CN-121997208-A - Sampling oscilloscope signal jitter analysis method based on TLC improved algorithm

Abstract

The invention discloses a signal jitter analysis method of a sampling oscilloscope based on a TLC improved algorithm, which can better separate periodic jitter from random jitter in total jitter, firstly obtain a jitter TIE data set of the total jitter through waveforms of detected signals, simultaneously perform TLC sample function calculation operation and FFT spectrum analysis operation on the TIE data set, screen out frequency points of each periodic jitter according to the obtained frequency spectrum, further construct a function to be fitted, obtain a fitting sample after interpolation operation is performed on the TLC sample function, perform least square fitting operation, obtain a periodic jitter amplitude through fitting, and finally solve the variance of the random jitter, thereby completing the jitter decomposition process. The invention improves TLC jitter decomposition algorithm, and the signal jitter analysis performance of the sampling oscilloscope is obviously improved by using the method provided by the invention.

Inventors

• ZHANG ZHENSHENG
• ZHANG ZHUOHANG
• ZHAO TONGGANG
• XU ZHENGSHAN
• Pan Dafa
• CHENG KUN
• GAO Ziang
• XIN XIANGJUN
• HU HAIYANG
• PAN WEI
• FANG SHUO

Assignees

• 北京邮电大学

Dates

Publication Date

20260508

Application Date

20241104

Claims (6)

1. 1. The signal jitter analysis method of the sampling oscilloscope based on the TLC improved algorithm is characterized by comprising the following steps of: and step 1, acquiring a TIE data set according to the measured waveform sample. And 2, performing TLC function calculation operation and FFT spectrum generation operation simultaneously. And step 3, processing the TLC sample function to obtain a fitting sample, screening the frequency spectrum and further constructing a function to be fitted. Step 4, performing least square fitting operation And 5, calculating the random jitter variance.
2. 2. The TLC function calculation operation of step 2 of claim 1, wherein: in calculating the TLC function, an unbiased autocorrelation function is used.
3. 3. The method for processing TLC sample functions according to claim 1, wherein the processing includes preprocessing and interpolation: wherein the pretreatment operation includes 1. Only the first half of the TLC sample function is truncated as a preliminary fit sample. 2. Modifying the zero point of the TLC sample function to be the value of the first function point thereafter; The interpolation algorithm used in the interpolation operation is particularly a cubic spline interpolation algorithm.
4. 4. The method for constructing a function to be fitted according to step 3 of claim 1, wherein the function to be fitted is a TLC function that removes a random jitter variance term and substitutes a periodic jitter frequency point value as follows: the number of PJ screened from the frequency spectrum obtained by FFT is the specific item number of the function to be fitted.
5. 5. The least square fitting operation according to step 4 of claim 1, wherein the fitting operation is specifically that the squares of the differences between the fitting samples and the function to be fitted are summed and used as criteria for the fitting effect, and the smaller the value is, the better the fitting effect is, and the fitting is divided into a first-order fitting and a second-order fitting in consideration of timeliness, wherein the fitting range of the first-order fitting is 0 to the maximum value of the fitting samples, the fitting step length is 5% of the fitting range, the fitting range of the second-order fitting is about 20% of the fitting result of the first-order fitting, and the fitting step length is 2% of the fitting range.
6. 6. The method for calculating random jitter variance as defined in claim 1, wherein the theoretical TLC function model is as follows: The zero point of the theoretical TLC function contains the variance information of the RJ, and the PJ amplitude obtained by least square fitting and the zero point of the TLC sample function are brought into the TLC function, so that the variance term value of the RJ is solved.

Description

Sampling oscilloscope signal jitter analysis method based on TLC improved algorithm Technical Field The invention relates to a signal quality analysis module of a broadband sampling oscilloscope, in particular to a method for decomposing signal jitter. Background With the development of electronic technology, the performance of digital circuit systems is rapidly improved, and particularly for high-speed serial communication systems, the transmission rate of the digital circuit systems can reach GHz, and for such high-speed signals, jitter can have a great influence on the error rate of the system, which has become a key factor for limiting the upper limit of the performance of the system. Jitter is often defined as the deviation of a signal edge from a threshold time from an ideal signal edge time. The generation of the total jitter is compounded by a plurality of factors, and the factors can be primarily divided into Random Jitter (RJ) and Deterministic Jitter (DJ) according to the different factors, wherein the DJ can be further divided into Periodic Jitter (PJ), data correlation jitter (DDJ) and Bounded Uncorrelated Jitter (BUJ). While DDJ can be subdivided into intersymbol interference (ISI) and Duty Cycle Distortion (DCD). In actual analysis, engineers can infer main factors for generating jitter according to the size of each jitter component in the tested circuit system, and the jitter is reduced more efficiently by symptomatic drug delivery, so that the performance of the circuit system is further improved. Currently, the dominant dither decomposition algorithms are Fast Fourier Transform (FFT) algorithm, tail-fitting (Tail-Fit) algorithm, empirical Mode Decomposition (EMD) algorithm, and time delay correlation (TLC) algorithm. However, the methods have some defects in different degrees, such as a frequency spectrum leakage problem of an FFT algorithm, a high requirement on the number of jitter samples is put forward by a Tail-Fit algorithm, the algorithm is difficult to run smoothly under the condition of small sample size, the smoothing operation in the method can lead the information of the original Probability Density Function (PDF) statistical histogram to be defective, so that errors are brought to fitting results, and modal aliasing in an EMD algorithm is unfavorable for analyzing the jitter signal containing transient components and can also influence the jitter decomposition precision. The TLC algorithm also has defects, because RJ is not an ideal Gaussian distribution in practice, the calculated TLC sample function has burrs, the burrs can greatly influence the selection of data points, so that the accuracy and stability of the algorithm are reduced, and meanwhile, the equation solving operation is very time-consuming, so that the timeliness of the algorithm is also insufficient, and the method is difficult to be practically applied to a signal jitter analysis method of a sampling oscilloscope. Disclosure of Invention The invention improves the original TLC algorithm, aims to overcome the defects of the original algorithm, and provides a jitter decomposition algorithm with more excellent performance so as to realize the application of the jitter decomposition algorithm in signal jitter analysis of a sampling oscilloscope. The invention provides an algorithm based on TLC improvement, which realizes the effective separation of periodic jitter and random jitter components, and comprises the following steps: and step 1, obtaining jitter Time Interval Error (TIE) data samples according to the waveform of the detected signal. And 2, carrying out TLC function operation and FFT operation according to the obtained TIE sample. And step3, preprocessing the obtained TLC sample function, and then performing cubic spline interpolation operation to obtain a fitting sample. And meanwhile, screening out the frequency points of the PJ according to the obtained sample frequency spectrum, and further constructing a function to be fitted. And 4, performing fitting operation based on a least square method to obtain the amplitude of the PJ. And 5, substituting the obtained frequency points and amplitude values of each PJ and zero points of the TLC sample function into the TLC function, so as to solve the variance of the RJ. The method has the main advantages that the existing thought of combining the TLC algorithm and the FFT algorithm can quickly and accurately obtain the frequency point value of each PJ, simultaneously reduce the unknown number in the equation, improve the timeliness of the algorithm, and on the basis, the method combines the least square fitting operation, and extracts effective information from the TLC sample function with burrs as much as possible by fitting the amplitude of the PJ, thereby improving the accuracy and the stability of the original algorithm, thoroughly avoiding the operation of equation knowing, greatly reducing the operation time of the algorithm, and improving the timelines