is obtained by calculating the interval span parameter of the amplitude distribution interval and the amplitude variation of the discrete amplitude sample value by a proportional coefficient.

[0015] In addition, the present application also provides a high-speed oscilloscope eye diagram real-time jitter detection system, which comprises: The original waveform sampling module acquires the digitized waveform of the measured signal through a high-speed analog-to-digital converter, detects the time stamp corresponding to the zero-crossing point based on a zero-crossing detection algorithm, calculates the time offset, groups the time points according to the bit period, and obtains the cross-period edge time sequence. The edge aggregation analysis module performs clustering analysis on the zero-crossing points of adjacent bit periods at the same phase according to the cross-period edge time sequence, merges the edge time points into a jump point cluster, calculates the deviation, and obtains a cluster drift trend vector. The drift compensation calculation module assigns an error compensation weight to the time points in the jump point cluster based on the cluster drift trend vector, calculates the time coordinate correction amount through weighted calculation, and adds it to the corresponding time coordinate to obtain the drift-compensated edge positioning coordinate. The period calibration monitoring module calls the drift-compensated edge positioning coordinate, monitors the time interval between the first and last zero-crossing points of the bit period through a period boundary detection circuit, calculates the difference from the standard period length, triggers a correction instruction when the difference exceeds a correction threshold, and obtains the period-corrected edge coordinate. The spectrum feature extraction module analyzes the spectrum characteristics of the edge position sequence based on the period-corrected edge coordinate using the Fourier transform algorithm, extracts the jitter frequency component and amplitude, calculates the jitter amplitude parameter, and generates an eye diagram jitter detection result.

[0016] Compared with the prior art, the present application has the following advantages and positive effects: In the present application, the phase correlation is established through the cross-period edge sequence, and the distribution structure of the time points is identified in a clustering manner, so that the cross-period offset forms a resolvable pattern. The systematic deviation gradually accumulated in the sequence is suppressed by extracting the drift trend and adjusting the correction amount of the time points with a difference weight. The influence of abnormal periods is eliminated locally by triggering period calibration through dynamic discrimination of the period time interval. The period structure and amplitude relationship of jitter variation are clearly separated by analyzing the frequency components in the corrected edge sequence, thereby improving the consistency of edge positioning and strengthening the overall stability and integrity of jitter expression. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description only constitute some embodiments of the present application, and for those skilled in the art, the other drawings can be obtained without creative effort based on the drawings.

[0018] Figure 1 The step flowchart of the present application is shown in the figure. Figure 2 The S1 refinement diagram of the present application is shown in the figure. Figure 3 The S2 refinement diagram of the present application is shown in the figure. Figure 4 The S3 refinement diagram of the present application is shown in the figure. Figure 5 The S4 refinement diagram of the present application is shown in the figure. Figure 6 The S5 refinement diagram of the present application is shown in the figure. Figure 7 The system module diagram of the present application is shown in the figure. DETAILED DESCRIPTION

[0019] The technical solutions in the present application will be described in detail below with reference to the drawings.

[0020] In the embodiments of the present application, the words such as "example", "for example" are used to represent as an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific manner. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two.

[0021] In the embodiments of the present application, "image" and "picture" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent. "Of", "corresponding" and "corresponding" can be used interchangeably at times. It should be pointed out that when the distinction is not emphasized, the meanings expressed are consistent.

[0022] In the embodiments of the present application, sometimes the subscript such as W1 is written in the form of non-subscript such as W1. When the distinction is not emphasized, the meanings expressed are consistent.

[0023] In order to make the technical problems, technical solutions and advantages of the present application more clear, the following will be described in detail with reference to the drawings and specific embodiments.

[0024] Referring to Figure 1 , the embodiment of the present application provides a high-speed oscilloscope eye diagram real-time jitter detection method, comprising the following steps: S1: acquiring the digital waveform of the measured signal by a high-speed analog-to-digital converter, detecting the zero-crossing point position based on a zero-crossing detection algorithm, calculating the time offset, grouping according to the bit period, and obtaining the cross-period edge time sequence; S2: according to the cross-period edge time sequence, clustering analysis is performed on the zero-crossing points of adjacent bit periods at the same phase, the edge time points are merged into a jump point cluster, the deviation is calculated, and a cluster drift trend vector is obtained; S3: based on the cluster drift trend vector, the time points in the jump point cluster are assigned an error compensation weight, the time coordinate correction amount is calculated by weighted calculation, and the corresponding time coordinate is superimposed to obtain the edge positioning coordinate after drift compensation; S4: calling the edge positioning coordinate after drift compensation, monitoring the time interval between the first and last zero-crossing points of the bit period through a period boundary detection circuit, calculating the difference value from the standard period length, and triggering a correction instruction when the correction threshold is exceeded to obtain the edge coordinate after period correction; S5: according to the edge coordinate after period correction, the Fourier transform algorithm is used to analyze the spectral characteristics of the edge position sequence, the jitter frequency component and amplitude are extracted, the jitter amplitude parameter is calculated, and the eye diagram jitter detection result is generated.

[0025] The cross-period edge time sequence includes time stamps, bit period grouping and cross-period jump points, the jump point cluster includes cluster centers, offsets and dispersion, the edge positioning coordinate after drift compensation includes correction amounts, compensation weights and positioning accuracy, the edge coordinate after period correction includes period length deviation, period consistency and correction deviation, and the eye diagram jitter detection result includes jitter frequency components, amplitude components and spectral characteristics.

[0026] Referring to Figure 2 , the specific steps of S1 are: S101: acquiring the digital waveform by a high-speed analog-to-digital converter, comparing the amplitude signs of adjacent sampling points based on a zero-crossing detection algorithm and positioning the sign change points, performing linear proportional interpolation on the sign change points according to the amplitude difference of two sampling points and determining the cross-time position, and generating a zero-crossing time stamp sequence; First, start the high-speed oscilloscope front-end acquisition module configured with a sampling rate, which locks the high-speed serial signal link to be measured , inputs the analog signal to the analog-to-digital converter with a bit width of , and the analog-to-digital converter continuously acquires the voltage amplitude with a fixed sampling time interval of , and quantizes the continuously acquired analog voltage to contain Discrete voltage sequence at each sampling point Then, the zero-crossing detection logic module in the central processing unit is called to initialize the traversal pointer. for Read data from two adjacent sampling points sequentially from the memory. and Extract respectively and The sign bit is used to perform an XOR logical operation. If the result is... The system determines that a sign change point exists at this location, meaning the signal waveform crosses the zero-level axis at this point, and records the index. Corresponding physical sampling time as well as Corresponding physical sampling time ,in Set as The corresponding amplitude value Measured as ,and Set as The corresponding amplitude value Measured as Next, based on the principle of linear proportional interpolation, a cross-time calculation model is constructed. This model calls the floating-point arithmetic unit to calculate the sum of the absolute values ​​of the amplitudes of the two sampling points. The calculation result is Calculate the proportion of the absolute value of the amplitude of the previous sampling point in the total amplitude difference. Substituting the numerical values, we obtain Simultaneously calculate the time step. The proportion coefficient With time step Multiply to obtain the zero-crossing point relative to the zero-crossing point. time offset Finally, this offset is added to the base sampling time. The above calculates the precise zero-cross timestamp. Following this process, the subsequent sampling points are traversed to identify the positions of sign changes in the sequence, and the interpolation calculations are performed sequentially. The calculated results are then... The values ​​are stored sequentially in a double-rate synchronous dynamic random access memory, forming a structure with a length of [missing information]. Zero-crossing timestamp sequence .

[0027] S102: Call the zero-crossing timestamp sequence, perform differential processing on the difference between adjacent time positions to form a time displacement correlation value, then perform proportional mapping on the correlation value according to the sampling time interval to obtain the time displacement set, and perform segmented aggregation according to the sampling time order to obtain the periodic time offset matrix; Calling the zero-crossing timestamp sequence built in memory , reading the first timestamp in the sequence , setting the standard unit interval of the signal under test as the reference phase , for , starting the differential processing engine, reading two adjacent timestamps in the sequence in turn and , performing subtraction to obtain the difference value of adjacent time positions , for example, reading and in a certain operation , the difference value is calculated , then the difference value is preprocessed by modulo operation relative to the standard unit interval , but in order to retain the multi-cycle jitter characteristics, the difference value is compared with the closest integer multiple , the calculation formula is , where is the positive integer that makes the smallest, in this example , substituting the numerical value into the calculation obtains the time displacement correlation value , which is a positive value indicating that the signal edge lags behind the ideal clock, and this difference and correlation operation is performed on adjacent points in the sequence in turn to form a set of time displacement correlation values, then the correlation values are proportionally mapped according to the sampling time interval, a two-dimensional mapping relationship table is established, taking the global time index of the sampling point in the original waveform as the horizontal coordinate and the calculated time displacement correlation value as the vertical coordinate, obtaining the time displacement set, then performing segmented aggregation in the order of sampling time, setting the segmented window length corresponding to the horizontal time base span of the oscilloscope screen, for example, setting each segment to contain , cutting the time displacement set into subsets in order, each subset representing the jitter distribution within a signal period, constructing a period time offset matrix , each row of the matrix corresponding to a segmented subset, each column of the matrix corresponding to the th edge position in the subset, filling the calculated value into the corresponding element of the matrix, the specific differential processing and mapping data are shown in Table 1. Table 1: Signal edge time displacement mapping data table.

[0028]

[0029]

[0030] ​As shown in Table 1, by comparing the difference of the original timestamp with the ideal multiple, the picosecond jitter component of each edge relative to the ideal clock was accurately extracted, and the component was then loaded into the periodic time offset matrix.

[0031] S103: Based on the periodic time offset matrix, sort multiple entries according to their intra-segment index positions, aggregate entries with the same index position according to the segment order, and obtain cross-period edge time series; Based on the constructed periodic time offset matrix The matrix column sorting and aggregation controller is activated. For each row of data vectors in the matrix, the intra-segment index position sorting is performed. Due to the existence of bit loss or bit error in the actual signal, the number of detected edges in each segment is inconsistent. First, the rows of the matrix are traversed to identify the valid non-empty data elements in each row, and intra-segment index values ​​are assigned according to the relative time order of the elements in the row. For example, for the first The row vector of the row, containing the first row. Each valid edge jitter value is marked as , No. Each marker is This process continues until the end of the line, after which a vertical scan is performed to locate the index position within the specific segment. For example, selecting From the first of the matrix Arriving at the Extract line by line within the row. The value of an entry, if a row is in position If data is missing, the row will be automatically skipped, and the extracted values ​​will be used. Serial aggregation is performed according to the row number order, i.e., segment order, of the matrix to construct a cross-cycle edge jitter subsequence for that specific edge position. Then, the pointer is... Incrementing, repeating the extraction and aggregation process described above, until... Once the maximum number of columns in the matrix is ​​reached, the generated subsequences are sorted by index. The data can be concatenated sequentially or output as independent data channels to ultimately obtain a complete cross-cycle edge time series. This sequence not only contains simple time information, but also implies the periodic jitter trend of the signal during long-term transmission through its arrangement structure. This result shows that discrete single-trigger sampling data has been successfully reconstructed into a statistically significant continuous eye diagram data source, which can directly support subsequent eye diagram rendering and jitter histogram analysis.

[0032] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the cross-cycle edge time series, perform a retrieval operation on the zero-crossing time value of the same phase in adjacent bit periods, and perform a difference operation between the time value and the phase reference time value. Then, arrange the difference items according to the phase index sequence to generate a zero-crossing difference matrix. Call the cross-cycle edge time series generated in memory Initialize the phase reference clock generator, and set the reference clock frequency to be strictly synchronized with the baud rate of the signal under test. The corresponding standard unit interval Fixed as Start sequence retrieval pointer point to Read the first zero-crossing time value from the address of the first address. For example, reading According to the preset code length (Corresponding to PRBS7 code pattern) Calculate the relative phase position of the current time point within the code pattern period using the modulo operation formula. Determine the phase index and substitute it into the numerical calculation to obtain... Take the integer part to lock the phase index Then, based on this index, the phase reference time lookup table is called to obtain the ideal reference time value corresponding to that phase. Assuming the current cumulative number of cycles is ,but Calculation A search operation is performed on the zero-crossing time value of the same phase in adjacent bit periods, scanning backwards in the sequence to lock the next marker as the index. zero crossing time point For example, the time value of the corresponding position in the next cycle is retrieved. Perform difference calculations to determine the deviation between the time value and the phase reference time value. The relative jitter calculation logic is used here, that is, the calculation... Relative to the deviation of the local recovery clock, if the local recovery clock is aligned at this point... The difference Continue performing this operation on the data points in the sequence to obtain massive amounts of difference data, and then construct a zero-crossing difference matrix. The row dimensions of this matrix correspond to the phase index. to The column dimension corresponds to the number of observation periods, and the calculated difference term According to its corresponding phase index Fill in the corresponding row of the matrix, for example, the above Fill in the first Line number The column will be used to detect subsequent events. Fill in the first row column, complete the classification loading of data, and finally generate the zero-crossing difference matrix The instantaneous jitter amplitude of 127 different phase points in the continuous monitoring process is recorded.

[0033] S202: Call the zero-crossing difference matrix, perform aggregation on the difference vector according to the difference amplitude and phase index, extract the offset sequence from the aggregated vector, and then sort the offsets according to the group order to obtain the offset distribution parameter set; Call the zero-crossing difference matrix constructed in the buffer area , activate the matrix aggregation processor, perform aggregation on the specific phase index vector represented by each row of the matrix, and set the difference amplitude aggregation interval step as , traverse the data vector of the first row of the matrix , map the difference amplitude to the corresponding quantization interval, count the number of falling points in each interval, for example, the value falls into the interval , and the interval count value is incremented by 1. After completing the statistics of the data in this row, the center value of the high probability distribution interval is extracted from the aggregation result as the characteristic offset of this phase, for example, the statistical peak value of the third row is located in the interval , and the characteristic offset (the midpoint of the interval) is extracted. This operation is performed on the rows of the matrix in turn to generate an original offset sequence containing elements , and then the offsets are sorted according to the group order. According to the data channel grouping standard in the high-speed serial bus protocol, the is divided into logical groups, each containing phase indexes (truncated or padded at the end), and the root mean square value of the offsets in the first group (index ) is calculated . Assuming that the set of offsets in the group is , the root mean square formula is calculated to obtain , and the maximum positive offset and the maximum negative offset in the group are extracted. The statistical characteristic parameters are packaged, and finally the offset distribution parameter set is obtained . This parameter set quantifies the jitter dispersion degree and central tendency of different phase groups in detail. The specific grouping and sorting data are shown in Table 2.

[0034] Table 2: Phase offset distribution parameter sorting table.

[0035]

[0036] As shown in Table 2, the massive point-by-point difference value data is compressed and regularized into distribution parameters with macroscopic statistical significance, clearly showing the signal quality differences of different logical channel regions.

[0037] S203: Call the offset distribution parameter set, perform the serialization arrangement action on the offset parameters in index order, perform the difference operation on adjacent offset parameters to form a difference sequence, extract the direction change item according to the sequence position and accumulate, and obtain the cluster drift trend vector; Call the offset distribution parameter set in the memory , start the trend analysis engine, first perform the serialization arrangement action on the offset parameters in index order, extract the aggregated characteristic value groups in Table 2 to form a one-dimensional feature vector , then perform the difference operation on adjacent offset parameters in the vector, the calculation formula is , the first-order difference , the second-order difference , the third-order difference , form a difference sequence , extract the direction change item according to the sequence position, set the positive drift threshold , the negative drift threshold , scan the elements in one by one, determine does not exceed the threshold, and mark it as a stable state , determine is less than , mark it as negative drift , determine is greater than , mark it as positive drift , accumulate the marked values to the corresponding drift counters, initialize the accumulation variable , and perform the accumulation operation , but in the weighted mode, give higher weight coefficients to the positive large drift , recalculate the accumulated value , the result shows that there is a slight positive divergence trend, finally combine the difference results of multiple groups with the accumulated trend value to obtain the cluster drift trend vector , which quantitatively describes the dynamic evolution law of eye diagram jitter in different phase intervals.

[0038] Please refer to Figure 4 , the specific steps of S3 are: S301: Based on the cluster drift trend vector, retrieve the time coordinates in the jump point cluster and call the offset parameter item in the offset distribution parameter set, compare the offset parameter item with the error compensation threshold, and aggregate the items greater than the threshold to obtain the weighted coefficient matrix value; The error compensation threshold is based on the statistical extraction of all offset parameter items in the offset distribution parameter set, and is a quantitative reference value obtained by performing a linear combination of the mean and standard deviation of all offset parameter items in the offset distribution parameter set. Based on the obtained cluster drift trend vector Initiate the abnormal transition point locking procedure, based on the markers in the vector. or The drift direction index is used to retrieve the corresponding time coordinates within the cluster of transition points in the cross-period edge time series, for example, to retrieve the set of edge times affected by positive drift. ,in Then call the generated offset distribution parameter set The phase offset parameter is extracted to construct a statistical sample space, and the sample size is [not specified]. Set as First, the floating-point arithmetic unit is called to perform an accumulation operation on the offset parameter and divide it by the sample size to obtain the arithmetic mean. Assuming the sum is Then the mean is calculated. Next, calculate the sum of squares of the differences between each parameter and the mean, and divide by . Obtaining the standard deviation by taking the square root Substituting the numerical values, we obtain At this point, based on the Gaussian distribution The principle is to set the error compensation threshold calculation logic and select weighting coefficients. Perform linear combination operations The error compensation threshold was calculated. After obtaining the quantization reference value, the offset parameter items corresponding to the jump point cluster are scanned one by one, and numerical comparison and judgment are performed. For example, for the corresponding offset parameter ,determination This item was identified as a significant source of jitter and retained, while for If an item is identified as natural noise, it will be removed. For the selected offsets exceeding a threshold, an aggregation operation will be performed, and a weighting coefficient will be calculated based on the magnitude of the deviation from the threshold. The calculation formula is as follows: ,for Calculated The weights of the retained items are calculated sequentially, ultimately generating a weighted coefficient matrix value for this cluster. This matrix precisely quantifies the contribution of each anomalous transition point to the overall time series drift.

[0039] S302: Call the weighting coefficient matrix value, perform proportional conversion on the time coordinates and weighting coefficients in the jump point cluster, aggregate the conversion results, calculate the offset based on the aggregated offset parameters, perform difference processing on the offset and the time coordinates, and obtain the time coordinate correction value; Call the weighting coefficient matrix value generated in the memory and the corresponding time coordinates in the jump point cluster, activate the timing error correction engine, perform proportional conversion for each controlled edge point, read the time coordinates corresponding to the original offset parameters and the weighting coefficients , first aggregate the conversion results to determine the effective drift component of the point, perform multiplication operation , calculate the effective drift value , which represents the non-random deviation of the time point caused by systematic jitter, then calculate the target theoretical offset based on the aggregated offset parameters, the goal in this model is to regress significant drift to the threshold boundary or ideal zero point, set the correction target to eliminate the effective drift component, i.e. calculate the offset , then perform difference processing on the offset and the original time coordinates to obtain the correction vector, since the previous detection shows positive lag drift (i.e. the time value is too large), the correction direction should be negative, construct the difference operation formula , substitute the values to obtain the time coordinate correction value , if it corresponds to negative lead drift, the sign is opposite, repeat this calculation process for abnormal points in the cluster , calculate the correction value , store the calculation results in the cache, form a complete sequence of time coordinate correction values, and the specific drift parameter processing and correction value calculation data are shown in Table 3.

[0040] Table 3: Abnormal jump point drift compensation calculation table.

[0041]

[0042] As shown in Table 3, the weighting correction value of each abnormal point is adaptively calculated according to the statistical threshold, which ensures that the compensation operation is only for significant drift error, and avoids excessive intervention on normal random jitter.

[0043] S303: Call the time coordinate correction value, perform superposition processing on the time coordinates and the correction value in the jump point cluster, perform time sorting and keep the index continuity, and obtain the edge positioning coordinates after drift compensation; The calculated time coordinate correction sequence and the original time coordinates within the jump point cluster are called, and a parallel adder array is started to perform superposition processing on the two to achieve physical coordinate relocation. For index 1, edge points, read the original coordinates With correction amount Perform algebraic addition The compensated coordinates are calculated. For indexes The point, calculate This process essentially pulls the diverging edge trajectory back to near the convergent ideal eye diagram trajectory. After updating the coordinates of the points, due to the significant correction causing a slight time sequence misalignment, the quicksort algorithm is immediately invoked with the updated time values. The entire sequence is reordered for the keywords, ensuring that edge events are strictly arranged in a monotonically increasing logical order. Simultaneously, an index continuity check is performed, traversing the sorted sequence. If index breaks are found due to filtering or merging, continuous logical indexes are reassigned. For example, the original index Mapped to the first in the new sequence Each valid bit outputs the final drift-compensated edge positioning coordinate sequence. The results indicate that the signal timing after statistical weighting compensation has reduced nonlinear drift interference, and the generated eye diagram data will exhibit higher clarity and opening when displayed in superimposed form.

[0044] Please see Figure 5 The specific steps of S4 are as follows: S401: Based on the edge positioning coordinates after drift compensation, detect the positioning coordinate time series, retrieve the first and last zero-crossing time positions of each bit period, calculate the first and last zero-crossing time interval, perform consistency judgment on the time interval sequence, and generate the zero-crossing interval amount. Call the edge positioning coordinate sequence generated after drift compensation processing First, activate the periodic retrieval engine in the timing logic analysis unit. This engine sets the bitstream length window to be analyzed to... Each standard unit interval positions the pointer to the first valid time coordinate of the sequence. ,For example And read the next adjacent time coordinate in sequence. ,For example The system executes the first and last zero-crossing point recognition logic to determine whether these two consecutive coordinate points constitute a complete signal level switching cycle, i.e., to verify whether there is a jump that crosses a logic threshold between them. Once confirmed... The starting point is zero and To terminate at zero, immediately start the floating-point subtractor to calculate the physical time span between the two and obtain the original time interval value. Substituting the numerical values, we obtain Then, a consistency check is performed on the calculated time interval sequence to filter it, and a reasonable fluctuation range of the effective bit period is preset. This range is based on the signal baud rate. Corresponding ideal cycle Set upper and lower limits, for example, set the lower limit as... And the upper limit is To exclude long-term connections or long-term connection Non-single-bit span data caused by the code pattern, if the calculated If a sample falls within this interval, it is marked as a valid consistency interval sample; otherwise, it is discarded as invalid data, and the traversal pointer continues to move and read. and Assuming The difference was calculated. The value exceeded the set range, was determined to be a non-single-bit period and skipped. Through scanning and filtering the entire sequence, zero-crossing interval data that met the single-bit characteristics were extracted one by one, and the data was stored in a high-speed buffer queue according to its original time-series index. Finally, a sequence containing... Sequence of zero-crossing intervals of valid elements This sequence not only records the actual duration of each independent bit, but also preserves the instantaneous pulse width modulation characteristics caused by high-frequency noise or inter-symbol interference during signal transmission, providing an accurate time-domain measurement basis for subsequent fine-grained correction.

[0045] S402: Call the zero-crossing interval amount, perform difference calculation on the zero-crossing interval amount according to the standard cycle length reference value, compare the difference with the correction threshold, when the difference exceeds the correction threshold, trigger the correction command and record the trigger identifier, and establish the cycle difference identifier amount; Call the zero-crossing interval sequence in the storage queue First, extract the rated frequency parameters of the clock reference signal. The nominal period parameter of the rated bit period, i.e., the standard period length reference value, is obtained by using the reciprocal operation relationship. Then select the first part of the sequence Using the zero-crossing intervals in a stable state as a statistical sample, mean and deviation analysis is performed to determine the correction threshold, and the statistical coprocessor is invoked to calculate the sample mean. with standard deviation Assuming the calculation yields , ,in accordance with Statistical principles set a fixed quantification threshold, and the calculation formula is as follows: Substituting the values ​​into the calculation yields the correction threshold. After establishing the baseline and threshold, the interval for each zero-crossing point in the sequence is... Perform the difference calculation and construct the difference operation formula. For example, for the first interval quantity Calculate the difference For another interval Calculate the difference The absolute value of the calculated difference is compared with the correction threshold. Perform numerical comparisons, targeting ,determination This is within the normal fluctuation range and will not trigger correction. ,determination Upon confirming a significant distortion in the bit period, a correction command is immediately triggered, and the trigger flag is recorded. Meanwhile, a periodic difference identifier containing the difference magnitude, direction sign and corresponding edge index is established. The entire sequence is traversed, and the excess difference events are recorded in detail in the correction register file. The specific periodic detection and threshold determination data are shown in Table 4.

[0046] Table 4: Bit Period Difference Detection and Correction Judgment Table.

[0047]

[0048] As shown in Table 4, by strictly comparing thresholds, abnormal bit periods that deviate too much from the nominal value were identified, thus locking in the target object for subsequent targeted geometric correction.

[0049] S403: Call the periodic difference identifier, and adjust the time coordinate of the edge positioning coordinate after drift compensation according to the difference direction and difference magnitude recorded in the identifier. Converge the adjusted edge time series to obtain the edge coordinate after periodic correction. The period difference flag in the calibration register file is invoked to start the time coordinate fine-tuning controller. Time coordinate adjustments are then performed on the specific edges marked as triggering calibration, using the indices in Table 4 as... Taking the anomaly as an example, the periodic difference is This indicates that the current bit period is too long, causing a zero-crossing point at the end of the period. A hysteresis drift occurred relative to the ideal position. The original drift-compensated edge positioning coordinates of that point were read. The adjustment strategy is determined based on the direction of the difference. For positive differences (larger period), a contraction adjustment is performed, and the adjustment amount is calculated. This means only pulling back the portion exceeding the threshold, or using a full regression strategy. Here, a full regression to the nominal value strategy is used to maximize the eye diagram opening, and the adjustment amount is set. Perform addition to update coordinates For indexes negative difference (Period is too short), perform an expansion adjustment and calculate the adjustment amount. Update the corresponding coordinates The above coordinate translation operation is performed sequentially on the recorded identifiers. For normal points that have not triggered correction, the original coordinates are kept unchanged. After completing the traversal and correction of the entire sequence, the adjusted edge time points are aggregated into a new memory space, and a timing integrity check is performed again to ensure that the adjustment operation has not caused a timeline reversal conflict. Finally, the periodically corrected edge coordinate sequence is obtained. Each data point in this sequence has been rigorously aligned to the ideal bit clock grid, eliminating periodic width and narrow distortion caused by non-uniform transmission medium or driver jitter, and can be directly used to generate high-precision jitter-free eye diagram models.

[0050] Please see Figure 6 The specific steps of S5 are as follows: S501: Based on the edge coordinate sequence after periodic correction, the discrete amplitudes within the sequence are decomposed in the frequency domain based on Fourier transform, and the sampled values ​​are compared with the frequency index. After comparison, amplitude positioning and index aggregation are performed, and serialization processing is carried out to generate a frequency domain amplitude array. Based on the generated periodically corrected back edge coordinate sequence First, the frequency domain signal processing engine is activated, and the discrete amplitude generation module within the sequence is called to transform the edge position deviation in the time domain into a discrete time interval error sequence. Set the number of sampling points The time resolution corresponds to the system clock cycle. For example, before extraction The error value at each point is Then, a frequency domain decomposition operation based on Fast Fourier Transform (FFT) is performed to map the time-domain sequence to the frequency domain space, and the complex spectral coefficients at each frequency point are calculated. For the fundamental frequency... (Assuming the sampling rate is) and ), calculate the real part With the imaginary part The amplitude of the sampled value is calculated based on the modulo operation logic. Calculate one by one to The amplitude corresponding to the frequency index is used to generate the original spectrum data. Then, the calculated sampled values ​​are compared with the frequency index using a scanning process, with the background noise floor threshold set as... It identifies frequency components that are significantly higher than the base, for example, in the index. (correspond The amplitude value is detected at index The amplitude value is detected at index The amplitude value is detected at index The amplitude value is detected at index The amplitude value is detected at index , , The amplitude value is detected at index The amplitude value is detected at index The amplitude value is detected at index The amplitude value is detected at index

[0051] S502: Call the frequency domain amplitude array, perform amplitude difference retrieval on the frequency sample value and adjacent sample value in the array, compare the difference value with the amplitude difference threshold value, perform aggregation encoding and sequence arrangement on the index greater than the amplitude difference threshold value, and obtain the jitter frequency index set; Call the frequency domain amplitude array in the memory, start the jitter component screening controller, perform amplitude difference retrieval on each frequency sample value and its adjacent sample value in the array, take the aggregation amplitude of index as an example, retrieve its left adjacent value (which has been processed for sidelobe suppression), calculate the difference value To accurately determine whether the difference is from deterministic jitter or random noise fluctuation, a dynamic amplitude difference threshold value needs to be constructed. First, perform statistical quantity extraction on all discrete amplitude sample values in the frequency domain amplitude array, identify the maximum amplitude and the minimum amplitude in the array, calculate the interval span parameter of the amplitude distribution interval, and calculate the amplitude variation of the discrete amplitude sample value, that is, the first-order difference standard deviation of the full sequence amplitude , select a proportion coefficient according to the set signal-to-noise ratio requirement, set the span weight and the variation weight , perform linear weighting calculation to obtain the quantitative value , substitute the numerical value into the numerical calculation to obtain After obtaining the threshold value, compare the calculated difference value with the amplitude difference threshold value , and determine , confirm as significant jitter frequency, for the determination result greater than amplitude difference threshold value index execution aggregation encoding, will index Marked as a specific interference source (such as power switch noise), and remove redundant harmonic components, sequence arrangement, specific frequency domain difference analysis and screening data as shown in table 5, finally get jitter frequency index set , the set accurately locks the main periodic interference frequency point affecting the eye diagram quality.

[0052] Table 5 frequency domain amplitude difference screening analysis table.

[0053]

[0054] As shown in table 5, through the differential screening of dynamic threshold, the deterministic jitter component with high energy characteristics is effectively extracted from the complex spectrum background.

[0055] S503: call jitter frequency index set, execute amplitude conversion on the amplitude parameters corresponding to the index frequency and the edge position sequence time interval parameters, and perform weighted superposition and amplitude sequence regularization on the converted amplitude value to obtain the eye diagram jitter detection result; Call the jitter frequency index set determined by screening And its associated frequency domain parameters, activate the eye diagram parameter synthesis unit, execute amplitude conversion on the amplitude parameters corresponding to the index frequency and the edge position sequence time interval parameters, and restore the single sideband amplitude In the time domain peak to peak jitter contribution , using the sine jitter conversion logic, the formula is , for example, for Component, calculate , for Component, calculate , then the converted amplitude value is executed weighted superposition, considering the tracking suppression ability of the receiver clock recovery circuit (CDR) to low frequency jitter, according to the position of the frequency point relative to the CDR loop bandwidth Set the weighting coefficient, assuming the loop bandwidth is , the above High frequency jitter is out of band and cannot be tracked and eliminated, so it is fully included, set the weighting coefficient , execute superposition operation This result is the total amount of deterministic jitter, then combined with the random jitter Estimate value, assuming , the peak to peak coefficient corresponding to the bit error rate Is , calculate the total jitter , finally perform amplitude sequence regularization, compare the total jitter value with the standard unit interval , calculate the eye opening , substituting the numerical values , while calculating the horizontal opening ratio , obtaining the eye diagram jitter detection result The result quantitatively reflects the actual effective sampling window width that the signal link can provide at the receiving end after eliminating systematic interference and comprehensively considering random noise.

[0056] Please refer to Figure 7 , the high-speed oscilloscope eye diagram real-time jitter detection system, comprising: The original waveform sampling module acquires the digitized waveform of the measured signal through a high-speed analog-to-digital converter, detects the zero-crossing point position based on a zero-crossing detection algorithm, calculates the time offset, groups according to the bit period, and obtains the cross-period edge time sequence; The edge aggregation analysis module performs clustering analysis on the zero-crossing points of adjacent bit periods at the same phase according to the cross-period edge time sequence, merges the edge time points into a jump point cluster, calculates the deviation, and obtains a cluster drift trend vector; The drift compensation calculation module assigns error compensation weights to the time points in the jump point cluster based on the cluster drift trend vector, calculates the time coordinate correction amount through weighted calculation, and adds it to the corresponding time coordinate to obtain the drift-compensated edge positioning coordinate; The period calibration monitoring module calls the drift-compensated edge positioning coordinate, monitors the time interval between the first and last zero-crossing points of the bit period through a period boundary detection circuit, calculates the difference value from the standard period length, triggers a correction instruction when the difference value exceeds the correction threshold, and obtains the period-corrected edge coordinate. The spectrum feature extraction module analyzes the spectrum characteristics of the edge position sequence based on the period-corrected edge coordinate using the Fourier transform algorithm, extracts the jitter frequency component and amplitude, calculates the jitter amplitude parameter, and generates the eye diagram jitter detection result.

[0057] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any skilled person in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for real-time jitter detection of an eye pattern of a high-speed oscilloscope, characterized in that, The method comprises the following steps: S1: acquiring a digital waveform of a measured signal by a high-speed analog-to-digital converter, detecting a zero-crossing position corresponding to a timestamp based on a zero-crossing detection algorithm, calculating a time offset, grouping according to a bit period, and obtaining a cross-period edge time sequence; S2: according to the cross-period edge time sequence, performing cluster analysis on the zero-crossing points of adjacent bit periods at the same phase, merging the edge time points into a jump point cluster, calculating a deviation, and obtaining a cluster drift trend vector; S3: based on the cluster drift trend vector, assigning an error compensation weight to the time points in the jump point cluster, calculating a time coordinate correction amount by weighting, and superimposing the correction amount on the corresponding time coordinate to obtain a drift-compensated edge positioning coordinate; S4: calling the drift-compensated edge positioning coordinate, monitoring the time interval between the first and last zero-crossing points of a bit period through a period boundary detection circuit, calculating the difference value from the standard period length, and triggering a correction instruction when the difference value exceeds a correction threshold to obtain a period-corrected edge coordinate; S5: according to the period-corrected edge coordinate, analyzing the spectral characteristics of the edge position sequence using a Fourier transform algorithm, extracting the jitter frequency component and amplitude, calculating the jitter amplitude parameter, and generating an eye diagram jitter detection result.

2. The high speed oscilloscope eye diagram real time jitter detection method of claim 1, wherein, The cross-period edge time sequence includes timestamps, bit period grouping, and cross-period jump points. The jump point cluster includes a cluster center, an offset, and a dispersion. The drift-compensated edge positioning coordinate includes a correction amount, a compensation weight, and a positioning accuracy. The period-corrected edge coordinate includes a period length deviation, a period consistency, and a correction deviation. The eye diagram jitter detection result includes a jitter frequency component, an amplitude component, and a spectral characteristic.

3. The high speed oscilloscope eye diagram real time jitter detection method of claim 1, wherein, The specific steps of S1 are as follows: S101: acquiring a digital waveform through a high-speed analog-to-digital converter, performing comparison on adjacent sampling point amplitudes based on a zero-crossing detection algorithm and positioning the sign change point, performing linear proportional interpolation on the sign change point according to the amplitude difference of two sampling points and determining the cross time position to generate a zero-crossing timestamp sequence; S102: calling the zero-crossing timestamp sequence, performing difference processing on the difference value of adjacent time positions and forming a time displacement correlation value, then performing proportional mapping on the correlation value according to the sampling time interval to obtain a time displacement set, and performing segmented aggregation according to the sampling time sequence to obtain a period time offset matrix; S103: based on the period time offset matrix, sorting the multiple segments according to the index position within the segment, aggregating the segment sequence of the same index position to obtain a cross-period edge time sequence.

4. The high speed oscilloscope eye diagram real time jitter detection method of claim 3, wherein, The specific steps of S2 are as follows: S201: based on the cross-period edge time sequence, performing retrieval action on the zero-crossing time values of adjacent bit periods at the same phase, and performing difference operation on the time values and phase reference time values, then arranging the difference values according to the phase index sequence to generate a zero-crossing difference matrix; S202: calling the zero-crossing difference matrix, performing aggregation on the difference vector according to the difference amplitude and phase index, extracting the offset sequence from the aggregated vector, and arranging the offsets according to the group order to obtain a set of offset distribution parameters; S203: calling the offset distribution parameter set, performing a serial arrangement action on the offset parameters in index order, performing a difference operation on adjacent offset parameters to form a difference sequence, and then extracting a direction change item according to the sequence position and accumulating it to obtain a cluster drift trend vector.

5. The high speed oscilloscope eye diagram real time jitter detection method of claim 4, wherein, The specific steps of S3 are: S301: based on the cluster drift trend vector, retrieving the time coordinates in the jump point cluster and calling the offset parameter item in the offset distribution parameter set, performing a comparison between the offset parameter item and the error compensation threshold, and aggregating the items greater than the threshold to obtain a weighted coefficient matrix value; S302: calling the weighted coefficient matrix value, performing a proportional conversion between the time coordinates in the jump point cluster and the weighted coefficient, and aggregating the conversion results, and calculating the offset based on the aggregated offset parameters, and performing a difference processing between the offset and the time coordinates to obtain a time coordinate correction value; S303: calling the time coordinate correction value, performing a superposition processing between the time coordinates in the jump point cluster and the correction value, performing time sorting and maintaining index continuity to obtain the edge positioning coordinates after drift compensation.

6. The high speed oscilloscope eye diagram real time jitter detection method of claim 5, wherein, The specific steps of S4 are: S401: based on the edge positioning coordinates after drift compensation, detecting the positioning coordinate time sequence, retrieving the first and last zero-crossing point time positions of each bit period, calculating the first and last zero-crossing point time interval, and performing a consistency judgment on the time interval sequence to generate a zero-crossing point interval amount; S402: calling the zero-crossing point interval amount, performing a difference calculation on the zero-crossing point interval amount according to the standard period length reference value, comparing the difference value with the correction threshold, and when the difference value exceeds the correction threshold, triggering a correction instruction and recording a trigger identifier, and establishing a period difference identifier amount; S403: calling the period difference identifier amount, performing a time coordinate adjustment on the edge positioning coordinates after drift compensation according to the difference direction and difference amplitude recorded in the identifier amount, and converging the adjusted edge time sequence to obtain the period-corrected edge coordinates.

7. The high-speed oscilloscope eye diagram real-time jitter detection method of claim 6, wherein, The correction threshold is a fixed quantization threshold determined according to the statistical dispersion degree of the zero-crossing point interval amount recorded in the stable segment of the original bit period, and the statistical dispersion degree is obtained by performing mean and deviation analysis on the zero-crossing point interval amount in the original bit period; The standard period length reference value is a reference period value obtained according to the nominal period parameter of the rated bit period, and the nominal period parameter is obtained by converting the rated frequency of the clock reference signal.

8. The high speed oscilloscope eye diagram real time jitter detection method of claim 6, wherein, The specific steps of S5 are: S501: based on the period-corrected edge coordinate sequence, performing a frequency domain decomposition based on Fourier transform on the discrete amplitude in the sequence and performing a comparison between the sample value and the frequency index, performing amplitude positioning and index aggregation after the comparison and performing a serialization processing to generate a frequency domain amplitude array; S502: calling the frequency domain amplitude array, performing an amplitude difference retrieval between the frequency sample value and the adjacent sample value in the array, and performing a comparison between the difference value and the amplitude difference threshold, performing an aggregation encoding for the indexes greater than the amplitude difference threshold and performing a sequence arrangement to obtain a jitter frequency index set; S503: calling the jitter frequency index set, performing amplitude conversion on the index frequency corresponding amplitude parameter and edge position sequence time interval parameter, and performing weighted superposition and amplitude sequence regularization on the converted amplitude value, to obtain an eye diagram jitter detection result.

9. The high speed oscilloscope eye diagram real time jitter detection method of claim 8, wherein, The amplitude difference threshold is based on the statistics of all discrete amplitude sample values in the frequency domain amplitude array, and is obtained by calculating the quantitative value of the interval span parameter of the amplitude distribution interval and the amplitude variation of the discrete amplitude sample value by a proportional coefficient.

10. A high-speed oscilloscope eye diagram real-time jitter detection system, characterized by, The system is used to implement the high-speed oscilloscope eye diagram real-time jitter detection method of any one of claims 1-9, and the system comprises: An original waveform sampling module acquires digitized waveforms of a measured signal through a high-speed analog-to-digital converter, detects time stamps corresponding to zero-crossing points based on a zero-crossing detection algorithm, calculates time offset, groups by bit period, and obtains a cross-period edge time sequence; An edge aggregation analysis module performs clustering analysis on zero-crossing points of the same phase in adjacent bit periods according to the cross-period edge time sequence, merges edge time points into a jump point cluster, calculates a deviation, and obtains a cluster drift trend vector; A drift compensation calculation module assigns error compensation weights to time points in the jump point cluster based on the cluster drift trend vector, calculates time coordinate correction amounts through weighted calculation, and adds the correction amounts to corresponding time coordinates to obtain drift-compensated edge positioning coordinates; A period calibration monitoring module calls the drift-compensated edge positioning coordinates, monitors the time interval between the first and last zero-crossing points of a bit period through a period boundary detection circuit, calculates the difference value from the standard period length, triggers a correction instruction when the difference value exceeds a correction threshold, and obtains period-corrected edge coordinates; A spectrum feature extraction module analyzes the spectral characteristics of the edge position sequence according to the period-corrected edge coordinates using a Fourier transform algorithm, extracts jitter frequency components and amplitudes, calculates jitter amplitude parameters, and generates an eye diagram jitter detection result.

Images