Data analysis method for phase change along with frequency of encoder operational amplifier circuit

Abstract

The invention belongs to the technical field of servo motor drivers, and particularly relates to a data analysis method for phase change along with frequency of an encoder operational amplifier circuit, which comprises the following steps: time domain delay between an input signal and an output signal of the operational amplifier circuit is acquired, and the input signal is generated by a sine wave signal source with adjustable frequency; calculating a phase delay angle under each frequency according to the time domain delay; drawing a phase-frequency curve of the phase delay angle and the frequency; segmenting the phase-frequency curve by taking a transition point of a local slope of the phase-frequency curve as a segmented trigger point; and performing linear fitting on each segment by adopting a least square method to obtain a phase compensation model of each segment, and compensating the phase delay angle according to the phase compensation model. According to the method, servo motor instruction identification errors caused by phase asynchronization are eliminated.

Description

Technical Field

[0001] The present application belongs to the technical field of servo motor drivers, and specifically relates to a data analysis method for phase changes of encoder operational amplifier circuits with frequency. Background Art

[0002] Servo motor drives play a crucial role in modern industrial automation and precision control systems, with applications spanning high-end manufacturing equipment such as injection molding machines, textile machinery, packaging machinery, and CNC machine tools. Servo motor drives ensure operational accuracy and production efficiency throughout the entire system by providing precise closed-loop speed and position control of the motor.

[0003] A typical servo motor drive system typically consists of two main components: a control board and a power board. The control board is responsible for executing the core control algorithm and generating the command signals required for motor operation; the power board drives the motor according to these commands. In the signal processing chain, encoder signals are key to achieving precise closed-loop control. However, before entering the main controller, these encoder signals typically need to pass through a conditioning circuit consisting of an operational amplifier (Op-Amp).

[0004] When encoder signals pass through op amp circuits, they introduce a phase delay that varies with the signal's frequency. If multiple signals from the same encoder (e.g., A / B / Z phase signals) exhibit inconsistent phase delays after processing, the encoder signals will become asynchronous. This asynchrony can cause the controller to misjudge the servo motor's position or speed, leading to command recognition errors. In severe cases, this can cause system oscillation, loss of control, or shutdown, posing a significant threat to equipment safety and product quality. Therefore, this application proposes a strategy that accurately identifies and compensates for the frequency-varying phase delay of encoder signals in op amp circuits to ensure signal synchronization and system stability. Summary of the Invention

[0005] The embodiment of the present application provides a data analysis method for the phase variation of an encoder op amp circuit with frequency, which solves the problem of the encoder signal phase delay varying with frequency caused by the characteristics of the op amp circuit in the existing servo motor driver.

[0006] This application is implemented in this way. A method for analyzing data of phase variation with frequency of an encoder operational amplifier circuit, comprising: Obtaining a time domain delay between an input signal and an output signal of an operational amplifier circuit, wherein the input signal is generated by a frequency-adjustable sinusoidal wave signal source; Calculate the phase delay angle at each frequency based on the time domain delay; Plot the phase-frequency curve of phase delay angle and frequency; The phase-frequency curve is segmented by taking the transition point of the local slope of the phase-frequency curve as the trigger point of segmentation; The least square method is used to perform linear fitting on each segment to obtain a phase compensation model for each segment, and the phase delay angle is compensated according to the phase compensation model.

[0007] Furthermore, obtaining the time domain delay between the input signal and the output signal of the operational amplifier circuit includes passing the input signal and the output signal of the operational amplifier circuit through a comparator to obtain the output signal of the comparator.

[0008] Furthermore, the phase delay angle at each frequency is calculated based on the time domain delay, including: Calculate the signal period based on the frequency; Calculate the proportion of time domain delay in the signal period to obtain the period proportion; Convert the period ratio into angle system to get the phase delay angle at the corresponding frequency.

[0009] Furthermore, the phase-frequency curve is segmented by taking the transition point of the local slope of the phase-frequency curve as the trigger point of segmentation, including: Calculate the rate of change of the phase delay angle with frequency to obtain the local slope; Determine whether the local slope change exceeds the set threshold or there is a sign change; When a set threshold is exceeded or a sign change occurs, the phase-frequency curve is segmented using the transition point of the local slope as the segmentation trigger point.

[0010] Furthermore, the phase compensation model of each segment is expressed as: , For the The phase after segment compensation, and For the The compensation coefficient of the segment, For the frequencies, and By minimizing the The sum of squares of the residuals for all data points in the segment Obtain: , The smoothed data points The phase delay angle.

[0011] Furthermore, the method further includes: performing outlier interpolation and data smoothing filtering on the time domain delay.

[0012] Compared with the prior art, the present application has the following advantages: This application uses segmented fitting of slope characteristics to improve the average accuracy of the phase compensation model from approximately 70% for a traditional single model to 95.92%. This high accuracy directly translates into precise perception of the servo motor's position and speed, fundamentally eliminating servo motor command recognition errors caused by phase asynchrony.

[0013] Precise phase synchronization ensures stable operation of the servo drive even under high-speed and high-acceleration conditions. This reduces system overshoot and oscillation, shortens settling time, and improves machining accuracy and production efficiency. It also avoids serious failures that could result from signal errors. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 Flowchart of a method for analyzing data of phase variation with frequency of an encoder operational amplifier circuit provided in an embodiment of the present application; Figure 2 A comparison chart of the effects of using single-segment fitting in the prior art and segmented fitting using the method of the present application is provided in the embodiments of the present application. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0016] See also Figure 1 As shown, the present application proposes a data analysis method for the phase variation of an encoder operational amplifier circuit with frequency, comprising: S1 measures the time domain delay between the input and output signals of the op amp circuit. The input signal is generated by a frequency-adjustable sinusoidal signal source. This frequency-adjustable sinusoidal signal source can be used to generate a frequency-adjustable 1Vpp sinusoidal wave as a simulated encoder signal source. The generated sinusoidal signal is input to the op amp circuit's conditioning circuit under test, which includes circuits that amplify, filter, and convert the input signal. The frequency of the input signal can be synchronously measured using an oscilloscope or data acquisition card. After the op amp circuit's input and output signals pass through a comparator, the time domain delay between the two signals is determined. The unit of time domain delay is microseconds. Time domain delay refers to the time lag of the output signal relative to the input signal.

[0017] The frequency covers a wide frequency range from 1 kHz to 2010 kHz, and the time domain delay corresponding to different frequencies is recorded.

[0018] S2 calculates the phase delay angle at each frequency based on the time domain delay; The original measurement is time domain delay. The time domain delay is converted into a phase angle that is independent of the period and is used to analyze the phase characteristics when the frequency changes.

[0019] S3 plots a phase-frequency curve of phase delay angle and frequency; S4 segments the phase-frequency curve by using the transition point of the local slope of the phase-frequency curve as the trigger point of segmentation; S5 performs linear fitting on each segment using the least squares method to obtain a phase compensation model for each segment, and compensates for the phase delay angle according to the phase compensation model.

[0020] Because a single linear model is used to fit phase delay data across the entire frequency band, its accuracy is low (approximately 70%) and cannot meet high-precision control requirements. The fundamental reason is that the frequency characteristics of the op amp circuit (such as the gain-bandwidth product limitation) result in a non-globally linear relationship between phase delay and frequency.

[0021] In the embodiment of the present application, the segmentation points are automatically determined based on the local slope change characteristics of the phase-frequency curve. First, at a frequency, the frequency is selected as the starting frequency, and one or more frequencies behind the frequency are used as the ending frequency. The ratio of the difference between the phase delay angles corresponding to the two frequencies and the difference between the two frequencies is used as the local slope. Multiple local slopes can be obtained by analogy. It is necessary to find the transition point where the local slope suddenly changes. For example, the first local slope is 1, the second local slope is 1, the third local slope is 3, and the fourth local slope is 1. Then the transition point corresponds to the third local slope. The frequency range before and after the starting frequency of the transition point is segmented based on the boundary. The transition point is the trigger point for the start of frequency segmentation.

[0022] A least-squares linear fit is performed on each segment to generate a phase compensation model for each segment. This results in multiple phase compensation models, each corresponding to a frequency segment. In servo drive control, for any operating frequency, the frequency segment to which it belongs is first determined. The corresponding phase compensation model is then used to calculate the phase delay angle to be compensated. Based on this phase delay angle, the controller advances or delays the signal sampling or processing timing, achieving real-time and precise compensation for the phase delay introduced by the op amp circuit.

[0023] In one embodiment, calculating the phase delay angle at each frequency based on the time domain delay includes: Calculate the signal period based on the frequency; Calculate the proportion of time domain delay in the signal period to obtain the period proportion; Convert the period ratio into angle system to get the phase delay angle at the corresponding frequency.

[0024] Calculate the signal period based on frequency: For any frequency ,in, The unit is Hertz (Hz), The unit is seconds (s). Indicates the frequencies, Indicates the corresponding The signal period of a frequency: Calculate the proportion of time domain delay in the signal period and get the period proportion: ,in For the The time domain delay of each frequency, For the The time domain delay of each frequency corresponds to the period ratio; thus, the time domain delay is converted into a ratio relative to the period.

[0025] Convert the period percentage to an angle system to get the phase delay angle at the corresponding frequency: Convert the period percentage to an angle system (0-360°) to get the phase delay angle at the frequency: ,in is the phase delay angle.

[0026] In one embodiment, the phase delay angles obtained from the original acquired data may contain measurement noise or outliers (e.g., zero values) due to instrument limitations, which can severely impact the accuracy of subsequent modeling. Therefore, rigorous data preprocessing is required. This application performs outlier interpolation and data smoothing filtering on the time-domain delay.

[0027] Among them, outlier interpolation processing: the phase delay angle is formed into a data sequence, the phase delay angle As the data points in the data sequence, linear interpolation is used to fill in the zero-value points or invalid data points in the data sequence. Interpolation phase delay angle The phase delay angle of the nearest valid data point before and after and Decide, Indicates the Interpolation phase delay angle The phase delay angle of the nearest valid data point, Indicates the Interpolation phase delay angle The phase delay angle of the nearest valid data point is interpolated as: ,in, For the Interpolation phase delay angle The frequency corresponding to the most recent valid data point, For the Interpolation phase delay angle The frequency corresponding to the most recent valid data point.

[0028] Data smoothing is to eliminate random measurement noise and use Savitzky-Golay filter to smooth the interpolated data. Savitzky-Golay filter can effectively smooth the data while retaining the main trend characteristics by fitting a low-order polynomial to the data points in the data sequence within a sliding window. The phase delay angle is calculated as follows:

[0029] in, is the filter window size, are pre-calculated filter coefficients, The interpolated data points The corresponding phase delay angle.

[0030] In one embodiment, the phase-frequency curve is segmented by using the transition point of the local slope of the phase-frequency curve as the trigger point for segmentation, including: Calculate the rate of change of the phase delay angle with frequency to obtain the local slope; Determine whether the local slope change exceeds the set threshold or there is a sign change; When a set threshold is exceeded or a sign change occurs, the phase-frequency curve is segmented using the transition point of the local slope as the segmentation trigger point.

[0031] The phase compensation model of each segment is expressed as: , For the The phase after segment compensation, and For the The compensation coefficient of the segment, and By minimizing the The sum of squares of the residuals for all data points in the segment Obtain: , The smoothed data points The phase delay angle.

[0032] First, calculate the local slope (i.e., the rate of change of the phase delay angle with frequency): ,in, Represents the smoothed data points The phase delay angle. For the frequencies.

[0033] Ideally, within a continuous frequency band, the local slope It should remain relatively stable or show monotonic changes. This application sets a slope change threshold criterion: when the value of the local slope changes dramatically, changes in sign, or deviates from the average slope of the current segment by more than the preset threshold, it is considered to be a transition point of physical properties and should be used as the boundary of the segmented fitting. For example, when the local slope A segment is triggered when a value suddenly changes from a stable positive area to a negative value or a value of a significantly different order of magnitude.

[0034] According to the above criteria, the entire frequency range Divided into multiple sub-intervals, is the minimum frequency, is the maximum frequency.

[0035] For example, the starting frequency corresponding to the transition point of the local slope shows two distinct linear characteristics after 50kHz, with an abnormal region with a negative local slope in the middle. Therefore, the effective fitting area is divided into two segments: the first segment: from 50kHz to the beginning of the abnormal region. The second segment: from the end of the abnormal region to the highest frequency. For each segment, the least squares method is used for linear fitting to establish a segmented phase compensation model. The resulting multiple phase compensation models form a piecewise function. In servo drive control, for any operating frequency, the frequency segment to which it belongs is first determined. Then, the corresponding phase compensation model is called to calculate the phase delay angle to be compensated. Based on this angle, the controller advances or delays the signal sampling or processing timing, achieving real-time and precise compensation for the phase delay introduced by the op amp circuit.

[0036] The segmentation criteria in the embodiments of this application are based on data-driven slope characteristics rather than fixed frequency points. This means that the method can automatically adapt to the unique frequency characteristics exhibited by op amp circuits of different models and batches. Users only need to re-collect data and run the algorithm to generate a highly customized phase compensation model for the new hardware, which is highly versatile and portable.

[0037] The accuracy of the phase compensation model can be evaluated by calculating the relative error between the fitted value and the true (smoothed) value. for: .

[0038] The relative errors of all fitting points are averaged to obtain the overall average error of the phase compensation model. Through the solution of this application, the final phase compensation model has an average relative error of less than 4.08%, that is, an average accuracy of up to 95.92%.

[0039] See also Figure 2 The embodiment of the present application provides a comparison diagram of the effects of using single-segment fitting in the prior art and segmented fitting by the method of the present application. The dotted line is the error between the curve fitted by the traditional method and the actual phase, and the solid line is the relative error of the phase after segmented fitting by the method of the present application. It can be found that the relative error of the segmented fitting (solid line) is greatly reduced in the low frequency band compared to the error of the single-segment fitting, and high precision is maintained in the high frequency band.

[0040] On the other hand, an embodiment of the present application provides a data analysis system for analyzing phase variations of an encoder operational amplifier circuit with frequency, comprising: a time domain delay module for obtaining a time domain delay between an input signal and an output signal of the operational amplifier circuit, wherein the input signal is generated by a frequency-adjustable sinusoidal wave signal source; A phase delay angle calculation module is used to calculate the phase delay angle at each frequency based on the time domain delay; A drawing module is used to draw a phase-frequency curve of phase delay angle and frequency; A segmentation module, configured to segment the phase-frequency curve by using the transition point of the local slope of the phase-frequency curve as a segmentation trigger point; The compensation module uses the least square method to perform linear fitting on each segment to obtain a phase compensation model for each segment, and compensates the phase delay angle according to the phase compensation model.

[0041] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements and improvements made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A data analysis method for the phase variation of an encoder operational amplifier circuit with frequency, characterized in that: include: Obtaining a time domain delay between an input signal and an output signal of an operational amplifier circuit, wherein the input signal is generated by a frequency-adjustable sinusoidal wave signal source; Calculate the phase delay angle at each frequency based on the time domain delay; Plot the phase-frequency curve of phase delay angle and frequency; The phase-frequency curve is segmented by taking the transition point of the local slope of the phase-frequency curve as the trigger point of segmentation; The least square method is used to perform linear fitting on each segment to obtain a phase compensation model for each segment, and the phase delay angle is compensated according to the phase compensation model.

2. The data analysis method of the phase variation of the encoder operational amplifier circuit according to claim 1, characterized in that: Obtaining the time domain delay between the input signal and the output signal of the operational amplifier circuit includes passing the input signal and the output signal of the operational amplifier circuit through a comparator to obtain the output signal of the comparator.

3. The data analysis method of the phase variation of the encoder operational amplifier circuit according to claim 1, characterized in that: Calculate the phase delay angle at each frequency based on the time domain delay, including: Calculate the signal period based on the frequency; Calculate the proportion of time domain delay in the signal period to obtain the period proportion; Convert the period ratio into angle system to get the phase delay angle at the corresponding frequency.

4. The data analysis method of the phase variation of an encoder operational amplifier circuit according to claim 1, characterized in that: The phase-frequency curve is segmented by taking the transition point of the local slope of the phase-frequency curve as the trigger point of the segmentation, including: Calculate the rate of change of the phase delay angle with frequency to obtain the local slope; Determine whether the local slope change exceeds the set threshold or there is a sign change; When a set threshold is exceeded or a sign change occurs, the phase-frequency curve is segmented using the transition point of the local slope as the segmentation trigger point.

5. The data analysis method of the phase variation of the encoder operational amplifier circuit according to claim 4 is characterized in that: The phase compensation model of each segment is expressed as: , For the The phase after segment compensation, and For the The compensation coefficient of the segment, For the frequencies, and By minimizing the The sum of squares of the residuals for all data points in the segment Obtain: , The smoothed data points The phase delay angle.

6. The data analysis method of the phase variation of the encoder operational amplifier circuit according to claim 1, characterized in that: Also includes: Perform outlier interpolation and data smoothing filtering on time domain delays.

Images