Jitter measurement method combining time domain, frequency domain and statistical domain in digital oscilloscope

Abstract

The present invention discloses a jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope. First, the digital oscilloscope collects a long series of continuous signals and forwards the sampled data to a host computer. Then, each jump moment of the signal to be measured is found. Then, a software phase-locked loop is used to recover the reference edge moment according to the jump moment of the signal to be measured. Then, the jump moment of the signal to be measured is subtracted from the reference jump moment to obtain a set of time interval error sequences. Finally, based on this sequence, the jitter is analyzed and processed from multiple information domains to obtain the magnitude of the overall jitter, deterministic jitter, random jitter, periodic jitter, data correlation jitter, intersymbol interference and duty cycle distortion.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital oscilloscopes, and more particularly, relates to a jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope. Background Art

[0002] As an instrument for electronic measurement and signal analysis, the digital oscilloscope represents a modern test instrument and is the most widely used and most widely used general-purpose test instrument in the electronic information field. From industrial production and teaching and research to national defense construction and equipment development, digital oscilloscopes are indispensable. As electronic signals become increasingly complex, the functions of digital oscilloscopes are also becoming more diverse. Jitter analysis, a key function of high-end digital oscilloscopes, can measure and analyze jitter in signals, thereby evaluating system performance and stability.

[0003] Jitter is one of the main factors affecting data transmission on high-speed serial links. Efficient and accurate jitter analysis, tracing its root causes, and implementing measures to mitigate its impact are crucial. The impact of different modules on the transmitted signal in a high-speed serial link can be characterized by different jitter components, thus facilitating diagnosis and debugging. Different jitter components and their characteristics in the time, frequency, and statistical domains provide information for analyzing the jitter components and their magnitudes during the jitter process.

[0004] Current jitter standards describe jitter phenomena based on jitter component analysis, and propose primary and secondary jitter analysis. Primary jitter analysis examines the system's bit error performance to determine whether it meets bit error rate requirements or conforms to certain specifications. Secondary analysis isolates individual jitter components from the total jitter to examine system characteristics and diagnose potential system issues. Summary of the Invention

[0005] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a jitter measurement method in a digital oscilloscope that combines the time domain, frequency domain, and statistical domain. The jitter is analyzed from multiple information domains to obtain the results of each jitter component, thereby realizing the separation of jitter components.

[0006] To achieve the above-mentioned object of the invention, the present invention provides a jitter measurement method in a digital oscilloscope that combines time domain, frequency domain, and statistical domain, characterized by comprising the following steps:

[0007] (1) Acquire the signal to be measured through the acquisition system of the digital oscilloscope to obtain continuous sampling data containing multiple edges of the signal to be measured;

[0008] (2) The FPGA of the digital oscilloscope forwards the sampled data to the host computer, and then performs amplitude histogram statistics on the sampled data to obtain the high and low levels of the signal to be measured;

[0009] (3) Calculate the hysteresis range and crossover level based on the high and low levels of the signal to be measured;

[0010] (4) Traverse all sampling points and find the sampling moments corresponding to the cross-level, which are the transition moments of the signal to be measured, and then store these transition moments in an array;

[0011] (5) Use the software phase-locked loop in the host computer to recover the reference jump time of the reference signal, and then subtract it from the corresponding jump time in the array to obtain a set of time interval error sequences;

[0012] (6) Analyze the time interval error sequence in the time domain, frequency domain, and statistical domain to obtain the jitter measurement value.

[0013] The object of the invention of the present invention is achieved like this:

[0014] The present invention discloses a jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope. First, the digital oscilloscope collects a long series of continuous signals and forwards the sampled data to a host computer. Then, each transition moment of the signal to be measured is found. Then, a software phase-locked loop is used to recover the reference edge moment based on the transition moment of the signal to be measured. Then, the transition moment of the signal to be measured is subtracted from the reference transition moment to obtain a set of time interval error sequences. Finally, based on the sequence, the jitter is analyzed and processed from multiple information domains to obtain the magnitude of the overall jitter, deterministic jitter, random jitter, periodic jitter, data correlation jitter, intersymbol interference and duty cycle distortion.

[0015] At the same time, the jitter measurement method combining time domain, frequency domain and statistical domain in the digital oscilloscope of the present invention also has the following beneficial effects:

[0016] (1) The clock recovery in the present invention adopts a software phase-locked loop with flexible and configurable parameters, and takes the signal jump time as input, which greatly reduces the amount of data required for the phase-locked loop calculation process, alleviates the storage burden and calculation pressure of the digital oscilloscope system, and improves the analysis efficiency.

[0017] (2) The present invention analyzes jitter from three information domains: time domain, frequency domain, and statistical domain, avoiding the problem of inaccurate and incomplete analysis in a single information domain, and realizing primary and secondary analysis of jitter, thereby achieving the purpose of comprehensive jitter analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flow chart of the jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope of the present invention;

[0019] Figure 2 It is the collected signal samples and statistical histogram;

[0020] Figure 3 Schematic diagram of signal edge detection;

[0021] Figure 4 It is the basic structure of software phase-locked loop;

[0022] Figure 5 is the statistical histogram of jitter;

[0023] Figure 6 is the bit error rate cumulative distribution function of jitter;

[0024] Figure 7 is the bathtub curve;

[0025] Figure 8 is the Q factor curve;

[0026] Figure 9 is the amplitude spectrum of the jitter; DETAILED DESCRIPTION

[0027] The following describes the specific embodiments of the present invention in conjunction with the accompanying drawings so that those skilled in the art can better understand the present invention. It should be noted that in the following description, when detailed descriptions of known functions and designs may dilute the main content of the present invention, such descriptions will be omitted here.

[0028] Example

[0029] In this embodiment, if Figure 1 As shown, the jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope of the present invention includes the following steps:

[0030] (1) Acquire the signal to be measured through the acquisition system of the digital oscilloscope to obtain continuous sampling data containing at least 100 edges of the signal to be measured;

[0031] The signal to be measured is a clock signal or a data signal. If it is a data signal, it must have a repeating data pattern.

[0032] (2) The FPGA of the digital oscilloscope forwards the sampled data to the host computer, and then performs amplitude histogram statistics on the sampled data. Generally speaking, the clock signal or non-return-to-zero code data signal has two peaks, such as Figure 2 As shown, we can obtain the high and low levels of the signal to be measured;

[0033] (3) Calculate the hysteresis range and crossover level based on the high and low levels of the signal to be measured;

[0034] The hysteresis range and crossing level are calculated as:

[0035]

[0036]

[0037] Among them, V top and V base Respectively represent the high and low level values ​​of the signal to be measured, V crossing Indicates the crossover level, V hystersis_high 、V hystersis_low Respectively represent the upper and lower limits of the hysteresis range.

[0038] (4) Traverse all sampling points and find the sampling moments corresponding to the cross-level, which are the transition moments of the signal to be measured, and then store these transition moments in an array;

[0039] In this embodiment, the specific method for obtaining each transition moment of the signal to be tested is:

[0040] (4.1) Traverse all sampling points and find all continuous sampling point segments that cross the hysteresis range, which are regarded as valid signal edges. If there is a sampling point whose level value is equal to the cross level value in these segments, the sampling moment corresponding to the sampling point is the signal jump moment. Otherwise, go to step (4.2);

[0041] (4.2) Find the two sampling points above and below the crossing level in the effective signal edge and perform linear interpolation. The number of the actual sampling point is S i , i=1,2,…;…

[0042] Calculate the signal transition time T(E n ):

[0043]

[0044] V(E n )=V crossing

[0045] T s =T(S i+1 )-T(S i )

[0046] Among them, V(S i ) represents the sampling point S i The level value, T(S i ) represents the sampling point S i The corresponding sampling time, V(E n ) represents the trip point E n The level value, T(E n ) represents the trip point E n The jump time, n = 1, 2, ... represents the jump point number, T s is the sampling time interval between two adjacent sampling points.

[0047] In this embodiment, since the signal to be measured is collected by the ADC, the sampled data is the voltage value of a discrete point, so there may not be an actual sampling point at the cross level. Figure 3 As shown in the figure, the black points S1, S2, ... are the actual sampling points, and the white points E1, E2, ... are the points where the cross level intersects the signal edge, that is, the points at the time of signal jump. 14 and S 15 、S 30 and S 31 ) to perform linear interpolation to obtain the time information T(E1) and T(E2) of E1 and E2.

[0048] (5) Use the software phase-locked loop in the host computer to recover the reference jump time of the reference signal, and then subtract it from the corresponding jump time in the array to obtain a set of time interval error sequences; in this embodiment, the structure of the software phase-locked loop is as follows: Figure 4 As shown;

[0049] In this embodiment, the specific method for obtaining each reference transition time is:

[0050] (5.1), let each reference jump time be T(E' n ), n=1,2,…, let T(E′1)=T(E1);

[0051] (5.2) The transition time of the signal to be measured is T(E n ) and the reference transition time T(E') of the reference signal n ) are the input and output of the software phase-locked loop, and T(E n ) and T(E' n ) is used as the input of the phase detector in the software phase-locked loop, and the output of the phase detector is then low-pass filtered through the loop filter;

[0052] The filtering result is compared with the reference jump time T(E' n ) and the duration of one symbol of the signal to be measured are added to obtain the next reference jump time T(E' n+1 );

[0053] (5.3), the reference jump time T(E' n+1 ) is fed back to the input terminal and the jump time T(E n+1 ) and then make a difference, repeat the above operation for the difference value to get the next reference jump time T(E' n+2 ), and so on, to obtain the reference transition moments that are consistent with the number of transition moments of the signal to be measured.

[0054] (6) Analyze the time interval error sequence in the time domain, frequency domain, and statistical domain to obtain the jitter measurement value;

[0055] (6.1), let the time interval error sequence obtained in step (5) be T;

[0056] (6.2), make a statistical histogram of the time interval error sequence T, as follows Figure 5 As shown;

[0057] (6.3), integrate the statistical histogram from both sides to the middle, and we get Figure 6 In the BER cumulative distribution function shown, the distance between the two points where a line parallel to the horizontal axis intersects the curve represents the overall jitter at a specific BER. Because the number of time interval error samples required for statistical processing is very large, directly obtaining them would be time-consuming, so an extrapolation method was used. The solid line represents the actual sample data, and the dashed line represents the extrapolated sample data.

[0058] (6.4), move the left and right branches of the bit error rate integral graph to the opposite sides respectively, and we get Figure 7 The bathtub curve shown;

[0059] (6.5), the bathtub curve is converted into a Q-factor curve with a linear tail through the Q function, such as Figure 8 As shown;

[0060] (6.6) Perform a linear fit on the tail of the Q-factor curve to obtain a fitted straight line. Then, calculate the deterministic jitter DJ and random jitter RJ based on the intercept and slope of the fitted line, and further calculate the total jitter TJ.

[0061] The calculation method of deterministic jitter DJ is:

[0062] DJ=μ L +μ R

[0063] Among them, μ L and μ R are the intercepts of the two straight lines fitted at the tail of the Q-factor curve;

[0064] The calculation method of random jitter RJ is:

[0065]

[0066] Wherein, slope1 and slope2 are the slopes of the two straight lines fitted at the tail of the Q-factor curve;

[0067] The total jitter TJ is calculated as:

[0068] TJ=DJ+14×RJ

[0069] (6.7) Determine the type of the signal to be measured. If the signal to be measured is a data signal, proceed to step (6.8); if the signal to be measured is a clock signal, proceed to step (6.9);

[0070] (6.8) If the signal to be measured is a data signal, the time interval error sequence T is averaged in the time domain according to the pattern length, and then the time interval error sequences before and after averaging are subtracted to remove random jitter and periodic jitter that are not related to the data pattern, and a set of time interval error sequences containing only data-related jitter is obtained. Among them, the time interval error sequence corresponding to the rising edge is recorded as The time interval error sequence corresponding to the falling edge is recorded as

[0071] Then, data correlation jitter DDJ, intersymbol interference ISI, and duty cycle distortion DCD are obtained based on these three time interval error sequences;

[0072] The calculation method of data dependency jitter DDJ is:

[0073]

[0074] The calculation method of inter-symbol interference ISI is:

[0075]

[0076] The duty cycle distortion is calculated as:

[0077]

[0078] Among them, max(x) means taking the maximum value in the sequence x, min(x) means taking the minimum value in the sequence x, max[a,b] means taking the larger value of values ​​a and b; mean(x) means calculating the mean of the sequence x.

[0079] Finally, the time interval error sequence T is subtracted from the time interval error sequence Get a set of time interval error sequences that do not contain data correlation jitter Then proceed to step (6.4);

[0080] (6.4), perform fast Fourier transform on the time interval error sequence to obtain the amplitude spectrum of the time interval error, as Figure 9 As shown;

[0081] Set a decision threshold, and record the amplitude above the decision threshold as periodic jitter, otherwise it is recorded as random jitter;

[0082] Finally, the periodic jitter and random jitter are subjected to inverse Fourier transform respectively to obtain the time interval error sequence T containing only the periodic jitter. PJ and the time error interval sequence T containing only random jitter RJ , and express periodic jitter and random jitter in peak-to-peak and RMS values ​​respectively.

[0083] The calculation method of periodic jitter PJ is:

[0084] PJ=max(T PJ )-min(T PJ )

[0085] Among them, max(T PJ )、min(T PJ ) are the time interval error sequences T containing only periodic jitter PJ The maximum and minimum values ​​of ;

[0086] The calculation method of random jitter RJ is:

[0087] RJ=RMS(T RJ )

[0088] Among them, RMS(T RJ ) represents the time interval error sequence T containing only random jitter. RJ The root mean square value of .

[0089] Although the above describes the illustrative specific embodiments of the present invention to facilitate understanding of the present invention by those skilled in the art, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations using the concepts of the present invention are protected.

Claims

1. A jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope, characterized in that: The following steps are involved: (1) Acquire the signal to be measured through the acquisition system of the digital oscilloscope to obtain continuous sampling data containing multiple edges of the signal to be measured; (2) The FPGA of the digital oscilloscope forwards the sampled data to the host computer, and then performs amplitude histogram statistics on the sampled data to obtain the high and low level of the signal to be measured; (3) Calculate the hysteresis range and crossover level based on the high and low levels of the signal to be measured; (4) Traverse all sampling points and find the sampling moments corresponding to the cross-level, which are the transition moments of the signal to be measured, and then store these transition moments in an array; (5) Use the software phase-locked loop in the host computer to recover the reference jump time of the reference signal, and then subtract it from the corresponding jump time in the array to obtain a set of time interval error sequences; The method for obtaining each reference jump time is as follows: (5.1), let each reference jump time be T(E' n ), n=1,2,..., let T(E1')=T(E1); (5.2) The transition time of the signal to be measured is T(E n ) and the reference transition time T(E') of the reference signal n ) are the input and output of the software phase-locked loop, and T(E n ) and T(E' n ) is used as the input of the phase detector in the software phase-locked loop, and the output of the phase detector is then low-pass filtered through the loop filter; The filtering result is compared with the reference jump time T(E' n ) and the duration of one symbol of the signal to be measured are added to obtain the next reference jump time T(E' n+1 ); (5.3), the reference jump time T(E' n+1 ) is fed back to the input terminal and the jump time T(E n+1 ) and then make a difference, repeat the above operation for the difference value to get the next reference jump time T(E' n+2 ), and so on, to obtain the reference transition time that is consistent with the number of transition times of the signal to be measured; (6) Analyze the time interval error sequence in the time domain, frequency domain, and statistical domain to obtain the jitter measurement value.

2. The jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope according to claim 1, characterized in that: The signal to be measured is a clock signal or a data signal. If it is a data signal, it must have a repeating data pattern.

3. The jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope according to claim 1, characterized in that: The calculation method of the hysteresis range and the crossover level is: Among them, V top and V base Respectively represent the high and low level values ​​of the signal to be measured, V crossing Indicates the crossover level, V hystersis_high 、V hystersis_low Respectively represent the upper and lower limits of the hysteresis range.

4. The jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope according to claim 1, characterized in that: The method for obtaining each transition moment of the signal to be tested in step (4) is: (4.1) Traverse all sampling points and find all continuous sampling point segments that cross the hysteresis range, which are regarded as valid signal edges. If there is a sampling point whose level value is equal to the cross level value in these segments, the sampling moment corresponding to the sampling point is the signal jump moment. Otherwise, go to step (4.2); (4.2) Find two sampling points above and below the intersection of the effective signal edge and perform linear interpolation. The sampling point number after linear interpolation is S. i , i=1,2,…; Calculate the signal transition time T(E n ): V(E n )=V crossing T s =T(S i+1 )-T(S i ) Among them, V(S i ) represents the sampling point S i The level value, T(S i ) represents the sampling point S i The corresponding sampling time, V(E n ) represents the trip point E n The level value, T(E n ) represents the trip point E n The jump time, n = 1, 2, ... represents the jump point number, T s is the sampling time interval between two adjacent sampling points.

5. The jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope according to claim 1, characterized in that: The method for obtaining the jitter measurement value in step (6) is: (6.1), let the time interval error sequence obtained in step (5) be T; (6.2) Make a statistical histogram of the time interval error sequence T; (6.3) Integrate the statistical histogram from both sides to the middle to obtain the bit error rate integral graph; (6.4), move the left and right branches of the bit error rate integral graph to opposite sides to obtain the bathtub curve; (6.5), convert the bathtub curve into a Q-factor curve with a linear tail through the Q function; (6.6) Perform a linear fit on the tail of the Q-factor curve to obtain a fitted straight line. Then, calculate the deterministic jitter DJ and random jitter RJ based on the intercept and slope of the fitted line, and further calculate the total jitter TJ. (6.7) Determine the type of the signal to be measured. If the signal to be measured is a data signal, proceed to step (6.8); if the signal to be measured is a clock signal, proceed to step (6.9); (6.8) If the signal to be measured is a data signal, the time interval error sequence T is averaged in the time domain according to the pattern length, and then the time interval error sequences before and after averaging are subtracted to remove random jitter and periodic jitter that are not related to the data pattern, and a set of time interval error sequences containing only data-related jitter is obtained. Among them, the time interval error sequence corresponding to the rising edge is recorded as The time interval error sequence corresponding to the falling edge is recorded as Then, data correlation jitter DDJ, intersymbol interference ISI, and duty cycle distortion DCD are obtained based on these three time interval error sequences; Finally, the time interval error sequence T is subtracted from the time interval error sequence Get a set of time interval error sequences that do not contain data correlation jitter Then proceed to step (6.9); (6.9) Perform fast Fourier transform on the time interval error sequence to obtain the amplitude spectrum of the time interval error; Set a decision threshold, and record the amplitude above the decision threshold as periodic jitter, otherwise it is recorded as random jitter; Finally, the periodic jitter and random jitter are subjected to inverse Fourier transform respectively to obtain the time interval error sequence T containing only the periodic jitter. PJ and the time error interval sequence T containing only random jitter RJ , and express periodic jitter and random jitter in terms of peak-to-peak value and root mean square value respectively.

6. The jitter measurement method combining time domain, frequency domain and statistical domain in a digital oscilloscope according to claim 5, characterized in that: The calculation method of the deterministic jitter DJ is: DJ=μ L +m R Among them, μ L and μ R are the intercepts of the two straight lines fitted at the tail of the Q-factor curve; The random jitter RJ is calculated as follows: Wherein, slope1 and slope2 are the slopes of the two straight lines fitted at the tail of the Q-factor curve; The total jitter TJ is calculated as follows: TJ=DJ+14×RJ The calculation method of the periodic jitter PJ is: PJ=max(T PJ )-min(T PJ ) Among them, max(T PJ )、min(T PJ ) are the time interval error sequences T containing periodic jitter PJ The maximum and minimum values ​​of ; The calculation method of the data dependency jitter DDJ is: in, are the time interval error sequences containing data correlation jitter The maximum and minimum values ​​of ; The calculation method of the inter-symbol interference ISI is: in, and They are The time interval error sequence corresponding to the rising and falling edges, max(x) means taking the maximum value in the sequence x, min(x) means taking the minimum value in the sequence x, and max[a,b] means taking the larger value of a and b; The calculation method of the duty cycle distortion is: Here, mean(x) means calculating the mean of the sequence x.

Images