Jitter extraction method for high-speed serial links based on cyclostationary characteristics

Abstract

The present invention belongs to the technical field of data jitter extraction of high-speed serial links, and specifically discloses a high-speed serial link jitter extraction method based on cyclostationary characteristics, comprising: using an adaptive algorithm to determine the turning point of an initial codeword; dividing the data according to a sampling multiple N, and calculating the cyclic autocorrelation function at the codeword rate; obtaining the phase of the cyclic autocorrelation function, which contains the jitter information of each codeword; fitting the above phase using the least squares method, removing the accumulated jitter error caused by the inherent phase shift of the channel, etc., and obtaining the accurate jitter information of each codeword of the high-speed serial link. The method of the present invention has the advantages of simple implementation and low calculation complexity.

Description

Technical Field

[0001] The invention relates to the technical field of high-speed serial communication link data signal jitter measurement, and in particular to a high-speed serial link jitter extraction method based on cyclostationary characteristics. Background Art

[0002] Jitter is a core indicator for measuring the stability and reliability of data transmission. As data transmission rates continue to increase in electronic and optical communication applications, signal integrity issues in many high-speed applications are related to jitter. Jitter reduces timing margins, causing functional problems in circuits / systems and affecting the reliability of data transmission. Jitter measurement, as an important part of high-speed serial communication link test systems, is a key indicator for measuring system performance and bit error rate. Fast and reliable jitter extraction technology plays an important role in system debugging and verification, and can accelerate system fault diagnosis and design iterations.

[0003] In high-speed serial link test and analysis systems, time interval error (TIE) jitter is usually used to characterize the jitter information of each data symbol. TIE jitter refers to the difference between the time when the high-speed serial data signal crosses the threshold voltage and the time when the clock signal crosses the threshold voltage. The clock signal refers to the clock signal recovered from the data signal under test.

[0004] Traditional TIE jitter extraction methods such as Figure 7 As shown, firstly, all sampling points are stored, and the reference level or threshold voltage of the current data is determined by the histogram statistics method; secondly, the clock signal embedded in the signal is extracted through the internal hardware / software clock recovery unit CDR to obtain the ideal / reference clock signal; the time difference between the data signal and the clock signal crossing the threshold voltage is determined, and in order to obtain accurate time information, an interpolation operation is required; the clock edge is subtracted from the data signal edge to obtain the TIE jitter trend function. The implementation method can be obtained through external instruments such as real-time sampling oscilloscopes, time interval analyzers, and on-chip jitter measurement circuits. When performing jitter TIE extraction in the above methods, the hardware circuit is complex and the software calculation is large. Since the measurement error of the instrument itself is difficult to quantify, for example, the CDR of each manufacturer is slightly different in specific implementation, resulting in slightly different jitter extraction and jitter decomposition results between different manufacturers. Summary of the invention

[0005] The purpose of the present invention is: to address the above-mentioned problems, the present invention provides a high-speed serial link jitter extraction method based on cyclostationary characteristics, based on the cyclostationary characteristics of the transmission signal, extracts the phase information of the cyclic autocorrelation function at a frequency of the symbol rate, and uses the least squares method to remove the cumulative error caused by the non-ideal clock to obtain the TIE jitter information of each symbol.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A high-speed serial link jitter extraction method based on cyclostationary characteristics, including:

[0008] Obtaining a cyclic autocorrelation function of the measured data signal at its symbol rate;

[0009] Acquire a phase of the cyclic autocorrelation function, where the phase includes jitter information of the measured data signal, and obtain coarse TIE jitter information;

[0010] The rough TIE jitter information is subjected to a de-accumulated error process to obtain accurate TIE jitter information.

[0011] Furthermore, the cyclic autocorrelation function acquisition includes:

[0012] Determining the number of sampling points of each data symbol in the measured data signal;

[0013] Adopting an adaptive algorithm to determine the starting point of the cyclic autocorrelation function of each data symbol;

[0014] The phase of the cyclic autocorrelation function of each data symbol is obtained by using FFT transformation.

[0015] Furthermore, the starting point includes finding the first extreme point of the measured data signal by a derivation method, which is the starting point.

[0016] Furthermore, after the starting point is determined, the measured data signal is segmented with a sampling multiple N to obtain a plurality of data symbols, and the cyclic autocorrelation function of each data symbol is calculated in turn.

[0017] Furthermore, the phase of the cyclic autocorrelation function is specifically expressed as follows:

[0018] (1)

[0019] In the formula, is the cyclic autocorrelation function of the measured signal at its symbol rate, is the phase of the cyclic autocorrelation function, is the phase offset of the channel, For accurate TIE jitter information.

[0020] Furthermore, obtaining accurate TIE jitter information includes:

[0021] The jitter information of the measured data signal is obtained by performing direct current removal processing on the phase information of the cyclic autocorrelation function and then removing the accumulated error jitter generated by the channel effect in the phase information.

[0022] Furthermore, the least square method is used to find the phase The curve is fitted and the phase is reversed Compensate to obtain accurate TIE jitter information .

[0023] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0024] The method proposed in the present invention is simple to implement. Compared with the traditional zero-crossing-based method for extracting TIE jitter, the traditional method needs to go through multiple processing links, such as first using a statistical method to determine the reference voltage, then recovering the clock signal from the data, and calculating the time difference between the data signal and the clock signal crossing the reference voltage by interpolation. The method of the present invention uses the collected data to obtain TIE jitter information through a simple digital signal processing link.

[0025] The method proposed in the present invention has low computational complexity. The traditional method uses a statistical method to determine the reference voltage, which requires calculation processing of all code elements to be analyzed. Secondly, when obtaining the time difference between the data signal and the clock signal crossing the reference voltage, an interpolation algorithm needs to be used, which has a large amount of calculation. The method of the present invention processes data at the sampling point of each code element, which is simple and fast and has low computational complexity.

[0026] The method proposed by the present invention can avoid the phenomenon that a non-quasi-clock signal causes TIE loss due to a code element not crossing a decision threshold.

[0027] In summary, the present invention can directly obtain the jitter information of each data codeword from the sampled data based on the cyclostationary characteristics of the high-speed serial link, without the need to pre-set the threshold voltage module; without the need to recover the clock signal from the data; without the need for an interpolation module to accurately solve the crossing moment of the signal and the threshold voltage; the implementation is simple, the calculation complexity is low, and the jitter information of the measured data can be quickly obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a flow chart of a high-speed serial link jitter extraction method based on cyclostationary characteristics of the present invention;

[0029] Figure 2 It is a principle block diagram of a high-speed serial link jitter extraction method based on cyclostationary characteristics of the present invention;

[0030] Figure 3 It is a phase result diagram of the cyclic autocorrelation function in the high-speed serial link jitter extraction method based on cyclostationary characteristics of the present invention;

[0031] Figure 4 It is a TIE jitter information result diagram in the high-speed serial link jitter extraction method based on cyclostationary characteristics of the present invention;

[0032] Figure 5 This is a diagram of oscilloscope jitter extraction results in a verification example of a high-speed serial link jitter extraction method based on cyclostationary characteristics of the present invention;

[0033] Figure 6 is a jitter extraction result diagram of a high-speed serial link jitter extraction method based on cyclostationary characteristics of the present invention;

[0034] Figure 7 This is the principle block diagram of the traditional method TIE jitter extraction. DETAILED DESCRIPTION

[0035] The present invention will be described in detail below in conjunction with the accompanying drawings.

[0036] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0037] Example

[0038] This embodiment provides a high-speed serial link jitter extraction method based on cyclostationary characteristics, such as Figure 1 and Figure 2 As shown, specifically including:

[0039] Acquiring data includes acquiring the measured data signal from an acquisition device such as a real-time oscilloscope or an ADC acquisition module, and determining the starting point:

[0040] In this embodiment, the first turning point of the high-speed serial link signal under test needs to be determined, and the first extreme point can be found by using the derivation method; after the starting point is determined, the data is segmented by the sampling multiple N, and the cyclic autocorrelation function of each data symbol is calculated in turn;

[0041] Calculate the cyclic autocorrelation function. High-speed digital serial links usually use baseband transmission and pulse amplitude modulation (PAM). The baseband pulse amplitude modulation signal model with jitter can be expressed as:

[0042] (2)

[0043] In the formula, is the first The value of a symbol is an independent and identically distributed random variable with zero mean and normalized variance of 1. is the duration of a code element, is the unit impulse response function of the system, is the normalized total timing jitter for each symbol, is the data currently received by the system, is additive white Gaussian noise.

[0044] After derivation, the jitter information of each data symbol is contained in the phase of the cyclic autocorrelation function at the symbol rate, as shown in the following equation:

[0045] (1)

[0046] In the formula, is the cyclic autocorrelation function of the input data information at the symbol rate, is the phase of the cyclic autocorrelation function, is the inherent offset of the channel. From formula (1), it can be seen that when the phase of the cyclic autocorrelation function of the measured signal is obtained, the jitter information of the code element can be obtained by simply removing the cumulative error from the phase information.

[0047] In order to quickly obtain the cyclic autocorrelation function of the current symbol, an algorithm such as Fast Fourier Transform FFT may be used to perform N-point FFT calculation on the autocorrelation function of each symbol to obtain the cyclic autocorrelation function of the current symbol.

[0048] According to formula (1), the phase information is obtained by performing a phase angle operation on the cyclic autocorrelation function, as shown in Figure 3 As shown, the acquired cyclic autocorrelation phase contains a negative phase cumulative phase error, which is the inherent offset of the channel. .

[0049] Get TIE jitter information. In order to obtain accurate TIE jitter information , the phase of the cyclic autocorrelation function needs to be To remove the accumulated error, within the passband, since the phase offset of the channel is fixed, the least square method can be used to calculate the phase The curve is fitted and the phase is reversed Compensation is performed to obtain accurate symbol jitter information, such as Figure 4 shown.

[0050] Verification Example

[0051] In order to verify the effectiveness of this embodiment, the following test platform was built. The vector signal generator R&S®SMW200A generated a 250Mbps PAM2-NRZ baseband signal, added a 100KHz sine phase modulation, and used a Tektronix MSO64 oscilloscope for jitter extraction analysis. The oscilloscope had a bandwidth of 500MHz, a sampling rate of 6.25GHz, and a storage depth of 1MSa. The collected samples were sent to the host computer through the GPIB bus. At the same time, this embodiment was used to compare the jitter extraction results.

[0052] Figure 5 The measurement results of the oscilloscope are shown, which are TIE time domain trend chart, TIE frequency domain spectrum chart at the lower left, and TIE statistical domain histogram and jitter quantification information at the lower right. The data collected by the oscilloscope is used in the jitter extraction method of this embodiment. The results are as follows Figure 6 As shown; the following Table 1 lists the jitter information errors extracted by this embodiment and the oscilloscope.

[0053] Table 1 Jitter information error between this embodiment and the one extracted by the oscilloscope

[0054]

[0055] Depend on Figure 6 It can be concluded that the TIE jitter information obtained by this embodiment is consistent with that obtained by the oscilloscope. Table 1 lists in detail the quantitative indicators of TIE jitter extracted by the present invention and the oscilloscope, and the relative error verifies the effectiveness and accuracy of this embodiment. Compared with the oscilloscope measurement results, this embodiment provides a simplified implementation framework, significantly reduces the computational complexity, and alleviates the uncertainty associated with modules such as clock recovery, threshold determination, and crosspoint interpolation.

[0056] The principles and implementation methods of the present invention are described in this article using specific embodiments. The description of the above embodiments is only used to help understand the method and core ideas of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

Claims

1. A high-speed serial link jitter extraction method based on cyclostationary characteristics, characterized in that: include: Obtaining a cyclic autocorrelation function of the measured data signal at its symbol rate; Acquiring the phase of the cyclic autocorrelation function, wherein the cyclic autocorrelation function acquisition comprises: Determining the number of sampling points of each data symbol in the measured data signal; Adopting an adaptive algorithm to determine the starting point of the cyclic autocorrelation function of each data symbol; The phase of the cyclic autocorrelation function of each data symbol is obtained by using FFT transformation; the phase includes jitter information of the measured data signal, and the coarse time interval error TIE jitter information is obtained; The phase of the cyclic autocorrelation function is specifically as follows: （1） In the formula, is the cyclic autocorrelation function of the measured signal at its symbol rate, is the phase of the cyclic autocorrelation function, is the phase offset of the channel, It is the accurate time interval error TIE jitter information; The coarse time interval error TIE jitter information is subjected to a de-accumulated error process to obtain accurate time interval error TIE jitter information.

2. The high-speed serial link jitter extraction method based on cyclostationary characteristics according to claim 1 is characterized in that: The starting point includes finding the first extreme point of the measured data signal by using a derivation method, which is the starting point.

3. The high-speed serial link jitter extraction method based on cyclostationary characteristics according to claim 1 is characterized in that: After the starting point is determined, the measured data signal is segmented according to the sampling multiple N to obtain multiple data symbols, and the cyclic autocorrelation function of each data symbol is calculated in turn.

4. The high-speed serial link jitter extraction method based on cyclostationary characteristics according to claim 1, characterized in that: The acquisition of the accurate time interval error (TIE) jitter information includes: The jitter information of the measured data signal is obtained by performing direct current removal processing on the phase information of the cyclic autocorrelation function and then removing the accumulated error jitter generated by the channel effect in the phase information.

5. The high-speed serial link jitter extraction method based on cyclostationary characteristics according to claim 4 is characterized in that: The phase The curve is fitted and the phase is reversed Compensation is performed to obtain accurate time interval error (TIE) jitter information. .

Images