Rapid frequency sweeping method and system in radio frequency acoustic filter design

Abstract

The invention provides a rapid frequency sweeping method and system in radio frequency acoustic filter design, and relates to the technical field of radio frequency acoustic filter simulation, and the method comprises the steps: determining a frequency sweeping band, and calculating the resonant frequency and anti-resonant frequency of an SAW device model in the frequency band; calculating a Q value to screen out a key characteristic frequency, determining a segmented node frequency of a sub-frequency band, selecting a Chebyshev zero point as a sampling frequency point in each sub-frequency band, and performing interpolation approximation on a sweep frequency curve by using a Maehly approximation method; according to the method, the accuracy and the reliability of a simulation result are ensured, and the calculation efficiency is improved while the simulation accuracy is ensured. According to the method, calculation of frequency points one by one in the to-be-researched frequency band can be avoided, and the calculation efficiency is remarkably improved under the condition that the calculation precision is ensured.

Description

Technical Field

[0001] The invention relates to the technical field of radio frequency acoustic filter device simulation, and in particular to a fast frequency sweeping method and system in radio frequency acoustic filter design. Background Art

[0002] RF acoustic filters are the mainstream filters in mobile RF front-ends, including surface acoustic wave (SAW) filters and bulk acoustic wave (BAW) filters. They are electronic devices that use the physical phenomenon of sound waves propagating on the surface of piezoelectric media to achieve signal processing. With the continuous development of 5G and 6G technologies, and the trend of continuous modularization, integration, and miniaturization of RF front-ends, RF acoustic filters have become an indispensable key component.

[0003] At present, numerical calculation methods such as Finite Element Method (FEM), Spectral Element Method (SEM), and Finite Difference Method (FDM) have the characteristics of accurate simulation when simulating RF acoustic filters with complex structures. However, as the scale of the structure increases, the solution becomes more and more time-consuming. Especially when analyzing broadband problems, it is necessary to solve each frequency point one by one, which requires huge computing resources. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a fast frequency sweeping method and system in the design of radio frequency acoustic filters, thereby improving the simulation efficiency.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] In a first aspect, a fast frequency sweep method in a radio frequency acoustic filter design comprises:

[0007] Determine the frequency sweep band and calculate the resonant frequency and anti-resonant frequency of the SAW device model;

[0008] First, the SAW model is built, including configuring the parameters of the SAW device model, including setting the model's geometric dimensions, material properties, and electrode configuration, and meshing the SAW device model. In order to simulate the real SAW device response, a perfectly matched layer (PML) absorption boundary is set around the model to simulate the real scene by absorbing elastic waves. Finally, the FEM method is used to construct the stiffness matrix K and mass matrix M of the entire model, and the characteristic frequency of the model is further solved;

[0009] In the simulation of the SAW device model, the electrode boundary conditions corresponding to the resonance / anti-resonance are set to calculate the characteristic frequency of the current model. When the electrode boundary conditions of the SAW device model are set to the short-circuit mode, that is, the potential on the electrode is set to 1 volt, the real part of the characteristic frequency calculated is the resonant frequency; when the electrode boundary conditions of the SAW device model are set to the open-circuit mode, that is, the total charge on the electrode is set to 0 coulomb, the real part of the characteristic frequency calculated is the anti-resonant frequency, and the steps are repeated to complete the calculation of the resonant frequency and anti-resonant frequency over the entire frequency band.

[0010] Filter out key characteristic frequencies as node frequencies of sub-bands, including:

[0011] When the PML absorption boundary is added to the SAW device model, the calculated characteristic frequency is the complex frequency f = f real +i·f imag . Among them, the quality factor Q = f real / 2f imag It can be calculated from the real and imaginary parts of the characteristic frequency. Since some resonant modes are absorbed by the PML layer, the Q value is small, and the peak characteristics will not be reflected in the frequency response curve; while the characteristic frequency with a larger Q value will show a sharper peak in the frequency response curve, so the characteristic frequency with a larger Q value must be retained as the segment node frequency of the subsequent sub-band.

[0012] The selection of key characteristic frequencies is crucial to the performance of the algorithm. By comparison, the change of conductance with frequency is greater than that of the modulus of admittance, so the node frequencies of the sub-bands are screened according to the conductance frequency response curve. Combining the characteristics of the Maehly approximation method, numerical experiments are performed to determine the relationship between different device structures (number of interdigital pairs, piezoelectric layer materials and thickness, etc.) and the number of peaks in the conductance frequency response curve, and the key characteristic frequencies are determined based on the order of Q values.

[0013] When studying a surface acoustic wave (SAW) model with 120 interdigital pairs, it was found that when the observation frequency range is 0.35GHz, its conductivity frequency response curve will roughly show 13 to 15 peaks. Therefore, after sorting the calculated resonant frequency and anti-resonant frequency from large to small according to the Q value, the first 25 to 30 frequencies can be selected for further screening in the subsequent steps.

[0014] In the observed frequency band, the resonant frequency and the anti-resonant frequency usually appear alternately. If they are close to each other, there is no need to distinguish them. The two closer resonant and anti-resonant frequencies can be combined as the node frequency of the sub-band. This position roughly corresponds to the peak of the conductance frequency response curve.

[0015] In the above characteristic frequency calculation process, the characteristic frequency calculation can be accelerated by limiting the search area with a smaller imaginary part of the characteristic frequency.

[0016] Furthermore, the node frequency is used to divide the entire frequency band into multiple sub-frequency bands. In each sub-frequency band, the Chebyshev zero point is selected as the sampling frequency point, and the Maehly approximation method is used to interpolate the frequency sweep curve, including:

[0017] A fixed number of Chebyshev zeros are selected within the frequency band, and the frequency response of the SAW device model at the current frequency is calculated. The calculation result is the admittance value.

[0018] Among them, the reasonable selection of the number of Chebyshev zeros is crucial to improving the performance of the algorithm. After numerical analysis of the frequency response curves of the SAW device models of various structures, it is generally found that taking 5 to 7 points in each sub-band can obtain a better interpolation effect without consuming too much computing resources.

[0019] The Chebyshev zero point is used as the sampling frequency point and its admittance value is passed into the Maehly approximation algorithm as a parameter to calculate the interpolation curve.

[0020] Furthermore, a fixed number of Chebyshev zeros are selected within the frequency band to calculate the frequency response of the SAW device model at the current frequency point. The calculation result is the admittance value, including:

[0021] Non-uniform sampling points are selected within the frequency band, and the sampling point positions correspond to the positions of Chebyshev zero points;

[0022] Perform frequency response simulation on the SAW device model. During the simulation, the frequency corresponding to the current Chebyshev zero point is used as input, and the response of the model is observed to obtain the simulation result.

[0023] The admittance value at each frequency point is extracted from the simulation results, which is the combination of conductance and susceptance.

[0024] Furthermore, the sampling frequency value and the admittance value corresponding to the Chebyshev zero point are passed as parameters into the Maehly approximation algorithm to calculate the interpolation curve, including:

[0025] Arrange the sampling frequency values ​​and admittance values ​​corresponding to the Chebyshev zeros into a format accepted by the Maehly approximation algorithm, that is, create a data set containing two columns: one column is the frequency value and the other column is the corresponding admittance value;

[0026] The frequency response of the SAW device is approximated by using the Chebyshev zero point as the frequency sampling point, and the formatted frequency value and admittance value are passed as parameters to the Maehly approximation algorithm. The parameters will be used for calculations within the Maehly approximation algorithm to construct an interpolation curve.

[0027] In a second aspect, a fast frequency sweeping system in a radio frequency acoustic filter design comprises:

[0028] A determination module is used to determine the frequency sweeping band and calculate the key characteristic frequencies of the SAW device model, including the resonant frequency and the anti-resonant frequency, and sort them from large to small based on the Q value, and select the resonant and anti-resonant frequencies with larger Q values;

[0029] The interpolation module is used to segment the frequency band using the key characteristic frequency. In each sub-band, the Chebyshev zero point is selected as the sampling frequency point, and the Maehly approximation method is used to interpolate the sweep frequency curve.

[0030] According to a third aspect, a computing device includes:

[0031] one or more processors;

[0032] The storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the method described.

[0033] In a fourth aspect, a computer-readable storage medium stores a program, and when the program is executed by a processor, the method described is implemented.

[0034] The above solution of the present invention includes at least the following beneficial effects:

[0035] By simulating the frequency response curve through SAW device model reduction or function approximation, the calculation time and computing resource consumption can be greatly reduced. Maehly approximation is a commonly used interpolation algorithm with good effect. It uses the Chebyshev points in the frequency band as sampling points, calculates the coefficients of the Chebyshev series, and then approximates it with rational functions. It has good function approximation effect. Combining this algorithm with numerical methods improves simulation efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Schematic diagram of the structure of a SAW single-port resonator

[0037] Figure 2 The typical FEM swept-frequency admittance modulus and conductance result curve of the SAW device model.

[0038] Figure 3 This is the flow chart of FEM swept frequency calculation based on Maehly approximation.

[0039] Figure 4 Typical resonant and anti-resonant mode vibration diagrams of the SAW device model.

[0040] Figure 5 The FEM frequency sweep curve and segmentation points of the SAW device

[0041] Figure 6are the sampling points and their calculated admittance modulus and conductance values.

[0042] Figure 7 is the Maehly approximate FEM fitting curve based on 7 Chebyshev sampling points.

[0043] Figure 8 This is a comparison curve of the admittance characteristics of the SAW single-port resonator.

[0044] Fig. 9 This is a comparison chart of the admittance characteristics of the commercial software COMSOL using the FEM and AWE methods.

[0045] Fig.10 This is an enlarged mesh diagram of the structure on the left side of the resonator. DETAILED DESCRIPTION

[0046] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in a form and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0047] An embodiment of the present invention provides a fast frequency sweeping method in radio frequency acoustic filter design, the method comprising the following steps:

[0048] Determine the frequency sweep band and calculate the key characteristic frequencies of the SAW device model, including the resonant frequency and anti-resonant frequency. After sorting them from large to small based on the Q value, select the frequency with a larger Q value, and merge the resonant and anti-resonant frequencies with a close distance to obtain the node frequency of the sub-band.

[0049] The frequency band is segmented by node frequencies. In each sub-band, the Chebyshev zero point is selected and the sweep frequency curve is interpolated using the Maehly approximation method.

[0050] In an embodiment of the present invention, determining a frequency sweep band and calculating a resonant frequency and an anti-resonant frequency of a SAW device model includes:

[0051] The electrode boundary condition of the SAW device model is set to the short-circuit mode, and the real part of the calculated characteristic frequency is the resonant frequency.

[0052] The electrode boundary condition of the SAW device model is set to open circuit mode. The real part of the calculated characteristic frequency is the anti-resonance frequency. Repeat the steps to complete the calculation of the resonance / anti-resonance frequency in the entire frequency band.

[0053] After sorting from large to small based on Q value, select the resonant or anti-resonant frequency with larger Q value. This is because some resonant modes are absorbed by the PML layer, so the resonant or anti-resonant frequency with smaller Q value will not reflect the resonant characteristics on the frequency response curve, so it is not considered. In addition, when the calculated resonant and anti-resonant frequencies are close to each other, there is no meaning of segmentation, and the closer resonant and anti-resonant frequencies can be merged. Finally, each processed frequency point is used as a node frequency, and the entire frequency band is divided into several sub-bands.

[0054] In an embodiment of the present invention, through the fast frequency sweeping method, the designer can quickly obtain the performance of the filter at different frequencies, which greatly shortens the design cycle and improves the design efficiency. The selection of segmented interpolation and Chebyshev zeros makes the frequency sweeping process more efficient and can quickly find the frequency band that meets the design requirements. By segmenting the frequency band using key characteristic frequencies (such as resonant frequency and anti-resonant frequency), the performance of the filter in each frequency band can be analyzed more accurately. This segmented analysis helps designers to optimize the performance of the filter in different frequency bands in a targeted manner, thereby improving the overall performance. By using the Maehly approximation method to interpolate the frequency sweeping curve, the complexity of the calculation can be reduced while ensuring a certain accuracy. This approximation method can maintain a high analysis accuracy while reducing the amount of calculation. The fast frequency sweeping method avoids unnecessary repeated calculations and waste of resources through effective approximate calculations.

[0055] In the embodiment of the present invention, the frequency band is segmented by using the key characteristic frequency, and in each sub-frequency band, the Chebyshev zero point is selected as the sampling frequency point, and the sweep frequency curve is interpolated by using the Maehly approximation method, including:

[0056] A fixed number of Chebyshev zeros are selected within the frequency band, and the frequency response of the SAW device model at the current frequency is calculated. The calculation result is the admittance value.

[0057] The Chebyshev zero point is used as the sampling frequency point and its corresponding admittance value is passed into the Maehly approximation algorithm as a parameter to calculate the interpolation curve.

[0058] In an embodiment of the present invention, by segmenting the frequency band using key characteristic frequencies (such as resonant frequency and anti-resonant frequency), it is possible to ensure that the filter characteristics in each sub-band are more consistent, thereby improving the accuracy of frequency band analysis. This segmentation method helps to understand the behavior of the filter in different frequency ranges in more detail. Selecting Chebyshev zeros as sampling points in each sub-band can effectively capture key information in the frequency band. Chebyshev zeros have the advantage of providing the best approximation characteristics with the least number of points in a given interval, so they can more accurately reflect the frequency response of the filter. By fitting the sweep frequency curve using the Maehly approximation method, a more accurate interpolation curve can be obtained using fewer sampling points. This method not only improves the interpolation efficiency, but also reduces the amount of data required for calculation, thereby speeding up the design iteration. Directly calculating the frequency response of the SAW device model at each Chebyshev zero to obtain the admittance value can avoid complex conversion processes and directly obtain parameters closely related to the filter performance. This simplifies the steps of data analysis, allowing designers to evaluate the performance of the filter more quickly.

[0059] In the embodiment of the present invention, a fixed number of Chebyshev zeros are selected in each sub-band, and the frequency response of the SAW device model at the current frequency is calculated. The calculation result is an admittance value, including:

[0060] Determine the number of Chebyshev zeros to be selected in each sub-band according to the required accuracy and computing resources;

[0061] The sampling points are selected non-uniformly in the sub-band, and the positions of the sampling points correspond to the positions of the Chebyshev zero points;

[0062] Perform frequency response simulation on the SAW device model. During the simulation, the frequency corresponding to the Chebyshev zero point is used as input, and the response of the model is observed to obtain the simulation result.

[0063] The admittance value at each frequency point is extracted from the simulation results, which is the combination of conductance and susceptance.

[0064] In an embodiment of the present invention, by predetermining the number of Chebyshev zeros in each sub-band, computing resources can be rationally planned to avoid unnecessary computing waste. This fixed-point sampling method reduces the amount of calculation while ensuring the accuracy of the analysis results, thereby improving the overall computing efficiency. As a special sampling point, the distribution of Chebyshev zeros in the frequency band can effectively capture the key changes in frequency response. This distribution method ensures that representative sampling data can be obtained even when the frequency band is wide or the response is complex, providing a reliable basis for subsequent interpolation and analysis. By simulating the frequency response of the SAW device model and observing the response of the model at a specific frequency, a high degree of simulation can be obtained. These results can accurately reflect the performance of the device under actual working conditions and provide valuable reference information for designers. Directly extracting the admittance value (a combination of conductance and susceptance) from the simulation results provides designers with intuitive electrical characteristic data. These data are not only easy to understand and analyze, but can also be directly used to evaluate the performance of the filter, simplifying the design verification process.

[0065] In the embodiment of the present invention, the sampling frequency value and the admittance value corresponding to the Chebyshev zero point are passed as parameters into the Maehly approximation algorithm to calculate the interpolation curve, including:

[0066] Arrange the sampling frequency values ​​and admittance values ​​corresponding to the Chebyshev zero points into a format accepted by the Maehly approximation algorithm, that is, create a data set containing two columns: one column is the frequency value, and the other column is the corresponding admittance value;

[0067] The frequency response of the SAW device is approximated by the sampling frequency value corresponding to the Chebyshev zero point, and the formatted frequency value and admittance value are passed as parameters to the Maehly approximation algorithm. The parameters will be used for calculations within the Maehly approximation algorithm to construct an interpolation curve.

[0068] In an embodiment of the present invention, by arranging the sampling frequency value and the admittance value corresponding to the Chebyshev zero point into a specific format and passing it to the Maehly approximation algorithm, efficient use of data is achieved. This method avoids redundant data collection and processing, and ensures that the data points used in the interpolation process are representative and accurate. The Maehly approximation algorithm is an effective curve interpolation method that can generate an accurate interpolation curve using limited data points. By passing the data of the sampling frequency corresponding to the Chebyshev zero point, the algorithm can capture the key features of the frequency response and generate an interpolation curve matching it, thereby more accurately reflecting the frequency response characteristics of the SAW device. Using the Maehly approximation algorithm to calculate the interpolation curve simplifies the design process. Designers do not need to manually perform complex curve approximation operations, and only need to pass the data of the sampling point to the algorithm to quickly obtain the interpolation result. This greatly saves design time and improves design efficiency. The method has certain flexibility and scalability. Designers can adjust the number and position of the Chebyshev zero point selection as needed to adapt to different design requirements. At the same time, the Maehly approximation algorithm itself also supports parameter adjustment, and the interpolation effect can be optimized according to actual conditions. Through precise interpolation curves, designers can more accurately predict and evaluate the performance of SAW devices at different frequencies. This helps improve the reliability of the design and reduce risks in actual production.

[0069] In the embodiment of the present invention, the electrode boundary condition of the SAW device model is set to an open circuit mode, the real part of the characteristic frequency calculated is the anti-resonance frequency, and the calculation of the anti-resonance frequency in the entire frequency band is completed by repeating the steps, including:

[0070] The open circuit mode is set to impose the charge value representing the electrodes in the finite element matrix of the SAW device model to be 0 coulomb;

[0071] After setting the open circuit boundary condition, start the simulation calculation of the characteristic frequency to find the resonant mode of the SAW device that meets the open circuit boundary condition, that is, the characteristic frequency;

[0072] After the simulation is completed, the characteristic frequency is extracted from the simulation results. The real part of the characteristic frequency calculated under the open circuit boundary condition is the anti-resonance frequency; repeat the steps to complete the calculation of all anti-resonance frequencies.

[0073] Using a similar method as above, by setting the electrode boundary condition to short-circuit mode, the calculation of all resonant frequencies within the frequency band can be completed.

[0074] In the embodiment of the present invention, by setting the electrode boundary condition to a short circuit or open circuit mode, the frequency response of the SAW device in the maximum or minimum energy exchange state can be accurately simulated, which is a key parameter in the filter design. In this way, the designer can accurately capture the characteristics of the resonant / anti-resonant frequency, and design the frequency response of the filter more accurately, thereby improving the performance and working efficiency of the filter, reducing the interference and distortion that may occur when the filter is working, and helping to optimize the overall performance of the filter and improve the quality of signal transmission.

[0075] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with specific embodiments and drawings.

[0076] First, it is necessary to clarify the SAW device model structure and frequency response characteristics of the SAW resonator. Figure 1 The structure of a single-port four-layer SAW resonator is shown. It includes an IDT electrode, reflective gratings on the left and right sides of the electrode, and a multilayer substrate. IDT is a transducer composed of periodically arranged metal fingers, each period containing a pair of alternately placed metal fingers. The multilayer substrate is a LT-42 / SiO2 / Poly-Si / Si structure, in which LT-42 is a piezoelectric substrate.

[0077] When a high-frequency electrical signal is applied to both ends of the IDT electrode, the surface of the piezoelectric substrate will produce mechanical vibrations and simultaneously excite surface acoustic waves with the same frequency as the external electrical signal.

[0078] When adding excitation of electrical signals of different frequencies to the electrodes of the SAW device model, the frequency response analysis of the SAW device model can be performed to obtain the Y parameter admittance curve. Through the Y parameter curve, the resonance and anti-resonance frequencies of the device and the frequency response information of other frequencies can be analyzed. Figure 2 This is a typical frequency response curve.

[0079] The specific example of the implementation of the present invention is to improve the traditional FEM frequency sweep using the Maehly approximation algorithm. Figure 3 As shown, Figure 3 The finite element fast frequency sweeping technology based on Maehly approximation provided in the embodiment of the present invention comprises the following steps:

[0080] (I) Steps for determining the frequency sweep segmentation points: Each vibration mode of a SAW device contains a resonant frequency and an anti-resonant frequency, such as Figure 4 The resonant frequency is the free vibration frequency when the electrical terminals are short-circuited, at which point the circuit impedance is minimum. The anti-resonant frequency is the free vibration frequency when the electrical terminals are open-circuited, at which point the circuit impedance is maximum. To calculate the eigenfrequency in FEM, you first need to construct the system matrix, and then solve the generalized eigenvalue problem to obtain the eigenfrequency.

[0081] Specifically, first establish the SAW device model and do mesh generation, such as Figure 1 The FEM method is used to discretize the piezoelectric substrate described by the piezoelectric medium constitutive equation and the electrodes and other substrates described by the elastic medium constitutive equation. Then, the overall stiffness matrix K and mass matrix M of the entire SAW device model are obtained by assembling the local matrices of each part. The degrees of freedom of the SAW device model are numbered from 1 to N. The generalized eigenvalue problem generally requires the participation of the K and M matrices in the calculation, see formula (1).

[0082] (K-ω 2 M)μ=0 (1)

[0083] Generally, large-scale finite element matrices can be calculated by subspace projection and other methods. If you want to speed up the calculation of eigenvalues, you can use a method that limits the search domain (such as the Chebyshev-Alnoldi algorithm). Finally, select the eigenfrequency with a large Q value.

[0084] Assume the sweep frequency band is [f a ,f b ], the specific steps of solving are as follows:

[0085] The resonant frequency is the free vibration frequency when the electrical terminals are short-circuited. By imposing a fixed potential value (such as 1V) on the positive electrode and an electrical boundary condition of grounding on the negative electrode, the SAW device model can be placed in a short-circuited state. In the specific calculation, we set the degrees of freedom of the positive and negative electrode potentials as i~j. Since the potential has been imposed, before calculating the characteristic frequency, the i~j rows and i~j columns of the stiffness matrix K and the mass matrix M are removed. Let the stiffness matrix after removing the degrees of freedom be K ij , the mass matrix is ​​M ij . Then the generalized eigenvalue problem becomes

[0086] (K ij -ω 2 M ij )μ=0 (2)

[0087] The real part of the characteristic frequency solved at this time is the resonant frequency. The anti-resonant frequency is the free vibration frequency when the electrical end is open. By imposing the electrical boundary conditions of zero charge on the positive electrode and grounding on the negative electrode, the SAW device model can be placed in an open-circuit state. At this time, the sum of the accumulated charges at the electrodes is zero, and since the positive electrode is an equipotential body, the potential degrees of freedom at the positive electrode are all the same value.

[0088] Based on the above characteristics, let the potential freedom of the positive electrode be i p ~j p , the stiffness matrix K and the mass matrix M i p -jp Row and i p -j p The columns are summed up and only the i-th p Line p The columns store the summed results, and the other rows and columns are removed. Let the number of the degree of freedom of the negative electrode potential be i n ~j n , since the ground boundary condition has been imposed, before calculating the eigenfrequency, the stiffness matrix K and the mass matrix M are n ~j n Row and i n ~j n The column is removed, so the SAW device stiffness matrix K is obtained when the electrical end is open ij and the mass matrix M ij . Similar to the calculation method of resonant frequency, by calculating K ij and M ij The generalized eigenvalue problem of , see formula 2, can be used to obtain the anti-resonance frequency.

[0089] Based on the above process, the resonant and anti-resonant frequencies of the SAW device model are solved. The resonant and anti-resonant frequencies are sorted from large to small based on the Q value. The frequency with a larger Q value is taken, and the resonant and anti-resonant frequencies with a closer distance are merged to obtain the node frequency to divide each sub-band. Specifically, Figure 5 For example, the frequency band is divided into:

[0090] [1.8000GHz,1.8112GHz],[1.8112GHz,1.8326GHz],[1.8326GHz,1.8518GHz],[1.8518GHz,1.8708GHz],[1.8708GHz,1.8884GHz],[1.8884GHz,1.9013GHz], [1.9013GHz,1.9073GHz],[1.9073GHz,1.9917GHz],[1.9917GHz,2.0677GHz],[2.0677GHz,2.0777GHz],[2.0777GHz,2.0924GHz],[2.0924GHz,2.1102GHz],

[0091] [2.1102GHz,2.1296GHz],[2.1296GHz,2.1500GHz].

[0092] (II) Maehly approximation steps for segmented curve construction:

[0093] In each sub-band, select the frequency points corresponding to the 7 Chebyshev zeros, and calculate the admittance value of each frequency point as the function value of the sampling point, such as Figure 6 The Maehly approximation algorithm uses Chebyshev polynomials to construct a rational fraction function with parameters. The sampling frequency points and their admittance values ​​are used as interpolation points to calculate the parameters in the rational fraction, and then calculate a continuous interpolation function, such as Figure 7 .

[0094] Specifically, first, for any frequency f∈[f a ,f b ], and we can get the following coordinate transformation:

[0095]

[0096] in So:

[0097]

[0098] Construct Maehly approximation of rational fraction function R in [-1,1] LM (f) to approximate the frequency response curve, see formula (5):

[0099]

[0100] where p L (f) and Q M (f) are the numerator and denominator polynomials of the Maehly rational fraction, L and M are the orders of the numerator and denominator polynomials. k (f) = cos(karccosf) is the Chebyshev polynomial, a i (i=1,...,L) and b i (i=1,...,M) are the coefficients of the numerator and denominator polynomial of the rational fraction respectively.

[0101] By comparing with the Chebyshev series, the polynomial coefficient a is solved by constructing a system of equations. i (i=1,...,L) and b i (i=1,...,M), and then obtain the Maehly approximate function in each sub-band, and obtain the frequency response admittance value of the unknown frequency point in the band, such as Figure 7 .

[0102] In the following, this specific embodiment uses a single-port SAW resonator as a SAW device model to provide a time comparison between an existing FEM algorithm based on Maehly approximation and a traditional FEM algorithm.

[0103] Firstly, the structural parameters of the basic unit of IDT are given as shown in Table 1. The grid division of its part is as follows: Fig.10 As shown in FIG. 8 , an equidistant grid density distribution is adopted in the direction from the surface to the bottom of the piezoelectric substrate (y direction).

[0104] Table 1: Structural parameters of IDT basic unit

[0105]

[0106] The piezoelectric material of the resonator substrate is set to LT-42, using aluminum electrodes, and there are 20 short-circuit reflective grating fingers with the same period on each side.

[0107] The sweep frequency range is 1.8GHz to 2.15GHz, and the sweep frequency step is 1MHz. Figure 8 The resonator admittance and conductance results obtained by traditional FEM calculation and Maehly approximate calculation are compared. It can be seen that the two are in good agreement. The FEM frequency sweep calculates 350 frequency points, while Maehly only calculates 98 frequency points, which restores the results of the traditional FEM frequency sweep. The computing platform used is a dedicated computing workstation with the following configuration: 11th Gen Intel(R)Core(TM)i9-11900K@3.50GHz, 3504MHz, 8 cores, 16 logical processors.

[0108] Finally, the time consumption of the Maehly approximate accelerated frequency sweep method proposed in the present invention is compared with the traditional FEM method and the AWE frequency sweep acceleration method adopted by the commercial software COMSOL. As shown in Table 2. The Maehly approximate accelerated frequency sweep method proposed in the present invention improves by 53.31%, which is because the present invention only needs fewer frequency points to fit the calculation curve of FEM. However, the conductivity curve of AWE has the problem of error, such as Fig. 9 .

[0109] Table 2 compares the time consumption of the two frequency sweeping methods: traditional FEM and FEM combined with Maehly

[0110] method FEM AWE Maehly Time[s] 1418 1159 662 Improve efficiency / 18.27％ 53.31％

[0111] Compared with the time-consuming frequency sweep calculation of traditional FEM, the present invention performs segmented frequency sweep processing by calculating the resonance point of the SAW device model, and restores the curve by segmented interpolation of the admittance curve using Maehly approximation. Compared with traditional FEM, only about 28% of the frequency points need to be calculated to restore the frequency response curve.

[0112] It should be noted that the system is a system corresponding to the above method, and all implementation methods in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0113] The embodiment of the present invention further provides a computing device, comprising: a processor, a memory storing a computer program, wherein when the computer program is executed by the processor, the method described above is executed. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0114] The embodiment of the present invention also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the method described above. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

Claims

1. A fast frequency sweeping method in the design of a radio frequency acoustic filter, characterized in that: The method comprises: Determine the sweep frequency band and calculate the key characteristic frequencies of the SAW device model, including the resonant frequency and anti-resonant frequency; Calculate the Q value of the resonant frequency and the anti-resonant frequency, sort them from large to small according to the Q value, and retain the frequency corresponding to the Q value as the segment node frequency of the sub-band; The frequency band is segmented by key characteristic frequencies. In each sub-band, the Chebyshev zero point is selected as the sampling frequency point, and the Maehly approximation method is used to interpolate and approximate the sweep frequency curve.

2. The fast frequency sweeping method in the design of the radio frequency acoustic filter according to claim 1, characterized in that: Determine the frequency sweep band and calculate the key characteristic frequencies of the SAW device model, including the resonant frequency and anti-resonant frequency, including: The electrode boundary condition of the SAW device model is set to the short-circuit mode. The calculated characteristic frequency is a complex number, and the real part is the resonant frequency. The Q value is calculated by the ratio of the real part to the imaginary part of the characteristic frequency. In the process of calculating the characteristic frequency, the search domain corresponding to the imaginary part of the characteristic frequency is limited to accelerate the solution of the characteristic value. Set the electrode boundary condition of the SAW device model to open circuit mode. The real part of the calculated characteristic frequency is the anti-resonance frequency. Repeat the steps to complete the calculation of the anti-resonance frequency.

3. The fast frequency sweeping method in the design of radio frequency acoustic filter according to claim 2, characterized in that: The frequency band is segmented using the key characteristic frequency. In each sub-band, the Chebyshev zero point is selected and the sweep frequency curve is interpolated using the Maehly approximation method, including: A fixed number of Chebyshev zeros are selected within the frequency band, and the frequency response of the SAW device model at the current frequency is calculated. The calculation result is the admittance value. The sampling frequency points corresponding to the Chebyshev zeros and their admittance values ​​are passed as parameters into the Maehly approximation algorithm to calculate the interpolation curve.

4. The fast frequency sweeping method in the design of the radio frequency acoustic filter according to claim 3, characterized in that: A fixed number of Chebyshev zeros are selected within the frequency band to calculate the frequency response of the SAW device model at the current frequency. The calculation result is the admittance value, including: Determine the number of Chebyshev zeros to be selected in each sub-band according to the required accuracy and computing resources; Non-uniform sampling points are selected within the frequency band, and the sampling point positions correspond to the positions of Chebyshev zero points; Perform frequency response simulation on the SAW device model. During the simulation, the frequency corresponding to the current Chebyshev zero point is used as input, and the response of the model is observed to obtain the simulation result. The admittance value at each frequency point is extracted from the simulation results, which is the combination of conductance and susceptance.

5. The fast frequency sweeping method in the design of radio frequency acoustic filter according to claim 4, characterized in that: The Chebyshev zero point is used as the sampling frequency value and its corresponding admittance value is passed into the Maehly approximation algorithm as a parameter to calculate the interpolation curve, including: The Chebyshev zeros are used as sampling frequency points and their corresponding admittance values ​​are sorted into a format accepted by the Maehly approximation algorithm, that is, a data set containing two columns is created: one column is the frequency value, and the other column is the corresponding admittance value; The frequency response of the SAW device is approximated by the Chebyshev sampling point, and the formatted frequency value and admittance value are passed as parameters to the Maehly approximation algorithm. The parameters will be used for calculations within the Maehly approximation algorithm to construct an interpolation curve.

6. A fast frequency sweeping system in the design of radio frequency acoustic filters, characterized in that: The system is used to perform the method according to any one of claims 1 to 5, comprising: A determination module, used to determine the frequency sweeping band and calculate key characteristic frequencies of the SAW device model, including the resonant frequency and the anti-resonant frequency; The interpolation module is used to segment the frequency band using the key characteristic frequency, select the Chebyshev zero point in each sub-band, and use the Maehly approximation method to interpolate the sweep frequency curve.

7. A computing device, characterized in that include: one or more processors; A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, the one or more processors implement the method as claimed in any one of claims 1 to 7.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program, which, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images