A calibration device and method for de-embedding a vector network analyzer

Abstract

The application discloses a kind of calibration equipment that can realize vector network analyzer deblocking, including multiple calibration pieces, the calibration piece includes straight-through piece, short-circuit piece, open-circuit piece, transmission line, material loading transmission line and PDMS polymer;The calibration equipment uses same coplanar waveguide structure, the straight-through piece, short-circuit piece, open-circuit piece, transmission line, material loading transmission line are manufactured on same low-loss substrate, with same cross-sectional dimension.This technical solution aims at improving the traditional vector network analyzer calibration technology, can better eliminate the influence of vector network analyzer and test fixture on measurement results, with the advantages of high precision, easy to manufacture, small chip occupation area, wide calibration frequency band, can realize the high-precision calibration of vector network analyzer in 100MHz-100GHz calibration frequency band.

Description

Technical Field

[0001] This invention belongs to the field of vector network analyzer testing, and specifically relates to a calibration device and method for de-embedding a vector network analyzer. Background Technology

[0002] A vector network analyzer (VNA) is a precision test instrument used to measure the scattering parameters (S-parameters) of high-frequency devices and systems, enabling accurate analysis and measurement of complex high-frequency signals and networks. Precise calibration of the VNA and the device under test (DUT) fixture is a prerequisite for high-precision measurement and characterization of devices. Calibration involves moving the reference plane from the VNA end to the DUT end to eliminate the influence of internal errors in the VNA and external errors such as those from cables, probes, and contact pads. The true performance of the device is obtained by calibrating the VNA measurement results.

[0003] Systematic errors caused by instrument imperfections are mainly described by error models. Depending on the number of receivers in the vector network analyzer (VNA), VNA systematic errors can be calibrated using either a 12-term or 8-term error model. By adding a switching error term, the 12-term model can also be solved using an 8-term algorithm. There is a correspondence between the error model, calibration algorithm, and calibration components. For example, the 8-term error model can be calibrated using the TRL or LRM calibration algorithm, with corresponding calibration components being direct-reflection-transmission line or transmission line-reflection-matched load, respectively. Furthermore, by normalizing the S_22 terms of the S-parameter matrix, the parameters of the 8-term model can be reduced to 7. When measuring a device, the errors on the left and right sides of the vector network analyzer and the device under test are represented by the error boxes X and Y in the model.

[0004] Currently, calibration methods for vector network analyzers include LRM and TRL calibration methods. LRM calibration involves load calibration components, requiring accurate load definition and symmetrical loads on both sides, a requirement difficult to meet. Manufacturing defects can also affect calibration accuracy. TRL, on the other hand, has simpler requirements for calibration components and is easier to define, giving it an advantage over LRM. However, TRL suffers from bandwidth limitations; the phase difference between the transmission line standard and the pass-through standard must be within 20 to 160 degrees. Otherwise, the transmission line and pass-through provide the same effect to the algorithm, significantly reducing calibration accuracy. All calibration components need to be fabricated on wafers. At lower operating frequencies, longer wavelengths result in longer transmission lines, occupying a larger chip area and increasing cost. Therefore, reducing the chip area occupied by calibration components and broadening the calibration frequency band warrants further research.

[0005] Therefore, how to design a calibration device and method that is highly accurate, easy to manufacture, and has a wide calibration frequency band to enable the embedding of vector network analyzers is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0006] In view of this, the present invention provides a calibration device and method that can realize the de-embedding of vector network analyzers, which can better eliminate the influence of vector network analyzers and test fixtures on measurement results, reduce the chip area occupied by calibration components and increase the width of calibration bandwidth.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] In a first aspect, the present invention provides a calibration device for de-embedding a vector network analyzer, comprising a variety of calibration components, including through-hole components, short-circuit components, open-circuit components, transmission lines, material-loaded transmission lines, and PDMS polymer; the calibration device employs the same coplanar waveguide structure, and the through-hole components, short-circuit components, open-circuit components, transmission lines, and material-loaded transmission lines are manufactured on the same low-loss substrate and have the same cross-sectional dimensions.

[0009] Preferably, the through component, short-circuit component, open-circuit component, and transmission line are bare coplanar waveguide transmission lines.

[0010] Preferably, the microfluidic channel in the middle of the material loading transport line is used to load materials with known parameters as a standard.

[0011] Preferably, the PDMS polymer is used to fabricate microfluidic channels and protective coatings.

[0012] Preferably, it also includes a probe station and a computer.

[0013] Secondly, the present invention provides a calibration method for de-embedding a vector network analyzer, comprising the following steps:

[0014] Calculate the complex permittivity of the reference material: Calculate the complex permittivity of the reference material using MATLAB for later data processing.

[0015] Calibrate the probe station and probes: Perform probe station temperature adjustment and probe leveling to calibrate the errors of the probe station and probes themselves;

[0016] Measurements were performed using calibration equipment: S-parameters of straight-through, short-circuit, open-circuit, and transmission lines were measured; S-parameters of unloaded material-loaded transmission lines were measured; and S-parameters of material-loaded transmission lines loaded with reference material were measured. All obtained S-parameters were transferred to the computer via USB data cable by connecting the calibration equipment, probe station, and vector network analyzer.

[0017] Data processing was performed using MATLAB software on a computer: the propagation constant, capacitance per unit length, inductance, conductance, and resistance data of the calibration component at each frequency point were calculated, and these data were substituted into the MATLAB mathematical model to obtain the error box parameters.

[0018] Preferably, the data processing using MATLAB software on a computer includes:

[0019] Input the basic information of the calibration equipment, and input the length of all calibration components, the length of the microfluidic channel of the material loading transmission line, and the complex permittivity of the loaded reference material in advance in the software program; calculate the complex propagation constant, distributed resistance per unit length, and distributed inductance per unit length of the bare coplanar waveguide transmission line based on the S-parameters of the through component, short-circuit component, open-circuit component, and transmission line.

[0020] Read the S-parameter file from the vector network analyzer, call the S-parameter S2P file, read the S_11, S_12, S_21 and S_22 data point by point, and pre-store them in the program array;

[0021] Obtaining the S-parameters of the equivalent model: By combining the transmission line principle, the conformal mapping theory of coplanar waveguides, and the formula for calculating the S-parameters of the equivalent model, the S-parameters of the equivalent model are calculated using the S-parameters of straight-through components, short-circuit components, open-circuit components, and transmission lines, as well as simulation results.

[0022] Obtaining error box parameters: Based on the nonlinear least squares algorithm, the residuals between the measured S-parameters and the equivalent model S-parameters are minimized using MATLAB to obtain the seven parameters of error boxes X and Y in the eight error models of the vector network analyzer.

[0023] As can be seen from the above technical solution, compared with the prior art, the present invention provides a calibration device and method for realizing the de-embedding of vector network analyzers. This technical solution aims to improve the traditional vector network analyzer calibration technology, and can better eliminate the influence of vector network analyzers and test fixtures on measurement results. It has the advantages of high precision, easy manufacturing, small chip area, and wide calibration frequency band, and can realize high-precision calibration of vector network analyzers in the 100MHz-100GHz calibration frequency band. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0025] Figure 1This is a top view of the calibration component provided by the present invention.

[0026] Figure 2 A flowchart of a calibration method for de-embedding a vector network analyzer provided by the present invention;

[0027] Figure 3 A flowchart of a data processing method using MATLAB software provided for this invention; Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Example 1;

[0030] A calibration device for de-embedding a vector network analyzer, comprising a variety of calibration components.

[0031] like Figure 1 As shown, the calibration component includes a straight-through component 1, a short-circuit component 2, an open-circuit component 3, a transmission line 4, a material-loaded transmission line 5, and a PDMS polymer 6. The performance design specifications of the calibration component are: the high-precision calibration frequency band is 100MHz-100GHz.

[0032] The five calibration components designed in this invention all employ a coplanar waveguide structure, with identical electrode cross-sections. This facilitates fabrication, and the geometry of the coplanar waveguide is compatible with the configuration of the GSG wafer probe (ground-signal-ground), simplifying calibration testing. The substrate is made of quartz, which offers low loss, low cost, and high reliability. Regarding the geometric parameters of the coplanar waveguide, the main factors affecting its performance are the center conductor width, trench width, and substrate thickness. The substrate thickness is set to 200 micrometers, the dielectric constant of quartz is 3.78, and the characteristic impedance is 50 ohms. ADS simulation yielded a center conductor-to-ground spacing of 8 micrometers, a center conductor width of 70 micrometers, and ground widths of 150 micrometers on both sides. The graphic design of the calibration component is as follows: Figure 1 The length of the straight-through component 1 is 500 micrometers. The spacing between the rod-shaped structures of the short-circuit component 2 and the open-circuit component 3 is 100 micrometers to reduce the coupling effect between the probes. The width of the rod-shaped structure is set to 250 micrometers to ensure that the probes are completely pressed on the calibration component. The transmission line 4 is 1.083 mm long, and the total length of the material-loaded transmission line 5 is 8.5 mm.

[0033] Straight-through, short-circuit, open-circuit, and transmission lines are used to extract the propagation constant, resistance per unit length, and inductance per unit length of bare coplanar waveguide transmission lines for traditional TRL calibration. Since the substrate is a low-loss material, the conductance per unit length is 0.

[0034] The material loading transmission line 5 is divided into two states: unloaded and loaded with material. The length of the PDMS polymer is 2.5 mm, and the loading position is located between two PDMS polymers. The length of the microfluidic channel is 0.5 mm, which is used to load materials with known parameters as a standard. The error box parameters are calculated by combining the measurement data of the straight-through part, short-circuit part, open-circuit part, and transmission line.

[0035] PDMS polymer 6 is a polydimethylsiloxane (PDMS) material, which is used to make microfluidic channels and protective coatings due to its hydrophobicity and inertness.

[0036] The calibration equipment for de-embedding the vector network analyzer also includes a probe station and a computer. The probe station is used to ensure accurate measurement and control and connects the vector network analyzer, probes, and calibration components. The computer is used for data processing and calculating the complex permittivity and error cell parameters of the loaded material.

[0037] Example 2;

[0038] like Figure 2 As shown, a calibration method for de-embedding a vector network analyzer includes the following steps:

[0039] The complex permittivity of the reference material was calculated using MATLAB. HFSS was used to simulate the distributed capacitance per unit length of a bare coplanar waveguide and the distributed capacitance and conductance per unit length of a transmission line loaded with the reference material, for later data processing. The complex permittivity of the liquid was calculated using relaxation equations.

[0040]

[0041] Where, ∈ ∞ Let represent the high-frequency dielectric constant, Δ∈ represent the relaxation increment, j is the complex unit, ω is the angular frequency, τ represents the relaxation time, and α represents the shape parameter.

[0042] The probe station and probes are calibrated, and the probe station temperature and probe leveling are performed. At the same time, the errors of the probe station and probes themselves are calibrated.

[0043] The S-parameters of straight-through, short-circuit, open-circuit, and transmission lines are measured using calibration equipment. The S-parameters of unloaded material-loaded transmission lines and material-loaded transmission lines with reference material are also measured. The obtained S-parameters are then transmitted to the computer via USB data cable by connecting the calibration equipment, probe station, and vector network analyzer.

[0044] S-parameters, or scattering parameters, characterize the relevant properties of microwave networks. The S-parameter matrix of a two-port network is: S11 is the reflection coefficient of port 1 when port 2 is matched; S22 is the reflection coefficient of port 2 when port 1 is matched; S12 is the reverse transmission coefficient from port 2 to port 1 when port 1 is matched; S21 is the forward transmission coefficient from port 1 to port 2 when port 2 is matched.

[0045] Data processing was performed using MATLAB software on a computer to calculate the propagation constant and unit capacitance data of the calibration component at each frequency point. The propagation constant, unit length distributed resistance, and unit length distributed inductance of the bare coplanar waveguide transmission line were extracted using the traditional TRL calibration algorithm. The calculated propagation constant and unit capacitance data were then substituted into the MATLAB mathematical model to obtain the error box parameters.

[0046] like Figure 3 As shown, the data processing using MATLAB software on a computer includes:

[0047] To input the basic information of the calibration components, the length of the line of all calibration components, the length of the microfluidic channel of the material loading transmission line, and the complex permittivity of the loaded reference material need to be entered in advance in the software program; the complex propagation constant, distributed resistance per unit length, and distributed inductance per unit length of the bare coplanar waveguide transmission line are calculated based on the measured S-parameters of the through components, short-circuit components, open-circuit components, and transmission lines.

[0048] The software program reads the S-parameter file from the vector network analyzer, calls the S-parameter S2P file, and reads the S_11, S_12, S_21 and S_22 data point by point, pre-stores them in the program array;

[0049] The S-parameters of the equivalent model are calculated. According to the transmission line principle, the relationship between the propagation constant and characteristic impedance and the distributed resistance R, inductance L, capacitance C, and conductance G can be obtained. According to the conformal mapping theory of coplanar waveguides, the distributed capacitance and conductance of the transmission line loaded with the reference material are linearly related to the complex permittivity of the reference material. Finally, the S-parameters of the equivalent model are calculated using the propagation constant, transmission line length, distributed capacitance, and conductance.

[0050] The transmission line principle states that the relationship between equivalent impedance, propagation constant, and distributed resistance R, inductance L, capacitance C, and conductance G always holds true:

[0051]

[0052]

[0053] Where Z represents the equivalent impedance, γ represents the propagation constant, ω represents the angular frequency, R represents the distributed resistance per unit length, L represents the distributed inductance per unit length, G represents the distributed conductance per unit length, and C represents the distributed capacitance per unit length.

[0054] The linear relationship between the distributed capacitance and conductance and the complex permittivity of the reference material is as follows:

[0055] C=k·ε′+k0; G=k·ε″

[0056] Where k is the slope, which is determined solely by the geometry of the transmission line and can be calculated from simulation data; ε′ is the real part of the complex permittivity; ε″ is the imaginary part of the complex permittivity; G represents the distributed conductance per unit length; and C represents the distributed capacitance per unit length.

[0057] The formula for calculating the S-parameters of the equivalent model first calculates the ABCD matrix from the propagation constant, transmission line length, distributed capacitance, and conductance, and then converts the ABCD matrix into the S-parameter matrix:

[0058]

[0059]

[0060] Where Z represents the equivalent impedance, γ represents the propagation constant, l represents the transmission line length, G represents the distributed conductance per unit length, and C represents the distributed capacitance per unit length.

[0061] The error box parameters were calculated. Based on the nonlinear least squares algorithm, the residuals between the measured S-parameters and the equivalent model S-parameters were minimized using MATLAB to obtain the seven parameters of the error boxes X and Y in the eight-term error model of the vector network analyzer. The eight-term error model of the vector network analyzer reduced the number of parameters from eight to seven by normalizing the S_22 terms of the S-parameter matrix.

[0062] The nonlinear least squares algorithm defines the residual between the calculated and measured S-parameter values ​​as...

[0063]

[0064] Where Serr is the residual, n is the total number of frequency points, Skm is the measured S-parameters, and Skc is the equivalent model S-parameters. Estimation is performed using the nonlinear least squares method, implemented with the lsqnonline function in MATLAB.

[0065] Based on the above embodiments, the working principle of the present invention is explained as follows:

[0066] The main factors affecting the S-parameters measured by the vector network analyzer are the propagation constant γ, distributed capacitance per unit length C, conductance G, resistance R, inductance L, and length l of different transmission lines. Since R and L are determined only by the geometry and properties of the electrodes, the R and L of all coplanar waveguide transmission lines in the aforementioned calibration components are identical. The R and L of the coplanar waveguide transmission lines can be obtained using the traditional TRL calibration algorithm. Based on the linear relationship between C and G and the complex permittivity of the reference material, the C and G of the transmission line with the reference material loaded can be obtained through simulation. According to the transmission line principle, the propagation constant and equivalent impedance can be calculated from R, L, C, and G. Based on the formula for calculating the S-parameters of the equivalent model, the equivalent model S-parameters can be obtained by calculating the propagation constant, transmission line length, C, and G. Finally, the residual is defined as the difference between the measured S-parameters of all calibration components and the equivalent model S-parameters. The seven parameters of the error boxes X and Y are determined by minimizing the residual at each measurement frequency point using the nonlinear least squares method.

[0067] This invention provides a calibration device for de-embedding a vector network analyzer, comprising various calibration components, including through-circuit components, short-circuit components, open-circuit components, transmission lines, material-loaded transmission lines, and PDMS polymer. All transmission lines employ the same geometric parameters and can be loaded with a reference material as a standard. Error box parameters are obtained through calibration. Simultaneously, the calibration method for de-embedding the vector network analyzer combines the transmission line principle, the conformal mapping theory of coplanar waveguides, and the equivalent model S-parameter calculation formula, and utilizes a nonlinear least squares algorithm for calibration. This method offers advantages such as high precision, ease of manufacturing, small chip footprint, and wide calibration frequency band, achieving high-precision calibration of vector network analyzers within the 100MHz-100GHz range.

[0068] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0069] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A calibration device capable of de-embedding a vector network analyzer, characterized in that, The calibration device includes various calibration components, such as through-hole components, short-circuit components, open-circuit components, transmission lines, material-loaded transmission lines, and PDMS polymers. The calibration device employs the same coplanar waveguide structure, and the through-hole components, short-circuit components, open-circuit components, transmission lines, and material-loaded transmission lines are manufactured on the same low-loss substrate and have the same cross-sectional dimensions. The microfluidic channel in the middle of the material loading transmission line is used to load materials with known parameters as a standard; The PDMS polymer is used to fabricate microfluidic channels and protective coatings; Calibration using the aforementioned calibration equipment includes: Calculate the complex permittivity of the reference material: Calculate the complex permittivity of the reference material using MATLAB for later data processing. Calibrate the probe station and probes: Perform probe station temperature adjustment and probe leveling to calibrate the errors of the probe station and probes themselves; Measurements were performed using calibration equipment: S-parameters of straight-through, short-circuit, open-circuit, and transmission lines were measured; S-parameters of unloaded material-loaded transmission lines were measured; and S-parameters of material-loaded transmission lines loaded with reference material were measured. All obtained S-parameters were transferred to the computer via USB data cable by connecting the calibration equipment, probe station, and vector network analyzer. Data processing was performed using MATLAB software on a computer: the propagation constant, capacitance per unit length, inductance, conductance, and resistance data of the calibration component at each frequency point were calculated, and these data were substituted into the MATLAB mathematical model to obtain the error box parameters.

2. The calibration device for de-embedding a vector network analyzer according to claim 1, characterized in that, The through-hole component, short-circuit component, open-circuit component, and transmission line are bare coplanar waveguide transmission lines.

3. The calibration device for de-embedding a vector network analyzer according to claim 1, characterized in that, It also includes a probe station and a computer.

4. The calibration device for de-embedding a vector network analyzer according to claim 1, characterized in that, The data processing using MATLAB software on a computer includes: Input the basic information of the calibration equipment: In the software program, input the line length of all calibration components, the length of the microfluidic channel of the material loading transmission line, and the complex permittivity of the loaded reference material in advance; Calculate the complex propagation constant, distributed resistance per unit length, and distributed inductance per unit length of the bare coplanar waveguide transmission line based on the S-parameters of the through component, short-circuit component, open-circuit component, and transmission line; Read the S-parameter file from the vector network analyzer: Call the S-parameter S2P file, read the S_11, S_12, S_21 and S_22 data point by point, and pre-store them in the program array; Obtaining the S-parameters of the equivalent model: By combining the transmission line principle, the conformal mapping theory of coplanar waveguides, and the formula for calculating the S-parameters of the equivalent model, the S-parameters of the equivalent model are calculated using the S-parameters of straight-through components, short-circuit components, open-circuit components, and transmission lines, as well as simulation results. Obtaining error box parameters: Based on the nonlinear least squares algorithm, the residuals between the measured S-parameters and the equivalent model S-parameters are minimized using MATLAB to obtain the seven parameters of error boxes X and Y in the eight error models of the vector network analyzer.

Images