RU2801297C1 - Method for dynamic calibration of mobile measuring stands in a wide frequency band - Google Patents

Abstract

FIELD: antenna measurements.

SUBSTANCE: invention relates to the measurements of the parameters of the scattering matrix to control the directivity and gain of the antennas. Essence: and the well-known method of two-port calibration of a vector network analyzer (VNA) is used, and the increase in measurement accuracy is achieved by switching to dynamic error correction of an extended moving measuring path. An extension of the analyzer model is proposed in such a way that each error parameter becomes a function of frequency, as well as a set of spatial coordinates, depending on the design and principle of operation of the measuring stand. Dynamic calibration allows you to determine the parameters of the error model and make a unique adjustment to the received data depending on the relative position of the measuring and tested antennas in space. The vector of error parameters of the model of the movable measuring stand is found, which depends on the frequency vector of the probing signal and the generalized vector of spatial coordinates, which includes information about the position of the measuring ports of the measuring and tested antennas in space. Calculation of calibration corrections is carried out for the data array coming from the VNA to a personal computer (PC) in the entire frequency band for the entire set of spatial coordinates. The method is especially useful in the case of serial measurements of the characteristics of a batch of antennas, when dynamic calibration is carried out once, and corrections are made to the results of all measurements.

EFFECT: increased accuracy of antenna measurements in a wide frequency band.

1 cl, 14 dwg

Description

SUBSTANCE: invention relates to the field of antenna measurements, mainly to measurements of scattering matrix parameters for monitoring the directivity characteristics and gain of antennas.

Antenna directivity characteristics have a significant impact on the range of electronic systems, such as high-precision navigation and control systems for unmanned aerial vehicles (UAVs), of which it is a part. The calculation of the characteristics is carried out in the process of developing the antenna, as well as when modeling it in various computer-aided design systems, which makes it possible to achieve the specified directivity parameters. Testing already manufactured antennas and verifying the compliance of the received directivity characteristics with the given ones, in a wide frequency band, is an essential part of the development of the entire system. Today, the main measurements of antenna characteristics are carried out at expensive, difficult-to-maintain antenna ranges, where they try to provide operating conditions in the far zone even for large antenna arrays. The main difficulty in conducting experimental studies of antennas in this way is the multipath nature of radio wave propagation, due to the influence of the underlying surface. As a result the connection between the measured and measuring antennas is carried out not only by a direct beam connecting their phase centers, but also by a beam reflected from the earth's surface. This factor causes the error introduced into the measurement result and, as a result, low measurement accuracy.

An alternative to polygons are specialized anechoic chambers (BEC). However, even in them, almost all elements involved in the process of measuring the amplitude and phase of the electromagnetic field in the working area become sources of errors during high-precision tests. This includes uncertainties associated with the measuring instrument, antenna positioning errors, changes in the characteristics of cables in the frequency band when moving during the test, the dependence of characteristics on environmental parameters (relative humidity, pressure and temperature), and so on. With an increase in frequency, the losses of the propagation path increase, the reflection coefficient from the elements of the path changes, and the frequency non-uniformity of its transmission coefficient manifests itself. All of these phenomena contribute to the measurement error. To determine the components (factors) of the error, the procedure for calibrating the main measuring device - a vector network analyzer (VNA) is used. To eliminate the error, mathematical correction of the measurement results is applied.

Increasing the accuracy of measuring antenna characteristics in a wide frequency band for mobile measuring stands is an urgent task.

A known method [1] to account for the influence of cables at an intermediate frequency for a movable measuring stand. In the process of its use, the difference between the reference signal from the measuring port of the fixed stand and the signal at the test output of the moving measuring port is determined. The disadvantages of the method include the difficulty of adapting the method to measuring stands based on combined measuring instruments (SI), for example, vector network analyzers (VNA), i.e. measuring instruments with local mixing of frequencies. The method allows taking into account only additional losses for signal transmission, but does not take into account the change in the reflection coefficient at the cable ports. The method is not applicable for long, many hours of testing. In addition, a multi-channel measuring signal receiver is required. All this leads to low accuracy of the calibration method, as well as the complexity and lack of automation when it is used in a wide frequency band.

There is also a method of calibration [2] based on measuring the phase error in the signal passing through a movable long cable, and then subtracting it numerically from the measurement results of the antenna characteristics. The phase delay is measured by the unmodulated reflection method from a short-circuited load located in one of the legs of a directional coupler, which is additionally introduced into the measuring setup, causing the signal to pass the flexible cable twice. The measured error is fed to one of the VNA ports and then to the computer, where it is actually taken into account when calculating the measured antenna parameters. The main disadvantage of the method is the presence of a directional coupler, which limits the dynamic range of measurements, as the authors themselves admit, as well as the impossibility of measurements in a wide frequency band. In addition, it is assumed in the calculated ratios that the characteristics of a directional coupler are ideal, and therefore, the use of a real coupler will introduce additional uncontrolled errors, both phase and amplitude, and hence reduce the measurement accuracy.

According to the technical essence, the closest to the proposed invention is a method for calibrating a two-port vector network analyzer (VNA), carried out before testing any device under study (DUT) in order to accurately measure its scattering parameters ( S -parameters), selected as a prototype (Fig. 13]. The method makes it possible to determine the components of the systematic measurement error introduced by the elements of high-frequency paths in the VNA itself, as well as by the elements of the measuring installation, including high-frequency cables located between the connection points of the ports of the DUT and the VNA. To account for the error, a mathematical correction of the measurement results is applied.

In the prototype method, it is assumed that the network analyzer model is presented as a set of devices: an ideal meter and distorting adapters (IA). In this case, the assumption is introduced that it is the IAs that are the sources of systematic measurement uncertainty, which leads to measurement errors.

The method is based on SOLT calibration and is applicable to the 12-parameter VNA model. For calibration, special load measures are used in the form of a short circuit (Short circuit, Short, SC ), idling (XX, Open, OC ), matched load (CH, Load, ML ) and a jumper adapter (A-P, THRU ) - calibration kit.

The VNA model is based on the graph representation of a two-port (2 P ) network analyzer - a meter of quadripole scattering matrix coefficients. It is modified for two cases:

- direct probing direction (Fig. 2), when the first port of the VNA is excited, and the second port is loaded on the ideal SN so that the reflection coefficient from it is equal to zero;

- the reverse direction of probing (Fig. 3), when the first port of the VNA is considered to be loaded on the ideal SN, and the second port of the analyzer is excited.

Error parameters are introduced to describe the model:

- orientation;

- interchange;

- source matching;

- load balancing;

- reflection loss;

- transmission loss,

where the indices "pz" and "oz" determine the parameters of the models for the forward and reverse directions of sounding, respectively.

These parameters are complex dimensionless quantities that depend on the frequency of the probing signal and are assumed to be constant during the measurement process. To extract the unknown parameters of the model, it is necessary to carry out independent measurements of the S -parameters ( with known measures connected to ports (KZ, XX, SN, A-P). After that, it is necessary to solve the problem using linear algebraic equations.

To determine the unknown parameters of the model in the prototype method, the following operations are performed.

1. A one-port calibration is performed for the first and second ports. At this stage, the parameters are determined, And (indices for forward and backward sounding are omitted) from the system of equations (1).

Where - complex reflection coefficient of the calibration measure, and the index "meas" indicates the measured value, and "ist" - the true corresponding to it (1-KZ, 2-XX, 3-CH).

2. The first and second ports are connected for passage through a jumper adapterTHRU, all are measuredS-options. Are determined And, in the general case, according to formulas (2); (3); (4); (5):

Where

3. If necessary, connect to the first and second ports of the CH measure. Decoupling is directly determined . This step is required when making precision measurements for small gains. Often this stage is omitted and it is considered that the port decoupling parameter is negligible: . Then the 12-parameter model is reduced to the 10-parameter one.

The true parameters of the DUT scattering matrix for the 12-parameter model are found using formulas (6); (7); (8); (9):

Where .

In modern VNAs, this procedure is standard and is carried out automatically when the difference between the true and measured values is stored in the analyzer itself and taken into account when presenting the final results.

The disadvantage of the known method is that it assumes a static measuring bench when short connecting cables are used. During antenna measurements (Fig. 4), it is necessary to place the measuring and test antennas as close as possible to the VNA. However, in real measuring installations, it is required to provide multiple long-term relative movement of these antennas. The effect of changing the characteristics of the connecting cable when it is twisted or bent on the measurement results when calibrating in a known way is not taken into account. As a result, there is an unacceptable decrease in the accuracy of antenna measurements in a wide frequency band.

The technical result of this invention is the creation of a method that does not require additional equipment and provides an increase in the calibration accuracy of a movable measuring stand in a wide frequency band.

The claimed technical result is achieved by the fact that in the known method of dynamic calibration of movable measuring stands, which consists in a two-port calibration of a vector network analyzer by first measuring the reflection coefficient from the calibration standards XX, short circuit and CH, connected in series to the measuring cables, and then all S -parameters, when the first and second ports of the analyzer are connected to pass through the calibration jumper adapter, after which the necessary error parameters of the vector analyzer model are determined and, using the measured S -parameters, the true ones are calculated, according to the claimed invention, dynamic calibration of an extended movable measuring path is carried out by expanding the vector analyzer model circuits, so that in a wide frequency band each error parameter is determined for a set of spatial coordinates in which all specified measurements of the reflection coefficient and S -parameters are made, determined by the design and principle of operation of the movable measuring stand; then, the parameter vector of the error model is determined and a unique adjustment is made to the obtained data depending on the relative position of the elements of the movable stand in space.

The task is achieved by switching to dynamic error correction of an extended mobile measuring path. An extension of the VNA model is proposed in such a way that each error parameter becomes a function of frequency, as well as a set of spatial coordinates, depending on the design and principle of operation of the measuring stand.

Dynamic calibration allows you to determine the parameters of the error model and make a unique adjustment to the received data depending on the relative position of the measuring and tested antennas in space (Fig. 4). Since the proposed method implies post-processing of data, the related measurements can be carried out both before and after the main studies of the characteristics of the DUT.

A positive effect in comparison with the prototype is achieved by the fact that the error parameter vector of the model of the movable measuring stand takes the form

Where - frequency vector of the probing signal, - a generalized vector of spatial coordinates, which includes information about the position of the measuring ports of the measuring and tested antennas in space. For measurements in the far field, this can be: a set of angular positions of the antenna under test on the turntable (Fig. 5) in a spherical coordinate system, depending on the type of positioner used, for example, in azimuth, roll or elevation. It also includes information about the current polarization of the measuring antenna. For measurements in the near field, this can be the spatial coordinates of the position of the probe carriage (Fig. 6) in Cartesian, cylindrical or spherical coordinates, depending on the positioner used, as well as information about the current polarization of the probe. Calculation of calibration corrections is carried out for the data array coming from the VNA to a personal computer (PC), in the entire frequency band for the entire set of spatial coordinates, using a specialized program containing calculations using formulas (1) to (10). The method is especially useful in the case of serial measurements of the characteristics of a batch of antennas, when dynamic calibration is carried out once, and corrections are made to the results of all measurements.

The invention is illustrated by drawings.

In FIG. 1 shows the general scheme of the static nominal calibration of a two-port vector network analyzer; in fig. 2 - graph of a two-port (2 P ) network analyzer - a meter of the coefficients of the scattering matrix of the measured device, the direct direction of probing; in fig. 3 - graph of a two-port (2 P ) network analyzer - a meter of the coefficients of the scattering matrix of the measured device, the reverse direction of probing; in fig. 4 is a diagram of the location of the measuring and test antennas relative to the VNA for the case of antenna measurements; in fig. 5 - scheme of antenna measurements in the far zone on the turntable; in fig. 6 is a diagram of antenna measurements in the near field using a near field scanner; in fig. 7 - calibration standards for full 2 P calibration with a known adapter; in fig. 8 - photograph of the tested antenna with a movable cable on a turntable; in fig. 9 - photograph of the tested antenna with a movable cable on the near field scanner; in fig. 10 - the measured modulus of the reflection coefficient from the mobile antenna in the frequency band; in fig. 11 - phase dependence reflection coefficient from a mobile antenna in the frequency band; in fig. 12 - dependence of the difference in the transmission coefficients of the tested antenna with a movable cable measured after static and dynamic calibration, from the azimuth angle of the stand for measurements in the far field; in fig. 13 - dependence of the transmission coefficient of the tested antenna with a movable cable from the frequency measured after static and dynamic calibration of the far-field stand; in fig. 14 - dependence of the transmission coefficient of the tested antenna with a movable cable on the frequency measured after static and dynamic calibration of the near-field test bench.

The proposed method for calibrating dynamic measuring stands is as follows.

To perform dynamic calibration, the reflection from the calibration standards from the corresponding calibration set is first examined (FIG. 7). The measurement takes place in the same way as in the prototype method, that is, the measurement process is repeated for existing measures of XX, short circuit and CH, connected in series to the measuring cable instead of any of the antennas. However, the difference lies in the fact that this procedure is repeated for each spatial position of the movable cable on the measuring stand in the frequency band of the probing signal. The application of this approach involves taking measures to ensure the repeatability of the measurement conditions in terms of the bench geometry. For each position, the components are determined , And (indices for forward and backward sounding are omitted) of the vector of the error parameters of the model of the mobile measuring stand for the system of equations (10).

2 P calibration is carried out on a dynamic stand, the difference from the prototype is the choice of jumper adapter. During the calibration process, the measuring ports move in space, the distance between them changes and can reach several meters or even tens of meters, which is true for a far-field stand. The jumper adapter can no longer be of small length or in the form of a coaxial adapter, as in the prototype. To ensure the electrical connection of the measurement ports, the jumper is implemented in the form of a cable assembly of considerable length. During testing, due to the bending and twisting it undergoes as it travels behind the test ports, the jumper adapter, like the test cable, will also experience similar mechanical stresses. The recommended option is to include phase-stable cables in its composition, for example, standard test cables from the meter kit (Fig. 7). They are connected to the connectors directly at the ends of the calibration plane, and a microwave cable of the required length is already connected between them, based on the configuration of the measuring stand, to ensure the required distance between the ports.

The coefficients of the scattering matrix of such a prefabricated jumper must be previously measured separately. The described approach is similar to calibration with a known adapter, as in the prototype, but differs in that the calibrated path is dynamic during the measurement process. Hence, the vectors of error factors are determined And (indices for forward and backward soundings are omitted).

If necessary, connect SN measures to the first and second ports to determine error factors (indices for forward and backward soundings are omitted). Otherwise, the assumption is made that the port decoupling parameter is negligible: . The resulting vector of error factors of the measurement stand model looks like:

To confirm the increase in the calibration accuracy of a mobile measuring stand in a wide frequency band, the proposed method carried out a number of experiments with calibrated measuring antennas. The studies were carried out using two mobile measuring stands at the Research and Production Center for Radio Information Metrology (SPC RIM) of the Moscow Aviation Institute in a certified anechoic chamber (BEC).

In FIG. 5 shows a diagram of antenna measurements in the far field on a turntable. A horn antenna from the measuring complex was chosen as the test one, which, together with a movable cable, made circular movements in azimuth on a turntable (Fig. 8). In FIG. 10 shows the dependence of the measured modulus , and in Fig. 11 shows the dependence of the phase reflectance from a mobile antenna in the frequency band from 1 to 6 GHz. The red curve corresponds to the measurement after static calibration, and the blue one after dynamic calibration. The stability of the parameter after dynamic calibration confirms the repeatability of the measurement conditions from the point of view of the bench geometry. The complex reflection coefficient measured using the proposed method corresponds to the passport data for the antenna.

In FIG. 12 shows the difference in the gain between the antennas, from the azimuth angle of the stand for measurements in the far field, when the antenna under test was connected to the analyzer through a moving path (Fig. 5). The measurements were compared after static and dynamic calibration and were carried out at a frequency of 3 GHz. The difference in the measured values reaches 0.4 dB, which has a significant impact on the accuracy of antenna tests. In FIG. 13 shows the dependencies of the transmission coefficient of the test antenna with a moving path in the frequency band from 2.5 to 6.5 GHz, measured after static and dynamic calibration on a far-field bench. The black curve corresponds to the measurement after static calibration, and the red one after dynamic calibration. Here, the difference reaches 4 dB at individual frequencies, which suggests an increase in the calibration accuracy of a movable measuring stand in a wide frequency band when using the proposed calibration method.

In FIG. 6 shows a diagram of antenna measurements in the near field on a planar movable scanner. The same horn antenna from the measuring complex was chosen as the tested antenna, which, together with the movable cable, performed linear movements in the scanning plane of the near-field scanner (Fig. 9). In FIG. 14 shows the dependence of the gain of an antenna with a moving path in the frequency band from 9.3 to 9.4 GHz, measured after static and dynamic calibration. The black curve corresponds to the measurement after static calibration, and the red one after dynamic calibration. In a bandwidth of about 10%, the difference in measurements reaches a value of more than 0.5 dB at high frequencies, which also confirms the increase in the accuracy of the calibration of a movable planar scanner.

Thus, a comparison of the measurement results of the S -parameters of the reference antennas, after static calibration by the prototype method and the proposed method of dynamic calibration, showed an increase in the accuracy of measuring the characteristics of the antenna in a wide frequency band for mobile measuring stands.

The essence of the invention lies in the fact that in the process of dynamic calibration, the VNA model is transformed in such a way that each error parameter becomes a function of frequency, as well as a set of spatial coordinates, depending on the design and principle of operation of the measuring stand.

Information sources

1. Rensburg, D.J. Millimeter Wave Near-Field Antenna Testing // Microwave Product Digest (MPD), 2010, pp. 1-6.

2. Tuovinen, J., Lehto, A. and Räisänen AV A new method for correcting phase errors caused by flexing of cables in antenna measurements // IEEE Trans. Antennas propagat. , June 1991, vol. 39, pp. 859-861.

3. Henze, A., Tempone, N., Monasterios, G. and Silva, H. Incomplete 2-port vector network analyzer calibration methods // IEEE Biennial Congress of Argentina (ARGENCON), 11-13 June 2014, pp. 810-815.

Claims (1)

A method for dynamic calibration of mobile measuring stands, which consists in a two-port calibration of a vector network analyzer by first measuring the reflection coefficient from the calibration standards XX, SC and SN, connected in series to the measuring cables, and then all S -parameters, when the first and second ports of the analyzer are connected to the pass through the calibration jumper adapter, after which the necessary error parameters of the vector analyzer model are determined and, using the measured S -parameters, the true ones are calculated, characterized in that they dynamically calibrate the extended movable measuring path by expanding the vector network analyzer model in such a way that in a wide band frequencies, each error parameter is determined for a set of spatial coordinates in which all specified measurements of the reflection coefficient and S -parameters are made, determined by the design and principle of operation of the movable measuring stand; then, the parameter vector of the error model is determined and a unique adjustment is made to the obtained data depending on the relative position of the elements of the movable stand in space.