Digital oscilloscope high-precision trigger correction method based on adaptive learning

Abstract

The invention discloses a digital oscilloscope high-precision trigger correction method based on adaptive learning, and the method comprises the steps: inputting an output signal of an arbitrary waveform generator into a digital oscilloscope, and completing the collection and caching of waveform data; and then setting a horizontal time base gear of the digital oscilloscope, starting trigger correction enabling of the digital oscilloscope, and correcting an unstable trigger waveform into a stable trigger waveform through self-adaptive high-precision trigger correction, so that the trigger precision of the collected trigger waveform is greatly improved.

Description

Technical Field

[0001] The invention belongs to the technical field of digital oscilloscopes, and more specifically, relates to a high-precision trigger correction method for a digital oscilloscope based on adaptive learning. Background Art

[0002] With the development and continuous application of technology in the field of electronic measurement, the requirements for distributed acquisition applications and transient testing have become increasingly stringent, and high-precision triggering has become a key indicator of electronic measuring instruments. For time domain test instruments represented by digital oscilloscopes, the trigger function has become an indispensable part in order to capture the waveforms that users are interested in and expect and to display the waveforms stably. The trigger function is a prominent feature of digital storage oscilloscopes (DSOs). As the sampling rate of digital storage oscilloscopes continues to increase, the demand for high-precision triggering has become increasingly urgent. In traditional digital oscilloscopes, waveform triggering uses hardware comparison signals to directly map to data, thereby obtaining a stable trigger waveform. This processing can be used on oscilloscopes with low sampling rates and low trigger accuracy requirements. However, in the application of high-performance oscilloscopes, direct mapping of trigger signals brings large trigger errors and trigger jitter problems, such as Figure 1 As shown, there are three trigger waveforms, namely, advanced trigger waveform, ideal trigger waveform and delayed trigger waveform;

[0003] On a high-sampling-rate oscilloscope, since the internal working clock of the FPGA is difficult to reach above GHz, and the sampling rate of the ADC is usually in the GSPS range, a common practice is to convert serial data into parallel data to ensure that the amount of data remains unchanged. At this time, the trigger signal will be mapped to the parallel paths at the same time, making it difficult to ensure correct triggering at a certain sampling point. The trigger jitter is related to the number of parallel paths of the data.

[0004] The disadvantages of traditional triggering methods mainly include the following aspects:

[0005] (1) The hardware circuit design is rigorous and sophisticated. Signal noise and trigger edge distortion can cause false triggering, resulting in the trigger waveform being a waveform that is not of interest.

[0006] (2) Parallel processing within the FPGA makes it difficult to determine the trigger point. When the oscilloscope processes the interpolation measurement gear, the trigger point jitter error will become larger and the trigger waveform will jitter significantly.

[0007] (3) The external trigger edge signal is asynchronous with the FPGA internal processing clock. During the internal processing of the FPGA, the setup time and hold time of different semiconductor devices may be different, resulting in incorrect trigger position during analysis.

[0008] (4) The minimum trigger accuracy is related to the ADC sampling rate. Under the real-time sampling mode, the trigger accuracy is the minimum sampling interval, which makes it difficult to achieve higher hardware trigger accuracy.

[0009] In summary, most of the current traditional triggering methods are difficult to meet the requirements of high sampling rate and high triggering accuracy. Especially in multi-channel oscilloscope triggering applications, the accuracy of the trigger position and trigger jitter cannot be guaranteed. Summary of the invention

[0010] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a high-precision trigger correction method for a digital oscilloscope based on adaptive learning to improve the accuracy of the trigger position and reduce the trigger jitter.

[0011] To achieve the above-mentioned object of the invention, the present invention provides a high-precision trigger correction method for a digital oscilloscope based on adaptive learning, characterized in that it comprises the following steps:

[0012] (1) Waveform data acquisition and caching;

[0013] The output signal of the arbitrary waveform generator is input to the digital oscilloscope, and the stable parallel acquisition waveform is obtained by sampling through the acquisition system in the digital oscilloscope, and the analog trigger signal is obtained by shaping through the comparator in the digital oscilloscope;

[0014] When the FPGA in the digital oscilloscope detects the arrival of the rising edge of the analog trigger signal, the FPGA controls the parallel acquisition of waveform data to be cached in the parallel FIFO until the parallel FIFO is full;

[0015] (2) Set the horizontal time base position of the digital oscilloscope and enable the trigger correction of the digital oscilloscope;

[0016] (2.1) Read parallel data from the parallel FIFO, and then perform parallel-to-serial conversion to obtain serial input data;

[0017] (2.2) Construct a two-dimensional waveform array through serial input data;

[0018] (2.3) Determine whether each row of waveform data in the two-dimensional waveform array is a single-cycle waveform. If not, proceed to step (2.4); otherwise, keep the two-dimensional waveform array unchanged, recorded as:

[0019]

[0020] (2.4) Extract the single-cycle waveform data of each row of waveform data in the two-dimensional waveform array X, and then construct a new two-dimensional waveform array

[0021] (2.5), construct the attribute matrix A;

[0022] (2.6) Obtain the actual trigger point;

[0023] In the attribute matrix A, the mean value x of the first column imean Use the majority method to find the majority x, and then map the position index value p in the corresponding row based on the majority x. iindex , remember p iindex The location is the actual trigger point and serves as the reference position index value for subsequent calibration;

[0024] (2.7) When the next sampling cycle arrives, the acquisition system in the digital oscilloscope calculates the position index value of the next trigger waveform according to steps (2.1) to (2.6) Then according to the reference position index value p iindex Correct the trigger waveform and display the waveform after the correction is completed;

[0025] Then, by analogy, in subsequent sampling cycles, the adaptive learning and correction process after the waveform is triggered is completed.

[0026] The object of the invention of the present invention is achieved in this way:

[0027] The invention discloses a high-precision trigger correction method for a digital oscilloscope based on adaptive learning. The output signal of an arbitrary waveform generator is input into a digital oscilloscope to complete waveform data acquisition and caching. Then, the horizontal time base gear of the digital oscilloscope is set, and the trigger correction enable of the digital oscilloscope is turned on. The unstable trigger waveform is corrected into a stable trigger waveform through adaptive high-precision trigger correction, thereby greatly improving the trigger accuracy of the acquired trigger waveform.

[0028] At the same time, the digital oscilloscope high-precision trigger correction method based on adaptive learning of the present invention also has the following beneficial effects:

[0029] (1) By adopting advanced adaptive learning mechanism and combining trigger correction technology, digital oscilloscopes can significantly improve trigger accuracy. Its actual performance not only exceeds the traditional boundaries, but also breaks through the physical limitations of ADC sampling accuracy, achieving more accurate signal capture capabilities.

[0030] (2) Implementing a high-precision trigger correction strategy for sampled data can significantly reduce the design threshold and challenges of hardware trigger circuits, simplify the hardware circuit structure, effectively save precious circuit layout area, and promote the miniaturization and cost optimization of equipment;

[0031] (3) By implementing a correction processing algorithm on the original sampling signal, the complexity of the trigger logic design inside the FPGA is simplified. This not only effectively avoids the logic processing problems that may be caused by the cross-clock domain between the external trigger signal and the sampled data, but also significantly optimizes the logic circuit architecture inside the FPGA, thereby improving the performance and stability of the overall system. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a waveform diagram of advanced triggering, ideal triggering and delayed triggering;

[0033] Figure 2 It is the relationship between the analog trigger position and the sampling point of a high sampling rate digital oscilloscope;

[0034] Figure 3 It is a flow chart of a high-precision trigger correction method for a digital oscilloscope based on adaptive learning of the present invention;

[0035] Figure 4 It is the uncorrected simulated waveform diagram;

[0036] Figure 5 It is the simulation waveform after adaptive learning trigger correction. DETAILED DESCRIPTION

[0037] The specific implementation of the present invention is described below 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 the detailed description of known functions and designs may dilute the main content of the present invention, these descriptions will be omitted here.

[0038] Example

[0039] In this embodiment, in order to better illustrate the technical solution of this scheme, the process of digital oscilloscope waveform from acquisition, triggering to display is briefly introduced. The digital oscilloscope system consists of a conditioning channel, an acquisition module, a signal processing module, a main control and display module. The digital oscilloscope acquisition framework is composed of an analog-to-digital converter (ADC) and a field programmable device (FPGA). Triggering is a prominent feature of a digital oscilloscope, and high-precision triggering is particularly important for fine acquisition.

[0040] like Figure 2 As shown in the figure, under the data clock, the relationship between the trigger point and the sampling point data is: one trigger moment corresponds to 8 sampling points. Since the trigger point and the sampling point in the sampling system are not in a one-to-one correspondence, the trigger jitter problem occurs, but its jitter range only exists in X(8)~X(16), so the actual trigger point is also in it.

[0041] The present invention solves the limitation of requiring additional hardware circuits to measure the triggering moment and the sampling clock interval, and directly obtains it from the analysis of the post-triggering data. Therefore, a high-precision triggering function can be realized with only a basic triggering structure and sampling data.

[0042] Next, we will describe in detail a high-precision trigger correction method for a digital oscilloscope based on adaptive learning according to the present invention. Figure 3 As shown, the specific steps include:

[0043] (1) Waveform data acquisition and caching;

[0044] In this embodiment, an arbitrary waveform generator or a radio frequency signal source is connected to a digital oscilloscope, and the arbitrary waveform generator or the radio frequency signal source generates an output signal that satisfies the sampling law (the sampling frequency is more than twice the sampled signal), and the output signal is input to the digital oscilloscope, and a stable parallel acquisition waveform is obtained by sampling through the acquisition system in the digital oscilloscope, and an analog trigger signal is obtained by shaping through the comparator in the digital oscilloscope;

[0045] According to the acquisition trigger principle of the digital oscilloscope, when the FPGA in the digital oscilloscope detects the arrival of the rising edge of the analog trigger signal, the FPGA controls the parallel acquisition of waveform data to be cached in 8 parallel FIFOs, each with a depth of 16k, until the parallel FIFO is full;

[0046] (2) Set the horizontal time base position of the digital oscilloscope and enable the trigger correction of the digital oscilloscope;

[0047] Digital oscilloscopes generally have horizontal time base gears such as interpolation, real-time and decimation. At a large decimation gear, the acquired waveform has completed stable triggering under low sampling rates, and the waveform display is also very stable, so there is no need for trigger correction.

[0048] In this embodiment, the present invention provides an enabling interface, which can automatically determine and perform enabling operations according to the horizontal time base gear set by the user, that is, to start the trigger correction function. The specific correction process is as follows:

[0049] (2.1) Read parallel data from the parallel FIFO, and then perform parallel-to-serial conversion to obtain serial input data;

[0050] (2.2) Construct a two-dimensional waveform array through serial input data;

[0051] Assume that the serial input data contains 1000 waveforms, and each waveform contains 1024 sampling points. Then each row of the two-dimensional waveform array represents a waveform, and the elements in each row are the number of sampling points of each waveform, thus obtaining a two-dimensional waveform array of size M×N=1000×1024;

[0052] (2.3) Determine whether each row of waveform data in the two-dimensional waveform array is a single-cycle waveform. If not, proceed to step (2.4); otherwise, keep the two-dimensional waveform array unchanged, recorded as:

[0053]

[0054] (2.4) Extract the single-cycle waveform data of each row of waveform data in the two-dimensional waveform array X, and then construct a new two-dimensional waveform array

[0055] Based on the data in the first row, find the edge of the waveform, extract the first row of waveform data in the two-dimensional waveform array X, then find the starting point and ending point of any cycle in the first row of waveform data, and then extract the elements between the starting point and the ending point, recorded as n is the number of elements between the starting point and the ending point;

[0056] By element Based on the first row in the two-dimensional waveform array X, take n element points at the corresponding positions in the remaining M-1 rows of waveform data in the two-dimensional waveform array X to construct a new two-dimensional waveform array.

[0057]

[0058] (2.5), construct the attribute matrix A;

[0059] For the convenience of description, let the two-dimensional waveform array X and The number of elements in each row is k, and then the average value x of each row of elements in the two-dimensional waveform array is calculated. imean , i=1,2,…,M;

[0060]

[0061] Among them, x ij Represents the element corresponding to the i-th row and j-th column in the two-dimensional waveform array, j = 1, 2, ..., k;

[0062] Find the position index value p corresponding to the element point closest to the average value in each row iindex

[0063]

[0064] Take the average value x of each row imean and position index value p iindex Construct attribute matrix A;

[0065]

[0066] (2.6) Obtain the actual trigger point;

[0067] In the attribute matrix A, the mean value x of the first column imean Use the majority method to find the majority x, and then map the position index value p in the corresponding row based on the majority x. iindex , remember p iindex The location is the actual trigger point and serves as the reference position index value for subsequent calibration;

[0068] (2.7) When the next sampling cycle arrives, the acquisition system in the digital oscilloscope calculates the position index value of the next trigger waveform according to steps (2.1) to (2.6) Then according to the reference position index value p iindex To correct the trigger waveform, When the trigger waveform needs to be moved forward when When the trigger waveform needs to be moved backward

[0069] After the correction is completed, the waveform is displayed, and then by analogy, in the subsequent sampling period, the adaptive learning correction process after the waveform is triggered is completed.

[0070] Experimental example

[0071] According to the above principle, the arbitrary waveform generator is set to output a 1kHz square wave signal, and the output signal is connected to the input channel of the digital oscilloscope. First, the high-precision trigger correction function of the invention is not enabled, and the digital oscilloscope is set to a suitable time base position. The oscilloscope trigger mode selects rising edge triggering, and is adjusted to display only a rising edge waveform on the digital oscilloscope interface. The afterglow mode of the digital oscilloscope is turned on, and the waveform display is observed, such as Figure 4 Then enable the high-precision trigger correction function of the invention, set the digital oscilloscope to the appropriate time base gear, select the rising edge trigger for the oscilloscope trigger mode, adjust it to the digital oscilloscope interface to display only one rising edge waveform, turn on the digital oscilloscope afterglow mode, and observe the waveform display, as shown in the figure below: Figure 5 As shown. Figure 4 and Figure 5 Comparative analysis shows that the triggering accuracy of the digital oscilloscope can be greatly improved by adaptively learning the triggering of the corrected analog waveform in the present invention.

[0072] Although the above describes the illustrative specific embodiments of the present invention to facilitate those skilled in the art to understand the present invention, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations using the concept of the present invention are protected.

Claims

1. A high-precision trigger correction method for digital oscilloscope based on adaptive learning, It is characterized in that The following steps are involved: (1) Waveform data acquisition and caching; The output signal of the arbitrary waveform generator is input to the digital oscilloscope, and the stable parallel acquisition waveform is obtained by sampling through the acquisition system in the digital oscilloscope, and the analog trigger signal is obtained by shaping through the comparator in the digital oscilloscope; When the FPGA in the digital oscilloscope detects the arrival of the rising edge of the analog trigger signal, the FPGA controls the parallel acquisition of waveform data to be cached in the parallel FIFO until the parallel FIFO is full; (2) Set the horizontal time base position of the digital oscilloscope and enable the trigger correction of the digital oscilloscope; (2.1) Read parallel data from the parallel FIFO, and then perform parallel-to-serial conversion to obtain serial input data; (2.2) Construct a two-dimensional waveform array through serial input data; (2.3) Determine whether each row of waveform data in the two-dimensional waveform array is a single-cycle waveform. If not, proceed to step (2.4); otherwise, keep the two-dimensional waveform array unchanged, recorded as: (2.4) Extract the single-cycle waveform data of each row of waveform data in the two-dimensional waveform array X, and then construct a new two-dimensional waveform array (2.5), construct the attribute matrix A; (2.6) Obtain the actual trigger point; In the attribute matrix A, the mean value x of the first column imean Use the majority method to find the majority x, and then map the position index value p in the corresponding row based on the majority x. iindex , remember p iindex The location is the actual trigger point and serves as the reference position index value for subsequent calibration; (2.7) When the next sampling cycle arrives, the acquisition system in the digital oscilloscope calculates the position index value of the next trigger waveform according to steps (2.1) to (2.6) Then according to the reference position index value p iindex Correct the trigger waveform and display the waveform after the correction is completed; Then, by analogy, in subsequent sampling cycles, the adaptive learning and correction process after the waveform is triggered is completed.

2. The high-precision trigger correction method for a digital oscilloscope based on adaptive learning according to claim 1, It is characterized in that The construction process of the two-dimensional waveform array is: Assume that the serial input data contains M waveforms, each waveform contains N sampling points, then each row of the two-dimensional waveform array represents a waveform, and the elements in each row are the number of sampling points of each waveform, thus obtaining a two-dimensional waveform array of size M×N.

3. The high-precision trigger correction method for a digital oscilloscope based on adaptive learning according to claim 1, It is characterized in that In step (2.4), a new two-dimensional waveform array is constructed The method is: Extract the first row of waveform data from the two-dimensional waveform array X, then find the starting point and ending point of any cycle in the first row of waveform data, and then extract the elements between the starting point and the ending point, recorded as n is the number of elements between the starting point and the ending point; By element Based on the first row in the two-dimensional waveform array X, take n element points at the corresponding positions in the remaining M-1 rows of waveform data in the two-dimensional waveform array X to construct a new two-dimensional waveform array.

4. The high-precision trigger correction method for a digital oscilloscope based on adaptive learning according to claim 1, It is characterized in that The method for constructing the attribute matrix in step (2.5) is: For the convenience of description, let the two-dimensional waveform array X and The number of elements in each row is k, and then the average value x of each row of elements in the two-dimensional waveform array is calculated. imean , i=1,2,…,M; Among them, x ij Represents the element corresponding to the i-th row and j-th column in the two-dimensional waveform array, j = 1, 2, ..., k; Find the position index value p corresponding to the element point closest to the average value in each row iindex Take the average value x of each row imean and position index value p iindex Construct attribute matrix A; 5. The high-precision trigger correction method for a digital oscilloscope based on adaptive learning according to claim 1, It is characterized in that In the step (2.7), according to the reference position index value p iindex The method to correct the trigger waveform is: when When the trigger waveform needs to be moved forward when When the trigger waveform needs to be moved backward

Images