Automatic electrostatic trimming method for resonant gyroscope

Abstract

The invention discloses an automatic electrostatic trimming method for a resonant gyroscope. According to the method, on the basis of a dichotomy of a trimming voltage square range, the square of the trimming voltage serves as an independent variable, a dichotomy iteration interval is determined through a frequency splitting value and a change rate thereof, the optimal shaft adjusting voltage and the optimal frequency adjusting voltage are quickly approached, and the trimming precision lower than 1 mHz can be quickly achieved. The method is simple in principle and high in robustness, does not need complex parameter debugging, is convenient for programming and hardware integration, can remarkably improve the modal matching efficiency and precision of the micro-hemispherical resonator gyroscope, and is suitable for high-precision electrostatic trimming of the resonator gyroscope.

Description

Technical Field

[0001] The present invention relates to the technical field of resonant gyro trimming, and specifically relates to an automatic electrostatic trimming method for a resonant gyro. Background Technique

[0002] The microhemispherical resonant gyro is a MEMS resonant gyro with broad application prospects, used to measure the angular velocity or angular displacement of rotation. Manufacturing errors and material defects, etc., result in unequal frequencies of the two working modes of the gyro. The frequency difference is called frequency splitting. The low-frequency and high-frequency modes respectively correspond to the low-frequency rigid axis and the high-frequency rigid axis. Both frequency splitting and rigid axis offset greatly limit the improvement of gyro performance.

[0003] Adopting the method of frequency trimming can significantly reduce frequency splitting. The main trimming methods include mechanical trimming and electrostatic trimming. At present, many units have significantly reduced the frequency splitting of the resonator through advanced processing technologies such as chemical etching, femtosecond laser, and ion beam. However, affected by the trimming accuracy, there is a bottleneck in the frequency splitting at the mHz level, and the frequency splitting generated by packaging cannot be eliminated by removing mass, and electrostatic trimming needs to be further adopted to achieve mode matching.

[0004] Electrostatic trimming is to apply a bias voltage on a specific electrode of the resonant structure, and trim the resonant frequency by introducing negative stiffness. By establishing a trimming control loop, automatic electrostatic trimming can be realized on the premise of ensuring trimming accuracy. Many units have developed relevant control loops, using methods such as phase-locked loop combined with mode switching and function fitting to adjust the voltage to achieve mode matching, and using algorithms such as particle swarm algorithm, improved fuzzy algorithm, and neural network for automatic trimming.

[0005] Currently, most studies use the electrical signals in the measurement and control loop as reference values for feedback, and all target frequency splitting as an index; the function fitting method is affected by fitting errors, and the PID control requires repeated parameter adjustment, and the efficiency and accuracy need to be improved. In addition, the existing automatic trimming algorithms rely on a certain amount of data for model training, and the preprocessing workload is large, and it cannot meet the high-precision automatic electrostatic trimming applicable to both the azimuth of the rigid axis and frequency splitting. Summary of the Invention

[0006] In order to solve the problems of complex operation and low trimming efficiency in current electrostatic trimming, the present invention proposes an automatic electrostatic trimming method based on the dichotomy of the square range of trimming voltage. By using the substitution idea, the square of the trimming voltage is used as the independent variable, and the interval iteration is used to quickly approximate the optimal trimming voltage to achieve high-precision mode matching of the resonant gyro.

[0007] In order to achieve the above object, the trimming method provided by the present invention mainly includes the following steps:

[0008] S1 Hardware System and Initial Parameters:

[0009] Build the relevant hardware system, including a programmable DC power supply, a double-phase-locked loop, a mode switching module, a signal processing circuit, and a host computer. According to the structure and electrode distribution of the microhemispherical resonant gyroscope, determine the axis adjustment electrode and the frequency modulation electrode; measure the initial frequency splitting Δf0, and set the initial effective domain of the square of the trimming voltage as [U l 2, U h 2], where U l and U h are the upper and lower limits of the trimming voltage, the domain of the square of the axis adjustment voltage is [U θl 2, U θh 2], and the domain of the square of the frequency modulation voltage is [U fl 2, U fh 2].

[0010] During the trimming process, Δf is equal to the initially set driving mode frequency minus the detection mode frequency, and its absolute value is not calculated, so Δf can be negative.

[0011] When used as the axis adjustment voltage, the signs of the Δf change rates corresponding to U θl and U θh must be opposite, that is, the first-order derivatives must be one positive and one negative; when used as the frequency modulation voltage, the modal states corresponding to U fl and U fh must be orthogonal, that is, Δf must be one positive and one negative. Ensure that the square of the optimal trimming voltage is within this interval.

[0012] In the S2 electrostatic axis adjustment stage, reduce the frequency splitting to a minimum value by changing the axis adjustment voltage U θ so that the rigid axis is aligned with the direction of the frequency modulation electrode.

[0013] 2.1) Measure the frequency splittings Δf θl corresponding to U θh 2 and U l and Δf h .

[0014] 2.2) Calculate the midpoint of the current interval, and the calculation formula is:

[0015]

[0016] Apply the corresponding voltage U θm , and measure the frequency splitting Δf m .

[0017] 2.3) If Δf l > Δf h , update the left endpoint U l 2 = U θm 2, Δf l＝ Δf m ; if Δfl <Δf h , update the right endpoint U h 2 = U θm 2. Δf h＝ Δf m .

[0018] 2.4) Repeat steps 2.2 and 2.3 until the difference between Δf l and Δf h is within the set threshold range. At this time, it is regarded as the rigid axis being aligned with the FM electrode direction, and at this time U θm is the optimal axis adjustment voltage.

[0019] In the S3 electrostatic frequency modulation stage, by changing the frequency modulation voltage U f reduce Δf to the required value.

[0020] 3.1) Measure U fl 2 and the corresponding frequency splitting Δf of U fh 2 l and Δf h .

[0021] 3.2) Calculate the midpoint of the current interval. The calculation formula is:

[0022]

[0023] Apply the corresponding voltage U fm , and measure the frequency splitting Δf m .

[0024] 3.3) If Δf m > 0, update the left endpoint U l 2 = U fm 2, Δf l＝ Δf m ; if Δf m < 0, update the right endpoint U h 2 = U fm 2, Δf h＝ Δf m .

[0025] 3.4) Repeat steps 3.2 and 3.3 until Δf m is less than the required value.

[0026] S4 Modal matching verification

[0027] Apply the optimal axis adjustment voltage and frequency modulation voltage, measure the real-time frequency splitting through the double phase-locked loop and mode switching, confirm that Δf meets the requirements, and complete the modal matching.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] Through the dichotomy iteration in the voltage square interval, the present invention avoids the function fitting error and the PID tuning parameter problem, and the trimming accuracy can reach the sub-millihertz level (<1mHz).

[0030] The initial interval of the present invention is determined by fast search with large steps, with fewer iteration times, greatly improving the trimming accuracy and efficiency.

[0031] The algorithm does not require complex control parameters, and only needs to update the interval based on the size values of frequency splitting, which is convenient for programming and hardware integration. Description of the Drawings

[0032] Figure 1 It is the schematic diagram of automatic trimming in the embodiment of the present invention.

[0033] Figure 2 It is the relationship curve between the trimmed voltage square and the frequency splitting during the axis adjustment process of the harmonic oscillator in the embodiment of the present invention.

[0034] Figure 3 It is the rigid axis trimming process of the harmonic oscillator in the embodiment of the present invention.

[0035] Figure 4 It is the linear relationship between the trimmed voltage square and the frequency splitting during the frequency modulation process of the harmonic oscillator in the embodiment of the present invention.

[0036] Figure 5 It is the frequency splitting trimming process of the harmonic oscillator in the embodiment of the present invention. Detailed Embodiment

[0037] To make the purpose, technical solution and advantages of the present invention clearer and more definite, the technical solution of the present invention will be further described in detail below with reference to the drawings.

[0038] The following detailed embodiments, related data and their descriptions are only used to explain the present invention and do not limit the present invention. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

[0039] The automatic electrostatic trimming method of a resonant gyro mainly includes three stages: initial interval determination, rigid axis trimming, and frequency splitting trimming. The specific steps are as follows:

[0040] S1 Hardware System and Initial Parameters:

[0041] 1.1) Build the relevant hardware system, including a programmable DC power supply, a double phase-locked loop module, a mode switching module, a signal processing circuit and a host computer, test the frequency splitting of the gyro and determine the axis adjustment electrode and the frequency modulation electrode. The relevant control loop is as shown in the appendix Figure 1 as shown.

[0042] 1.2) Determine the left endpoint \(U\) of the shaft alignment voltage θl = 8.3V, and the right endpoint \(U\) θh = 25.5V. The effective domain of the initial voltage squared is [68.89V 2 , 650.25V²].

[0043] In the S2 electrostatic shaft alignment stage, align the high-frequency rigid shaft with the direction of the frequency modulation electrode.

[0044] 2.1) As shown in the appendix Figure 2 , the left and right endpoints of the initial effective domain are A and B respectively. Measure the corresponding frequency splitting \(\Delta f\) A and \(\Delta f\) B .

[0045] 2.2) Calculate the midpoint \(U\) C of the voltage squared in the interval A - B. The calculation formula is:[[]]

[0046]

[0047] Apply the corresponding voltage \(U\) C , and measure the frequency splitting \(\Delta f\) C .

[0048] 2.3) As shown in the appendix Figure 2 , the interval is updated to C - B.

[0049] 2.4) Repeat steps 2.2 and 2.3, and gradually narrow the domain of the voltage squared through bisection iteration until the difference in \(\Delta f\) corresponding to the left and right endpoints is within 1mHz. As shown in the appendix Figure 3 , the minimum value of the frequency splitting is 333.335mHz.

[0050] In the S3 electrostatic frequency modulation stage, the frequency splitting is reduced to below 1mHz.

[0051] 3.1) As shown in the appendix Figure 4 , the left and right endpoints of the initial frequency modulation voltage squared are F and G respectively. Determine the left endpoint \(U\) F = 25.0V, and the right endpoint \(U\) G = 43.7V. The effective domain of the initial voltage squared is [625.00V 2 , 1909.69V²]. Measure the frequency splitting \(\Delta f\) F corresponding to \(U\) G ² and \(\Delta f\) F and \(\Delta f\) G .

[0052] 3.2) The midpoint \(U\) H of the voltage squared in the current interval F - G. The calculation formula is:[[]]

[0053]

[0054] Apply the corresponding voltage U H , and measure the frequency splitting Δf H .

[0055] 3.3) As shown in the appendix Figure 4 , the update interval is F - H.

[0056] 3.4) Repeat steps 3.2 and 3.3, and gradually narrow the interval through dichotomy iteration until the frequency splitting is less than 1 mHz. As shown in the appendix Figure 5 , at this time the frequency modulation voltage is 32.355 V.

[0057] S4 Mode matching verification

[0058] Apply the axis adjustment voltage of 21.056 V and the frequency modulation voltage of 32.355 V, and measure the real-time frequency splitting through a double-phase-locked loop and mode switching. Δf < 1 mHz, meeting the target requirements.

Claims

1. An automated electrostatic trimming method for a resonant gyroscope, characterized in that, Including the following steps: S1 Determine the trimming electrode and initial parameters: According to the structure and electrode distribution of the microhemispherical resonant gyroscope, the axis-adjusting electrode and the frequency-modulating electrode are determined; the initial frequency splitting Δf0 is measured, and the initial effective domain of the square of the trimming voltage [U l 2 , U h 2 is set, where U l and U h are the upper and lower limits of the trimming voltage, the domain of the square of the axis-adjusting voltage is [U θl 2 , U θh 2 , and the domain of the square of the frequency-modulating voltage is [U fl 2 , U fh 2 . During the S2 static axis adjustment phase, by changing the axis adjustment voltage U θ reduce the frequency splitting to a minimum value, that is, align the rigid axis with the direction of the frequency modulation electrode. 2.1) Measure U θl 2 and U θh 2 corresponding frequency splitting Δf l and Δf h . 2.2) Calculate the midpoint of the current interval, and the calculation formula is: Apply the corresponding voltage U θm , and measure the frequency splitting Δf m . 2.3) If Δf l > Δf h , update the left endpoint U l 2 = U θm 2 , Δf l＝ Δf m ; If Δf l < Δf h , update the right endpoint U h 2 = U θm 2 , Δf h＝ Δf m . 2.4) Repeat steps 2.2 and 2.3 until the difference between Δf l and Δf h is within the set threshold range. At this time, it is regarded as the rigid axis aligned with the frequency modulation electrode direction, and U θm is the optimal axis adjustment voltage. In the S3 electrostatic frequency modulation stage, by changing the frequency modulation voltage U f reduce Δf to the required value. 3.1) Measure U fl 2 and U fh 2 corresponding frequency splitting Δf l and Δf h . 3.2) Calculate the midpoint of the current interval, and the calculation formula is: Apply the corresponding voltage U fm , and measure the frequency splitting Δf m . 3.3) If Δf m > 0, update the left endpoint U l 2 = U fm 2 and Δf l＝ Δf m ; if Δf m < 0, update the right endpoint U h 2 = U fm 2 and Δf h＝ Δf m . 3.4) Repeat steps 3.2 and 3.3 until Δf m is less than the required value. S4 Modal matching verification Apply the optimal axis adjustment voltage and frequency modulation voltage, measure the real-time frequency splitting through a double phase-locked loop and modal switching, confirm that Δf meets the requirements, and complete the modal matching.

2. The electrostatic trimming method according to claim 1, wherein This method is applicable to a resonant gyroscope with a rotating body structure and a working mode of "n = 2" wine glass mode.

3. The electrostatic trimming method according to claim 1, wherein Through the method of binary iteration, update the interval endpoints according to the magnitude or positive / negative of the frequency splitting value, and quickly approach the optimal trimming voltage.

4. The electrostatic trimming method according to claim 1, wherein The frequency splitting in the axis adjustment stage has a quadratic function relationship with the square of the axis adjustment voltage, and the optimal axis adjustment voltage is determined by using the function symmetry; the frequency splitting in the frequency modulation stage has a linear relationship with the square of the frequency modulation voltage, and the optimal square of the frequency modulation voltage is determined by using the sign change.

5. According to the electrostatic trimming method described in claim 1, the initial effective domain is determined by a large-step voltage scan to ensure that it includes the square of the optimal trimming voltage. Among them, the signs of the frequency splitting change rates corresponding to the left and right endpoints in the axis adjustment stage must be opposite; the modal states corresponding to the left and right endpoints in the frequency modulation stage must be orthogonal.

Images