RU2717904C1 - Method of measuring using differential sensor - Google Patents

Abstract

FIELD: measuring equipment.

SUBSTANCE: invention relates to measurement equipment and can be used for measurement using differential sensor. Differential transducer comprises at least two transducers connected to common output of transducer and provided with possibility to supply excitation signals. To rectify the output signal of the sensor, a synchronous detector synchronized with the excitation signal and configured to generate a pair of quadrature signals is used. To carry out the measurement cycle, the sensor is simultaneously supplied with the first and second excitation signals which are antiphase. Using the registration device, values of the first quadrature signal pair are recorded at the output. Then supply of the first excitation signal is cut off while maintaining supply of the second excitation signal and the value of the second pair of quadrature signals is recorded at the output. Further, on the basis of the obtained values of the first pair, a scalar unbalance value of the sensor corresponding to the measured value is found. Values of the second pair are used to compensate for errors in finding scalar value of unbalance of the sensor. Measurement cycle is performed at least once.

EFFECT: reduction of measurement error of useful signal of imbalance of differential sensor intended for excitation with high-frequency antiphase signals.

10 cl, 4 dwg

Description

The invention relates to the field of measuring technology, and more particularly, to the field of measuring differences in physical quantities using an electric differential sensor. Various types of electrical differential sensors are well known - resistive, inductive, capacitive. Their principle is based on the fact that at least two electrical transducers of the measured value are used in the sensor. These transducers are connected to a common sensor circuit in such a way that the output signal of the sensor is determined by the difference in the electrical parameters of these transducers, which is usually called the imbalance of the sensor. Thus, the output signal of the sensor is determined by the difference in the values of the magnitude acting on each of the transducers. This output signal is often called differential to emphasize the way it is formed.

Differential sensors are used for those measurements in which it is important to find the difference in magnitude, rather than its absolute value. Often, the absolute value of a quantity that acts equally on the converters has the character of an interference, against the background of which it is necessary to isolate a useful signal corresponding to the difference in magnitude. In addition, differential sensors are used to suppress external disturbing factors in the form of an electromagnetic field, mechanical vibrations, or temperature changes that equally affect the transducers and therefore are not transmitted to the differential output signal of the sensor.

Differential electric sensors are used to measure the difference in spatial position, temperature, dielectric constant, magnetic permeability and other values. Indirectly, using these values, the difference is measured between the thickness of the material, the concentration of a particular substance, the gas flow rate and many other factors. Differential electrical sensors use both direct and alternating current. Sensors with capacitive and inductive converters use alternating current, since the electrical effects of changing capacitors, inductance or mutual induction of the coils are observed when exposed to alternating current or voltage. Differential sensors are usually used to record the small response of the converters to the difference in the measured value, against the background of the large output signal of each of the converters. Specialists in the field of measurement technology know that the use of differential sensors allows you to more accurately measure the difference in the measured value than when two independent measurements of this quantity and obtaining the desired difference by numerically subtracting the results. This is due to the fact that the relative measurement errors of the large output signal of an individual converter lead to large absolute errors comparable with the values of the differential signal. Of course, the absolute accuracy of two independent measurements can be improved by a more sophisticated measuring device. However, when using similar converters, a differential sensor is more preferable.

AC sensors can be divided into at least three classes. The first class includes the so-called bridge sensors. The bridge sensor circuit is powered by a single source of alternating voltage or current, and the output signal is removed between two points in the circuit. An example of a bridge inductive sensor is given in US patent US 8026716 (publ. 01.10.2009, IPC G01N 27/72). The bridge connection scheme is constructed in such a way that when the converters have the same electrical characteristics and there is no sensor imbalance, the output signal is also absent. The output signal appears when an imbalance occurs, which, in turn, is due to differences in the measured value acting on the transducers.

The second class includes sensors, which are also powered by a single source of alternating voltage or current, and in which the converters are electrically connected in such a way that their output signals are subtracted from the common sensor output. This class mainly includes sensors with inductive coupling between the coils of the transducers, since the polarity of the response of such a transducer can be easily reversed by changing the direction of winding of its coil. The converters are connected counter-series or counter-parallel, which provides subtraction directly in the conductors, which are made of the coil. An example of such a sensor with an in-series connection is given, for example, in US Pat. No. 6,995,021 (publ. 02.09.2004, IPC G01N 27/72).

The third class can conditionally be called the class of sensors with antiphase excitation. Such a sensor is supplied with currents or voltages of the same frequency and amplitude, but opposite in phase. Antiphase currents or voltages are supplied to individual converters, and the output signals of the converters are summed up in their connection diagram. As a result of the summation, the total output signal of the sensor is formed.

Due to the antiphase relationship between the exciting signals, in fact, the output of the transducers is mutually subtracted from the output signals of the transducers. When the measured value of the same value is applied to the transducers, mutual subtraction leads to the complete compensation of the transducer signals, so that the sensor output signal is absent. If the measured values acting on each of the transducers have a difference, then, with mutual subtraction, the transducer signals do not fully compensate each other, and a signal proportional to the specified difference appears at the output of the sensor. This class, for example, includes an inductive differential sensor for the number of ferromagnetic particles on the sheet material described in patent RU 2675405 (publ. 12/19/2018, IPC G01D 5/12). In it, antiphase currents are supplied to two exciting coils of the inductive sensor, and the output signal is removed from the receiving coil, which is magnetically coupled to both exciting coils. Magnetic particles change the coupling coefficient between the exciting and one of the receiving coils, which leads to a change in the output signal of the sensor. The output signal of the differential sensor of the alternating current arises due to subtraction (i.e., summation in antiphase) of the output signals of the converters. In some sensors, electronic circuits containing active elements, such as transistors or operational amplifiers, and sometimes rectifiers or pulse generators, are used for summing. The electronic circuit introduces its own errors due to its inherent asymmetry in input, nonlinearities, limitation of the dynamic range, frequency and phase distortions, as well as interference and noise. Sensors in which the transducer signals are summed passively in a physical medium that is highly linear, almost does not introduce phase and frequency distortions, has a low noise level and no level limitation are deprived of this drawback. Such a medium may be an electric conductor or the space inside the coil. Passive summation (subtraction) is used in the already mentioned sensors: according to the patent RU 2675405 the summation of the magnetic flux occurs in the receiving coil, and according to US patent US 6995021 antiphase subtraction occurs in the series connection of the conductors of the coils.

In many cases, in order to increase the sensitivity of sensors, they try, as far as possible, to increase the frequency of the exciting voltage or current. This is due to the fact that the response of the sensor output signal to changes in inductance or capacitance, in the general case, increases with increasing frequency. For the most highly sensitive sensors, the operating frequency can exceed one hundred megahertz. Another reason for increasing the frequency of the exciting signal is the desire to increase the speed of the sensor.

Differential sensors are widely used for highly sensitive measurements. A change in capacitance or inductance corresponding to the full sensitivity range of the sensor is often within a few hundredths of the nominal value of the capacitance or inductance of the converter. Processing the output signal of such a sensor is fraught with a number of problems, among which the main ones are the suppression of noise and interference, and adequate compensation for spurious imbalance of the sensor. A high degree of suppression of noise and interference is usually achieved by accumulating the measured signal over many periods of oscillation of the exciting signal. The higher the frequency, the greater the number of periods over which you can accumulate the measured signal for a given time.

In many cases, the measurement process is a sequence of measurement cycles, in each of which a separate measured value is found. When measuring a variable over time, the sequence of measurement cycles allows you to build the dependence of the measured value on time. If the measured value is static, then averaging the results obtained in several measurement cycles allows to reduce the contribution of noise to the measurement result. So, when exposed to white noise, its contribution to the measurement result decreases in proportion to the square root of the number of measurements.

The phenomenon of spurious imbalance is well known for all types of differential sensors, and is usually associated with the difference in the electrical characteristics of the transducers when they are exposed to the same measured value. The cause of parasitic imbalance, as a rule, is the geometric difference between the transducers within the design tolerance of their manufacture, or the inaccuracy of the installation of the converter. For example, for capacitive sensors, the source of spurious imbalance is often a small difference in the area of the plates of the measuring capacitors, or a small skew of these plates. Also, spurious imbalance arises due to differences in the characteristics of the materials from which the transducers are made, for example, dielectric or magnetic permeability.

The occurrence of imbalance is the principle of operation of a differential sensor, and, in itself, is not harmful. In the overwhelming majority of cases of using a differential sensor, from a technical point of view, it is necessary to finally obtain an imbalance corresponding to the scalar difference in the amplitudes of the output signal of the converters. In this description, we will talk about a useful imbalance when it comes to the response of a differential sensor to a measured quantity. Useful imbalance is opposed to parasitic imbalance, which is added to the useful imbalance and impairs the accuracy of the sensor.

To combat spurious imbalance, before the measurement, when the transmitters are affected by the same measured value or there is no measured value, the so-called sensor balancing is carried out. To do this, register the sensor output signal and then subtract it from the obtained measurement result. Or, before the measurement, the parameters of the sensor elements or its switching circuits are promptly changed in order to compensate for the parasitic imbalance due to the intentional introduction of small asymmetries.

However, in the case of a differential sensor excited by a high-frequency voltage or current, the fight against spurious imbalance is complicated. This is due to the fact that, as a rule, the phase difference between the antiphase exciting signals arriving at the transducers is not exactly 180 degrees. This is due to the imperfect design of the sensor, for example, due to the different lengths of the lead-in conductors and different phase shifts in the circuits along which exciting signals are transmitted. In addition, the converters themselves may have slightly different phase characteristics, which leads to a phase difference of 180 degrees when summing the output signals. All these effects are almost imperceptible at low excitation frequencies. However, at high excitation frequencies, starting from about 1 MHz and above, the phase relationships become significant for the accuracy of the sensor. At a frequency of about 100 MHz, as practical experience shows, the phase shift between the output signals of the converter can reach 20 degrees even with the adoption of special design measures to reduce the phase difference. As a result of a combination of these random factors, the output signal due to spurious imbalance may have an arbitrary phase shift with respect to the excitation signal. To compensate for it by subtracting the signal corresponding to spurious imbalance from the signal received during the measurements, vector scalar rather than scalar subtraction should be used.

The described mechanism for the occurrence of spurious imbalance is shown in FIG. 1, which shows vector diagrams of the formation of the differential output signal of the sensor. Vector diagrams allow you to visualize the formation of the output signal of the sensor in the form of a vector sum of exciting signals that have passed through the converters and, at the same time, in a certain way changed their amplitude and phase. For definiteness, in this description we will apply the concept of unbalance magnitude only to scalar values. To describe the vector quantity resulting from the imbalance phenomenon, the term “mismatch vector” will be used. We point out here that the ultimate goal of measurements with a differential sensor is always a scalar imbalance, with the exception of the rare special case of measuring the complex permeability of a dielectric or magnetic material. The mismatch vector, as a rule, has value only for intermediate calculations. It will be shown below that, for high-frequency sensors, the imbalance cannot be considered equal to the modulus of the mismatch vector.

In many cases, the output of the sensor transducers is further filtered and amplified, so that the ac signal processing circuitry is an integral part of the sensor. Then, the imbalance and the vector of the output signals must take into account the changes in the amplitude and phase introduced by the named processing scheme.

The sensitivity of the sensor can be expressed as the ratio of the useful imbalance to the measured value or to the measured differential value.

соответствует выходному сигналу первого преобразователя в отсутствии измеряемой величины, а вектор

представляет собой сигнал второго преобразователя. Вектор

соответствует сигналу первого преобразователя при появлении измеряемой величины. В отсутствие измеряемой величины выходной сигнал датчика соответствует вектору паразитного рассогласования

Выходной сигнал датчика при появлении измеряемой величины соответствует вектору

Полезный сигнал датчика

нельзя измерить непосредственно, но он получается вычитанием из выходного сигнала вектора паразитного рассогласования:

In FIG. 1a and FIG. lb, vector

corresponds to the output signal of the first converter in the absence of a measured quantity, and the vector

represents the signal of the second converter. Vector

corresponds to the signal of the first converter when the measured value appears. In the absence of a measurable quantity, the sensor output signal corresponds to the parasitic mismatch vector

The output signal of the sensor when the measured value corresponds to the vector

Useful Sensor Signal

cannot be measured directly, but it is obtained by subtracting the parasitic mismatch vector from the output signal:

и является паразитным, а при измерении он равен

и является скалярной суммой паразитного дисбаланса и полезного скалярного значения дисбаланса, вызванного измеряемой величиной. Тогда, полезный дисбаланс легко получить как скалярную разность модуля вектора выходного сигнала датчика в момент измерения и модуля вектора паразитного рассогласования:

Такой подход к компенсации паразитного дисбаланса наиболее широко применяется на практике. Однако, если вектора не коллинеарны, подобное упрощение недопустимо.In FIG. Figure 1a shows a degenerate case characteristic of a low excitation frequency at which phase shifts can be neglected. Due to this, all vectors are collinear, and you can restrict yourself to a scalar solution to the problem, without involving vectors. The scalar value of the imbalance in the absence of the measured value is calculated as

and is parasitic, and when measured it is equal to

and is the scalar sum of the parasitic imbalance and the useful scalar value of the imbalance caused by the measured quantity. Then, a useful imbalance can be easily obtained as the scalar difference of the module of the vector of the sensor output signal at the time of measurement and the module of the parasitic mismatch vector:

This approach to compensation for spurious imbalance is most widely used in practice. However, if the vectors are not collinear, such a simplification is unacceptable.

в отсутствие измеряемой величины можно представить суммой

, где

является значимым компонентом, а

является незначимым. Значимый компонент направлен по биссектрисе

угла

между векторами выходных сигналов преобразователей и образует с ними малый угол. Поэтому, значимый компонент практически равен по величине скалярному значению дисбаланса. Незначимый компонент почти ортогонален векторам выходных сигналов преобразователей и практически не изменяется при изменении дисбаланса.As the frequency increases, significant phase shifts occur between the output signals of the converters in the sensor, which cannot be ignored. Therefore, the problem should be solved in vector form. This case is shown in FIG. lb. The output vector of the sensor can be decomposed into orthogonal components, one of which can conditionally be called significant, and the other insignificant. For example, the output vector

in the absence of a measurable quantity can be represented by the sum

where

is a significant component, and

is insignificant. The significant component is directed along the bisector

angle

between the vectors of the output signals of the converters and forms a small angle with them. Therefore, a significant component is almost equal in magnitude to the scalar value of the imbalance. An insignificant component is almost orthogonal to the vectors of the output signals of the converters and practically does not change when the imbalance changes.

значительно превысит по модулю значимый компонент

Наличие угла α≤20° между сигналами отдельных преобразователей

дает незначимый компонент

Note that a seemingly small phase shift between the output signals of the converter, not exceeding 20 degrees, can lead to the fact that an insignificant component

significantly exceed in magnitude the significant component

The presence of an angle α≤20 ° between the signals of the individual converters

gives an insignificant component

несмотря на достаточно малую величину sinα≤0,35.The signal amplitude of each converter in a differential sensor is many times greater than the difference in the amplitudes of the converters, which determines the typical ratio

despite the rather small value of sinα≤0.35.

From the above explanation, it becomes clear that, with increasing frequency, the sensor output signal module, by itself, should not be used to obtain a scalar imbalance value, since the main contribution to it is usually made by an insignificant component. It is possible to separate a significant component from an insignificant only due to phase-sensitive detection of the sensor output signal, which is implemented in the already mentioned patents US 6995021, US 8026716 and RU 2675405.

Вслед за этим, из вектора полезного сигнала извлекают значимый компонент

Например, в упомянутом выше патенте США US 8026716 описан дифференциальный индуктивный датчик для измерения малых концентраций вещества, обладающего магнитными свойствами. В этом датчике, с целью компенсации паразитного дисбаланса, синтезируют компенсирующий сигнал, равный по модулю и в точности противонаправленный сигналу паразитного рассогласования, и вводят его в схему моста для суммирования с выходным сигналом. В результате аналогового вычитания переменных сигналов, эквивалентного векторному вычитанию, выходной сигнал датчика представляет собой чистый полезный сигнал. Этот сигнал подвергают синхронному детектированию и, в результате, получают постоянное напряжение, которое соответствует значимому компоненту. Постоянное напряжение соответствует скалярному значению дисбаланса моста с устраненным паразитным дисбалансом.In the prior art, the elimination of spurious imbalance when operating at high frequency is sometimes implemented by subtraction in a vector form to obtain a vector of a useful signal:

Following this, a significant component is extracted from the vector of the useful signal

For example, the aforementioned US Pat. No. 8,026,716 describes a differential inductive sensor for measuring low concentrations of a substance having magnetic properties. In this sensor, in order to compensate for stray imbalance, a compensating signal is synthesized that is equal in magnitude and exactly opposite to the stray mismatch signal, and introduced into the bridge circuit for summing with the output signal. As a result of analog subtraction of variable signals equivalent to vector subtraction, the sensor output signal is a pure useful signal. This signal is subjected to synchronous detection and, as a result, a constant voltage is obtained, which corresponds to a significant component. The constant voltage corresponds to the scalar value of the unbalance of the bridge with the eliminated spurious imbalance.

и, затем, вычитая их друг из друга, как скалярные величины. Так, в патенте RU 2483276 (опубл. 27.05.2013, МПК G01B 7/06) описан емкостной дифференциальный датчик для измерения различия в толщине листового материала, возбуждаемый противофазными напряжениями высокой частоты. Выходной сигнал с датчика усиливается и подается на синхронный детектор. Применение синхронного детектора позволяет сразу же выделить компонент вектора выходного сигнала, имеющий заданное направление. Результатом измерения является скалярная величина дисбаланса, получаемая после оцифровки выходного сигнала синхронного детектора. В этом решении запоминают паразитный дисбаланс, полученный перед проведением измерений, и скалярным образом вычитают его из дисбаланса, полученного при проведении измерений.Spurious imbalance can be eliminated in another way, by removing significant components

and then subtracting them from each other as scalar quantities. So, in patent RU 2483276 (publ. 05.27.2013, IPC G01B 7/06) a capacitive differential sensor for measuring the difference in the thickness of the sheet material, excited by antiphase high-frequency voltages, is described. The output signal from the sensor is amplified and fed to a synchronous detector. The use of a synchronous detector allows you to immediately isolate the component of the output signal vector having a given direction. The measurement result is a scalar imbalance value obtained after digitizing the output signal of the synchronous detector. In this decision, the parasitic imbalance obtained before the measurements is stored, and it is subtracted from the imbalance obtained during the measurements in a scalar manner.

In both of the solutions described here, the accuracy of the measurement depends on the extent to which the direction in which the significant component of the sensor output signal is extracted corresponds to the real direction of the significant component. A significant difference between these directions leads to a decrease in the sensitivity of the sensor, since only part of its output voltage is used in the measurements. The actual direction of a significant component cannot be obtained from the output of a balanced sensor. Therefore, it is determined during the manufacture of the device by adjusting or preliminary measuring the phase incursion in the measurement circuit. In addition, these decisions rely on the immutability of the transducers' sensitivity to the measured value over time.

The solutions used in these patents are usually effective if disturbing factors do not act on the sensor, which lead to a synchronous change in the response of the transducers. An example of such factors can be called vibration, which leads to the same change in the distance between the plates of the measuring capacitors of the capacitive differential sensor, which causes a synchronous change in the capacitance of the capacitors. In inductive differential sensors, the vibration of some ferromagnetic structural element of the device located in the far zone of the sensor sensitivity region also leads to a synchronous change in the inductance of the transducers. The instability of the sensor supply voltage is another type of disturbing factor that simultaneously affects both transducers. For differential sensors used to control the differential of certain material properties, balancing is performed in the absence of material, and the measurement is performed after the material is introduced into the sensor. The introduction of material into the sensor leads to a synchronous change in the response of the transducers. For example, when a uniform sheet material is introduced into the gap of a capacitive sensor, the capacitance of the measuring capacitors increases simultaneously. If the design of the sensor involves its separation into parts and reverse articulation for loading the measured material or for maintenance, then after such an operation the characteristics of the transducers of the sensor slightly change, usually equally.

В силу конструктивной идентичности преобразователей наборы их коэффициентов зависимости s_i тоже можно считать идентичными.The parameters of inductive and capacitive converters, such as inductance, inductive coupling coefficient, capacitance or current flowing, are differentiable functions of many independent variables, which represent both the measured value and disturbing influences. So, for example, the capacitance of the measuring capacitor is proportional to the area of its plates and the dielectric constant of the material between the plates, and inversely proportional to the distance between the plates. The current through the capacitor, in a similar way, depends on the same variables, and, in addition, is proportional to the amplitude and frequency of the applied voltage. That is, the parameter of the converter, for example, the current flowing through it, can be approximately represented through a certain initial value and the sum of the changing components proportional to the change Δ of each individual variable X _i : I≈I ₀ (1 + ∑s _i ΔX _i ). For small changes ΔX _i, this linear expression can be considered almost accurate. The coefficient of dependence of s _i on changes in a certain variable can be expressed in terms of the partial derivative of the converter parameter with respect to this variable X _i :

Due to the structural identity of the converters, the sets of their dependence coefficients s _i can also be considered identical.

Вектор паразитного рассогласования, который запоминается на этой стадии измерений, равен

К моменту начала измерений эти вектора пропорционально увеличиваются под воздействием возмущающего фактора и превращаются в вектора

а соответствующий вектор паразитного рассогласования превращается в

При попытке получить полезный сигнал вычитанием ранее запомненного паразитного рассогласования возникает неверный результат:

На самом деле, верным является выражение

но действующее в момент измерения значение паразитного рассогласования

недоступно для его использования в вычислениях. Отсюда возникает погрешность нахождения выходного сигнала, которая выражается вектором изменения паразитного рассогласования

Длина этого вектора часто превосходит длину вектора полезного сигнала

The influence of the same perturbing factor X _i on both transducers leads to scaling of the output signal vector of each transducer in accordance with the coefficient of dependence of s _i on this factor. That is, in fact, the entire vector diagram is scaled. As a result of this influence, the phenomenon of spurious imbalance during measurement cannot be suppressed by subtracting the previously measured initial value, neither in scalar nor in vector form, since the parasitic mismatch vector itself also scales. In FIG. 1c shows a vector diagram in which the output signals of the sensors, before taking measurements, in the absence of a measured quantity, also correspond to vectors

The parasitic mismatch vector, which is stored at this stage of the measurements, is

By the time the measurements begin, these vectors proportionally increase under the influence of a disturbing factor and turn into vectors

and the corresponding parasitic mismatch vector turns into

When you try to get a useful signal by subtracting a previously stored spurious mismatch, an incorrect result occurs:

In fact, the expression is true

but the value at the time of measurement of the parasitic mismatch

not available for use in calculations. Hence the error in finding the output signal, which is expressed by the vector of variation of the parasitic mismatch

The length of this vector often exceeds the length of the useful signal vector.

In the most difficult cases, the disturbing factor also affects the phase of the output signals, as shown in FIG. 1d. This behavior is characteristic of sensors in which, to combat induced interference, the sensor output is connected to a frequency-selective resonance circuit, for example, an LC circuit. A similar solution is used, for example, in patent RU 2483276. The capacitance and / or inductance of the converters in such a circuit are included in the LC circuit. When an external disturbing factor leads to a change in the capacitance and / or inductance of the converter, the resonant frequency of the circuit shifts from the original value. If the circuit has a high Q factor, then even a small shift in the resonant frequency leads to a significant rotation of the phase of the output signal. This rotation concerns all components, and changes both the direction of the parasitic mismatch signal and the useful signal. A similar effect occurs when a specially created resonant LC circuit is absent, but stray capacitances and inductances in the sensor form a resonant circuit. Such an effect is described, for example, in patent US 8026716.

It can be seen from the foregoing that the parasitic mismatch vector can change during the time between balancing and measurements not only in length, but also in direction. Along with spurious mismatch, the direction of the useful signal also changes. Since both the parasitic mismatch and the direction of the useful signal, at the time of measurement, are hidden parameters, the exact reconstruction of the scalar value of the useful imbalance is not possible.

Another difficulty when using the sensors described in RU 2483276 and US 8026716 to measure a small difference in magnitude is associated with the influence of noise and interference. In many cases, the use of a capacitive or inductive sensor, the electromagnetic environment at the place of its installation leads to interference at the output, which introduce a significant measurement error. The random noise vector has no distinguished direction. As is known from the theory of electrical measurements, when a component of a random noise vector corresponding to a given direction is extracted, the result does not depend on this direction. If the direction of the significant component of the output signal at the time of the measurement is unknown or only approximately known, this can reduce the useful signal obtained, despite the fact that the noise level does not depend on the direction. That is, the lack of accurate information about the direction of a significant component can significantly reduce the signal-to-noise ratio compared to the best possible.

As follows from the above reasoning, to extract, with the smallest errors, a significant component of the sensor output signal corresponding to the measured value, it is necessary to obtain additional data. Namely, at the time of the measurement, it is necessary to know both the direction of the significant component and the transmission coefficient of the sensor. This data is required to take into account the synchronous change in the response of the sensor transducers and changes in the direction of a significant component of the sensor output signal, which occur under the influence of external disturbing factors. Under the transmission coefficient refers to a purely electrical characteristic, connecting the values of the input and output signal. In relation to the sensor, it connects the magnitude of the excitation signal and the output signal.

Patent RU 2483276 was selected as a prototype of the claimed invention. The named patent describes a multi-element differential sensor, which contains many measuring capacitors, united by a common plate, which is used as the output of the sensor. The measurement process includes a sequential search of several measurement modes, each of which involves only a part of the measuring capacitors. This makes it possible to measure differences in thickness for the area of the sheet in which the involved measuring capacitors are located. Enumeration of modes allows you to measure all areas in the sensor sensitivity zone. In general, multi-element differential sensors allow simultaneous measurement of differences in magnitude in several places or on several channels. They are a simpler alternative to using many conventional differential sensors.

The disadvantage of the prototype is that it does not allow, in the course of measurement, to obtain relevant information about the direction of a significant component of the output signal of the sensor and the transmission coefficient of the sensor. Without the above information, it is not possible to reliably measure the amplitude of the useful signal and compensate for stray imbalance. This leads to poorly removable measurement errors when exposed to such a disturbing factor as the vibration of the plates of the measuring capacitors. However, even in the absence of vibration, the spread of the phase characteristics of the sensor and circuit elements leads to the fact that the real direction of the significant component of the output signal differs significantly from the calculated one, as a result of which the sensor sensitivity is lower than the calculated one. For the same reason, in the prototype the signal-to-noise ratio is significantly lower than theoretically possible.

The prototype places high demands on the factory setting of circuit elements, as this setting allows you to adjust the calculated direction of registration of a significant component, and bring it closer to the real direction of a significant component. Inaccuracies in the adjustment significantly reduce the sensitivity and signal-to-noise ratio.

Another drawback is that, for each individual mode, the true direction of the significant component is individual due to the different phase incursion in the circuit of each of the measuring capacitors. The device does not imply the ability to quickly change the direction of registration of a significant component when switching from one mode to another. This leads to a large spread in sensitivity and signal-to-noise ratio between different modes.

In the prior art, solutions are known in which the so-called common-mode output signal, which is formed together with the usual differential output signal of a differential sensor, is used as additional data. A common-mode signal, in contrast to a differential signal, is not a difference, but the sum of the responses of the transducers of the differential sensor. So, in patent RU2289789 (publ. 20.12.2006, IPC G01C 19/5712), the differential signal is detected by a synchronous detector, and the common-mode signal is passed through an amplitude detector. The value of the rectified common-mode signal is used to normalize the output signal of the synchronous detector. Thus, it is possible to compensate for the influence of disturbing factors on the amplitude of the output differential signal, but the influence of these factors on the phase of the output signals of the converters is not compensated and continues to be a source of errors. In patent RU 2514158 (publ. 04/27/2014, IPC G01B 7/00) a differential signal is obtained by mutually subtracting the rectified signals from two converters, and the common-mode signal is obtained by adding them. Further, the common mode signal is used to correct the level of the differential signal. This solution is also characterized by ignoring the influence of disturbing factors on the phase of the output signals of the converters. Both of these patents refer to low frequencies at which the phase effects are quite small and create small errors. With increasing frequency, such technical solutions cease to be effective.

Another important feature of the solutions according to the patents RU 2289789 and RU 2514158 is that the differential and common-mode signals in them are obtained by processing individual signals of the converters in an electronic circuit with active elements. Such an electronic circuit, as already indicated, introduces numerous intrinsic errors. Sensors with passive summation are deprived of such errors, examples of which were given earlier. Another example is given in the prototype, where the current of the measuring capacitors is passively summed in the receiving plate common to all capacitive converters. The flip side of passive summation is the inability, without additional efforts, to receive a common-mode signal simultaneously with a differential signal.

The already mentioned patent RU 2675405 uses passive summation and a special method of phase-sensitive detection. It uses a simple amplitude detector, but the sensor is intentionally tuned with a large value of spurious imbalance. Due to this, the output voltage of the sensor is the sum of the differential signal and a more powerful excitation signal that passed through one of the converters. This excitation signal, which has passed through one of the converters, synchronizes the switching of the diode in the amplitude detector due to the significant amplitude, so that synchronous detection is actually provided for the differential signal. The constant part associated with a powerful synchronizing signal is subtracted from the output current of the amplitude detector, and the remainder of the current is a synchronously rectified differential signal. This solution provides the passage of the differential signal and the synchronizing signal along the same conversion path, thereby compensating for changes in phase incursion. The resulting rectified differential signal is a significant part of the sensor mismatch vector. However, in such a solution it is impossible to measure and take into account the change in the transmission coefficient, which does not allow to compensate for the changes in the sensitivity of the sensor associated with this due to external disturbing factors.

The technical result of the claimed invention is to reduce the measurement error of the useful signal of the imbalance of the differential sensor, designed to excite out of phase high frequency signals. An additional technical result is the reduction of the measurement error arising from the scatter of the parameters of the sensor elements and its switching circuit. These results are achieved due to the fact that in the method of measurement using a differential sensor designed to excite high-frequency antiphase signals, the differential sensor contains at least two transducers that are connected to the common output of the sensor and are configured to supply signals to them excitement; to rectify the output of the sensor, a synchronous detector is used, synchronized with the excitation signal and configured to generate a pair of quadrature signals, where the first and second excitation signals, which are out of phase, are fed to the sensor at the same time, and the values of the first pairs of quadrature signals, then, turn off the supply of the first excitation signal while maintaining the supply of the second excitation signal, and register and the output value of the second pair of quadrature signals, after which, on the basis of the obtained values of the first pair, find the scalar value of the sensor imbalance corresponding to the measured value, and the values of the second pair are used to compensate for errors in finding the scalar value of the imbalance of the sensor, and carry out a measurement cycle, at least , once.

The differential output resulting from the measurement is recorded as the first pair of quadrature signals. In general, the use of quadrature signals allows the use of vector calculations in the processing of sensor signals.

Obtaining additional data in the claimed invention is implemented by turning off the supply of one of the antiphase excitation signals, which is quite simple to implement in practice, for example, by blocking the shaper of the corresponding excitation signal. In the absence of the first of the excitation signals, the second excitation signal will pass to the input of the synchronous detector, taking into account the total phase shift and the transfer coefficient all the way from the signal generator. It will be recorded as a second pair of quadrature signals.

The ultimate goal of measurement is to obtain the scalar value of the imbalance of the sensor. The output signal of the sensor is formed as the vector sum of the output signals of the transducers of the sensor. The first pair of quadrature signal values represents the coordinates of the sensor error signal vector taking into account additional amplification and phase shift along the path from the sensor output to the input of the synchronous detector. This vector consists of a significant component and an insignificant component.

Measurements are carried out in the form of one or many cycles. The use of multiple measurement cycles allows you to build the time dependence of the measured value, or to reduce the contribution of noise to the measurement result by averaging the results obtained in separate cycles. In the case of using a multi-element sensor containing more than two transducers, several measurement cycles can be used to obtain several measurement results with the supply of exciting signals, in each cycle, only to a part of the sensor transducers.

The second pair of quadrature signal values represents the coordinates of the output signal vector of one sensor transducer, taking into account additional amplification and phase shift along the path from the sensor output to the input of the synchronous detector. In the claimed invention, this pair of values has a dual function. Firstly, it allows, in real time, to indicate the direction for highlighting a significant component of the sensor mismatch vector, since this direction forms a small angle with the output signal vector of one sensor transducer. A small directional error provides close to the maximum possible values of the sensor sensitivity and signal-to-noise ratio. Secondly, the second pair of values allows you to quickly measure the transmittance of the signal along the same path from the generator to the input of the synchronous detector, through which the differential output signal of the sensor mismatch passes. This allows, in real time, to amend the value of the measured scalar value of the imbalance of the sensor to compensate for changes in the specified transmission coefficient over time. A change in the transmission coefficient due to disturbing factors, without compensation, affects the sensitivity of the sensor to the measured value. That is, the registration of the values of the second pair allows you to select a significant component from the values of the first pair with small errors in the direction of its selection, and compensate for errors in its length in order to avoid the influence of the spread of element parameters, inaccuracies in the factory settings and disturbing influences. This leads to increased measurement accuracy, ensuring the achievement of both the main and additional technical result.

It is important to note that the second pair of quadrature signals has large absolute values, in comparison with the small differential signal obtained from the first pair of quadrature signals. This is due to the very principle of a differential sensor, which measures a weak level of difference between the output signals of the converters against the background of a high signal level of an individual converter. Therefore, the values of the second pair are almost not affected by noise and interference, and can be recorded with high accuracy. This feature makes an additional contribution to reducing the error of the claimed method.

The direction of the significant component of the sensor unbalance vector, in the claimed invention, is approximately determined by the direction of the signal vector of one of the converters corresponding to the second pair of quadrature signals. Such an approximation is valid, since the angle between the signal vectors of the first and second converters, in real designs, is small and does not exceed 20 degrees. Accordingly, the angular approximation error is less than 20 degrees.

Косинус угла ошибки определения направления, меньшего 10 градусов, весьма близок к единице, что соответствует небольшой погрешности измерения амплитуды полезного сигнала, не превышающей 0,015.Vector analysis involves well-known methods for finding the length of a component of a vector when the direction of this component is known. In one of these methods, the projection of a vector onto a known direction is calculated. In another method, the scalar product of a vector and a unit vector directed in a known direction is calculated. In all these methods, the result is determined by the cosine of the angle between the vector and the known direction of its component. If the direction is not determined accurately, then the error resulting from this is proportional to 1 - cosδ, where δ is the angular error in determining the direction. In fact, we are talking about half the angle formed by the vectors of the output signals of the individual converters, since the significant component is directed along the bisector of this angle. In this way,

The cosine of the angle of error in determining the direction of less than 10 degrees is very close to unity, which corresponds to a small error in measuring the amplitude of the useful signal, not exceeding 0.015.

In the future, we will call the output a vector formed on the basis of the values of the first pair of quadrature signals. The sensitivity vector we will call the vector reflecting the direction of the vector of the useful signal, and the sensitivity of the sensor transducers. For example, a vector formed based on the values of the second pair is a sensitivity vector. Its direction almost coincides with the direction of the significant component of the output vector, and its modulus is the greater, the higher the sensitivity of the converters. To find the imbalance of the sensor, find the projection length of the output vector on the direction of the sensitivity vector. Such a projection excludes from consideration the insignificant component perpendicular to the sensitivity vector, which does not carry useful information about the imbalance. The imbalance resulting from the projection, in the general case, is the sum of the significant components of the parasitic imbalance and the useful imbalance.

The length of the sensitivity vector is used to compensate for the time variation of the sensitivity of the sensor transducers. When an external perturbing effect is applied to the transducers of the sensor, resulting in a synchronous change in the characteristics of these transducers, the sensitivity vector module changes with them. The magnitude of the parasitic imbalance and the magnitude of the useful imbalance under the influence of an external disturbance change to the same extent. Thus, taking the sensitivity vector into account allows one to find the influence of the disturbing effect and compensate for the change in the parasitic imbalance.

In possible implementations of the invention, the second excitation signal is additionally turned off while maintaining the first excitation signal and the values of the third pair of quadrature signals are recorded, and the values of this pair are used to compensate for errors in finding the imbalance of the sensor. Preferably, the sensitivity vector of the sensor is formed as the difference of vectors corresponding to the values of the second and third pairs of quadrature signals. That is, the sensitivity vector turns out to be directed along the bisector of the angle formed by two vectors corresponding to the values of the second and third pairs of quadrature signals. This approach allows to increase the accuracy of compensation due to the orientation of the sensitivity vector in the direction of the significant component of the output vector, and also by taking into account the sensitivity changes of both transducers.

The sensitivity vector thus formed will be called in-phase. The differential response of the sensor is excluded from the common-mode sensitivity vector, almost completely. Thus, the length of the in-phase sensitivity vector characterizes the transmission coefficient much more accurately on the way from the generator to the input of the synchronous detector than the length of the sensitivity vector, constructed on the basis of only the values of the second pair of quadrature signals.

Before measuring, balance the sensor. The balancing of the sensor consists in recording quadrature signals and calculating on their basis the initial sensitivity vector and the initial imbalance as the projection of the differential output signal vector on the direction of the sensitivity vector. During balancing, the resulting initial imbalance is a stray imbalance. During the subsequent measurement, a complete measured imbalance and a measured sensitivity vector are obtained. The module of the sensitivity vector between the moments of balancing and the moment of measurement changes under the influence of a disturbing factor. The sensitivity measurement coefficient, equal to the ratio of the measured sensitivity vector module to the initial sensitivity vector, reflects both the measure of the parasitic imbalance and the measure of the useful imbalance corresponding to the measured value. Therefore, to find a useful sensor imbalance, the measured sensor imbalance is divided by the coefficient of sensitivity change, and then the initial sensor imbalance is subtracted. Thus, the sensitivity of the sensor to the measured value is reduced to the sensitivity at the time of balancing, and the parasitic imbalance is fully compensated.

When applying the inventive method, difficulties may arise associated with a large difference in order of magnitude between the values in the first pair, on the one hand, and in the second and third pairs, on the other hand. As already indicated, the values in the first pair are small differential signals measured against the background of large output signals of single converters, which correspond to the second and third pairs. Registration of these signals may require a very wide dynamic range of the measuring equipment used for registration. Accordingly, the use of measuring equipment with a wide dynamic range can significantly complicate the device for implementing the inventive method.

To reduce the dynamic range of the recorded signals when registering the values of the second and third pairs, the transmission coefficient is reduced in the signal conversion section from the sensor output to the value recording device, in order to avoid signal limitation in this section and in the registration device. The transfer coefficient can be switched both at the signal conversion section from the input of the synchronous detector to the recording device, and at the signal conversion section from the sensor output to the input of the synchronous detector. Each of these options for implementation has its own characteristics.

When switching is performed at the signal conversion section from the quadrature detector input to the registration device, it does not affect the phase incursion of the high-frequency signal. Both the output vector and the sensitivity vector are transformed by the same circuit, therefore, the phase shift for both of these vectors is the same. Accordingly, the direction of the sensitivity vector almost coincides with the significant component of the output vector, which allows adequate correction of errors. On the other hand, a synchronous detector must have a wide dynamic range of input signals, which requires careful design of this section of the circuit. Switching the transmission coefficient can be realized either by switching the sensitivity of the quadrature detector, or by switching the gain in the DC amplifier, which amplifies the detector output signal for supply to the registration device.

On the other hand, switching can be implemented on the conversion section of the high-frequency signal, even before it is fed to the input of the synchronous detector. In this case, switching the transmission coefficient is likely to cause some change in the phase incursion in the amplifying elements, which will violate the consistency between the directions of the sensitivity vector and the significant component of the output vector. To avoid this mismatch, when finding the imbalance of the sensor, it is necessary to take into account the previously known change in the phase shift during switching, to determine which may require calibration of the high-frequency part of the signal conversion circuit. The advantage of this implementation can be considered low requirements for the applied synchronous detector, which does not require a wide dynamic range of input signals.

A synchronous detector can be implemented in different ways. For example, it can be a true quadrature detector and at the same time receive from the generator two matched reference signals shifted 90 degrees relative to each other. In such a detector, in fact, two independent synchronous detectors are implemented, each of which is connected to a separate reference signal and generates a separate quadrature signal. A certain technical complexity is ensuring the equality of the transmission coefficients of each of the synchronous detectors, which is necessary for the accurate execution of vector signal processing.

A simpler solution could be to use one synchronous detector, alternately connected to one and the other reference signals shifted 90 degrees relative to each other. Such a synchronous detector alternately generates first one and then another quadrature signal. This solution reduces the complexity of the device and guarantees the same transmission coefficient for both quadrature signals, but reduces the measurement speed. The described solution is equivalent to using the same reference signal, and alternately connecting the driver of the exciting voltage on the plates to the orthogonal outputs of the generator of the driving signal, the voltage on which is 90 degrees shifted relative to each other.

Однако, этот метод отличается низкой точностью, поскольку малая величина выходного вектора

находится как сумма двух больших величин векторов

при измерении которых имеются относительные погрешности, порождающие абсолютные отклонения, пропорциональные длине данных векторов. Эти отклонения оказываются достаточно большими в сравнении с

что снижает ценность подобного косвенного измерения. Фактически, речь идет об использовании дифференциального датчика в недифференциальном режиме, когда результат дифференциального измерения и связанные с ним преимущества не используются. При непосредственном получении

из небольших по величине дифференциальных квадратурных сигналов первой пары, относительные погрешности измерения вызывают малые абсолютные отклонения. Поэтому, в заявленном способе обязательно проводится измерение первой пары квадратурных сигналов.The values of the second and third pairs of quadrature signals allow you to indirectly, without using the values of the first pair, obtain the output vector if vector addition is performed:

However, this method is characterized by low accuracy, since the small value of the output vector

is found as the sum of two large quantities of vectors

when measured, there are relative errors that generate absolute deviations proportional to the length of these vectors. These deviations turn out to be quite large in comparison with

which reduces the value of such an indirect measurement. In fact, we are talking about using a differential sensor in non-differential mode, when the differential measurement result and the associated benefits are not used. Upon direct receipt

of the small differential quadrature signals of the first pair, the relative measurement errors cause small absolute deviations. Therefore, in the claimed method, the first pair of quadrature signals is necessarily measured.

Particular advantages are provided by the application of the indicated method for measurement using a multi-element differential sensor containing more than two transducers. A multi-element differential sensor is used to find several imbalance values at once. As a rule, the transducers are divided into groups consisting of two or more transducers, and sequentially measure in several modes. Each mode corresponds to the supply of exciting voltages to only one group of converters. An example of such a sensor is considered in the prototype of the claimed invention, where the set of unbalance values obtained corresponds to the differences in the effective dielectric constant in different places of the controlled sheet, corresponding to the location of the groups of converters.

In multi-element sensors, due to the large number of transducers, the equalization of the phase shift along the path of the exciting signal from the generator to the converter is of particular difficulty. At an operating frequency of 100 MHz, even a small delay of 560 ps leads to a phase incursion of 20 degrees. It is technically not possible to completely eliminate phase incursions between different transducers of the sensor. If individual tracking and compensation of such incursions is not provided for different measurement modes, a difference in sensitivity and signal to noise ratio arises between these modes. The specified difference in sensitivity was one of the main disadvantages identified in the practical application of the prototype.

In one implementation of the claimed invention, specifically oriented to a multi-element sensor, when using a sensor containing more than two transducers, a set of excitation signal supply modes is preliminarily determined for measurements, so that at least one transducer for supplying of the first excitation signal, and at least one converter is indicated in advance for supplying a second excitation signal to it, and the signal to the remaining converters zbuzhdeniya not applied, and perform measurement cycle in each of the defined modes.

For each mode, its own sensitivity vector is determined, which allows one to obtain a measurement result with high accuracy by compensating for the influence of external disturbing factors, as already described. In addition, tracking an individual sensitivity vector for each set of transducers involved in a particular mode neutralizes the harmful effects of significant and different phase shifts occurring in different modes. Due to this, it is possible to achieve the same sensor sensitivity in different modes. For the measurement in each individual mode of the multi-element differential sensor, all the improvements listed above for the differential sensor can be applied. Of particular benefit is the reduction of the sensitivity in each mode to the sensitivity at the time of balancing. In one implementation, for each of the modes, before the measurement, in the absence of the measurement object, the initial sensitivity vector and the initial sensor unbalance corresponding to the selected mode are determined, and during the measurement cycle, the measured sensitivity vector and the measured sensor unbalance corresponding to the selected mode are determined, and calculate the coefficient of sensitivity change corresponding to the selected mode and equal to the ratio of the measured sensitivity vector module corresponding to the selected mode, to the module of the initial sensitivity vector corresponding to the selected mode, and to find the imbalance of the sensor corresponding to the measured value in the selected mode, the measured sensor imbalance corresponding to the selected mode is divided by the sensitivity change coefficient corresponding to the selected mode, and the initial sensor unbalance corresponding to selected mode.

The described implementation is very effective when the converters involved in each mode are located in close proximity to each other and the same disturbing factors act on them. At the same time, quite often it happens that, due to the large size of the multi-element sensor, disturbing factors that vary in magnitude act on the transducers involved in different modes. However, since the change in sensitivity is monitored individually for each of the modes, it neutralizes this difference in disturbing factors. In practice, this means that the harmful effect of the geometric instability of the design of the sensor itself, which manifests itself, for example, in the deflections and torsions of the capacitive sensor plates under the influence of vibration, is largely neutralized. Similarly, in the case of magnetic sensors, the influence of a large ferromagnetic mass located closer to one of the edges of the sensor sensitivity zone is neutralized.

In FIG. Figure 1 shows vector diagrams of signals generated by a differential sensor excited by high-frequency antiphase signals.

In FIG. 2 shows a connection diagram of a differential sensor used in an example implementation of the inventive method.

In FIG. 3 is a vector diagram explaining the receipt of a measurement result.

In FIG. 4 shows a circuit for switching on a multi-element differential sensor.

The practical implementation of the method is shown on the example of a differential capacitive sensor of heterogeneity of the rolled material. This sensor is used in a paper web quality control machine, where the web is advanced through the sensor using a transport mechanism. The task of controlling the thickness of the paper using a capacitive sensor makes high demands on the measurement method. As indicated in the prototype, with a gap between the plates of the measuring capacitor of 0.5 mm and a paper thickness of 0.1 mm, a twofold decrease in thickness will lead to a 5% change in the signal of the capacitive transducer. At the same time, a change in the gap by 0.1 mm, for example, due to vibration, will lead to a signal change of 20%. These values give an idea of the typical magnitude of the change in signals that need to be processed.

The electrical circuit of the device used in practical implementation is shown in FIG. 2. The sensor is installed in a flat channel along which the paper web moves (not shown in the figure). Plates 6 and 7 of the measuring capacitors of the capacitive sensor are placed on one side of the channel, and the combined plate 8 is placed strictly opposite to the plates 6 and 7 on the other side of the channel. One transducer of the sensor is a capacitor with plates 6 and 8, and the second transducer has plates 7 and 8. When the web passes through a flat channel between the plates, the capacitance of the capacitor 6-8 and the capacitor 7-8 changes depending on the thickness of the web. In the absence of heterogeneity, these capacities will be the same, but in the presence of heterogeneity of the fabric, which is within the gap of one of the capacitors, the capacitance of the capacitors will vary. The principle of operation of this sensor is based on the detection of the difference between the capacitances of the capacitors 6-8 and 7-8. A similar principle of detecting heterogeneity is used in the prototype of the claimed invention.

Plates 6 and 7 are connected to a source 1 of high-frequency exciting voltage. It contains a crystal oscillator 2, providing a rectangular signal, having the shape of a meander, at the output OUT_0. The signal frequency is 100 MHz. At the outputs OUT_90 and OUT_180, the generator generates signals of the same frequency and shape, but phase shifted relative to the signal OUT_0 by 90 degrees and 180 degrees, respectively.

The out-of-phase signals OUT_0 and OUT_180 are supplied from the generator 2 to the OR logic elements 3 and 4, made using CMOS technology on a common semiconductor chip. These logic elements are used as drivers of excitation signals supplied in antiphase to the sensor plates 6 and 7. A feature of the logic elements made by the CMOS technology is a small difference in the amplitude of the output voltage. The magnitude of the output voltage, in fact, is equal to the supply voltage, subject to a small load current. In addition, gates of this type have a fairly low and stable output impedance, which practically does not change depending on the output signal.

The generator 2 has logic outputs and is connected to the inputs of the logic elements, therefore, small changes in the amplitude of the output signals of the generator do not affect the operation of the circuit. For ease of calculating the transfer coefficient on the way from the generator to other elements of the signal processing path, we will assume that the output voltage levels of the generator are constant and equal to the nominal values of 0 V and 3.3 V. The output voltages of logic elements 3 and 4, on the contrary, through the plates 7 and 8 of the sensor 5 directly affect the voltage on the elements of the signal processing path, and depend on the supply voltage. Due to this, the instability of the supply voltage causes the instability of the transmission coefficient on the way from the generator to other elements of the signal processing path.

The receiving plate 8, common to both measuring capacitors, is connected to the receiving circuit 9. This circuit contains an input resonant circuit 10, an aperiodic high-frequency amplifier 11, a synchronous detector 12, a low-pass filter 13, and an operational amplifier 14. When applying antiphase excitation signals to the plates 6 and 7, the summing of the currents of the measuring capacitors in the receiving plate 8. As a result, the sum of the currents of the measuring capacitors arrives at the resonant circuit.

The sensor is controlled by a microcontroller 17, which has a built-in analog-to-digital converter (ADC). The ADC input is indicated by ADC_IN. In addition, the microcontroller generates control logic signals designated PIO0-PIO3. The microcontroller also provides measurement results via a serial communication channel (not shown in the figure).

The resonant circuit 10 serves to suppress interference that penetrates the input of the receiving circuit 9 from the output of the sensor 5. In addition, due to the inductance, the resonant circuit 10 allows you to neutralize the stray capacitance of the shielding surfaces (not shown in the figure), which inevitably occur when the sensor is shielded from external interference and could reduce the transmission coefficient. The low-pass filter 13 further narrows the overall sensitivity bandwidth of the receiving circuit 9 and also allows you to deal with interference from the output of the sensor 5. Additionally, this filter reduces harmful high-frequency ripple of the signal at the input of the ADC of the controller 17, which reduces the repeatability of the conversion result.

Using the control signals PIO1 and PIO2 connected to the inputs of logic elements 3 and 4, the microcontroller can independently enable and disable the output of an alternating excitation voltage from the generator 2 to the sensor plates 6 and 7. By issuing both exciting signals, or each of them individually, the sensor operates in the mode for receiving the first, second or third pairs of quadrature signals, respectively.

The control signal PIO0 is output to the reference signal switch 18. Depending on the logic level of PIO0, the switch generates a signal OUT_0 from the generator 2 to the input of the synchronous detector 12, or the signal OUT_90, phase shifted 90 degrees relative to it. Switching the switch 18 makes it possible to obtain both a sine and a cosine quadrature signal at the output of the synchronous detector 12.

The PIO3 signal controls the closing and opening of switch 19. This switch switches the gain of the DC amplifier performed on the operational amplifier 14 and resistors 15 and 16. When the switch is closed, the DC amplifier provides a gain of 20 for the quadrature signal received by the detector 12 and smoothed by a filter 13. When the switch is open, the DC amplifier operates as a repeater, with a gain equal to 1. The output signal of the DC amplifier under fed to the input of the ADC of the microcontroller 17. The ADC converts the input voltage into a 12-bit digital code.

To measure the quadrature signals, the microcontroller 17, using the switch 18, first supplies the voltage OUT_0 from the generator 2 to the input of the reference signal of the synchronous detector 12. Further, the microcontroller 17 performs a delay to stabilize the voltage at the output of the receiving circuit 9. The duration of this pause is 10 times longer than the filter time constant 13, which allows you to complete the transition process, so that the deviation from the stationary value does not exceed the price of the least significant bit of the ADC. Then, the microcontroller 17 starts the analog-to-digital conversion process, as a result of which a code value of the sine quadrature signal is obtained.

Then, the microcontroller 17, using the switch 18, supplies the voltage of the output OUT_90 of the generator 2 to the input of the reference voltage of the synchronous detector 12. After the delay steps, to stabilize the voltage at the output of the receiving circuit 9 and analog-to-digital conversion, which are exactly the same as previously described, a value code is obtained cosine quadrature signal. During the measurement of quadrature signals, the microcontroller does not change the state of the outputs PIO1 - PIO3, due to which the sensor circuit works with constant excitation and a constant gain. The values of the sine and cosine quadrature signals form a pair.

как показано на Фиг. 3.Before taking measurements, in the absence of a paper web in the gap of the sensor 5 and the transportation mechanism turned off, the microcontroller 17 carries out balancing. To do this, he sequentially receives the first, second and third pair of quadrature signals. First, it sets the logic zero state at the outputs PIO1 and PIO2, which ensures the supply of antiphase excitation signals to both plates 6 and 7. Using the PIO3 signal, the microcontroller 17 puts the switch 19 in the closed state, which provides a constant current gain of 20. Next , the quadrature signals are measured to obtain the first pair of values, as described previously. The microcontroller forms the coordinates of the output vector from this pair

as shown in FIG. 3.

After that, the microcontroller 17 sets the signal PIO1 to the state of the logical unit, which prohibits the supply of the excitation signal to the plate 6 and saves the excitation on the plate 7. At the same time, the microcontroller 17 outputs the signal PIO3 to open the switch 19, due to which the DC amplifier is converted to repeater mode with a gain of 1. Following this, a quadrature signal is measured to obtain a second pair of values. From this pair, the microcontroller forms the coordinates of the vector

Then, the microcontroller 17 sets the signal PIO1 to the logic zero state and PIO2 to the logic one state, which prohibits the supply of the excitation signal to the plate 7 and saves the excitation on the plate 6. Following this, the quadrature signals are measured to obtain a third pair of values. From this pair, the microcontroller forms the coordinates of the vector

Длина этого вектора характеризует, в момент балансировки, коэффициент передачи на пути от генератора 1 до входа АЦП микроконтроллера 17, и чувствительность датчика к измеряемой величине. Направление вектора соответствует значимому компоненту выходного вектора

датчика. Отметим, что во время балансировки значимый компонент выходного вектора является паразитным дисбалансом. Чтобы найти его, микроконтроллер вычисляет проекцию выходного вектора на вектор чувствительности для режима балансировки

Отметим, что получаемое при такой проекции скалярное значение имеет знак, то есть, оно может быть как положительными, так и отрицательными. Для последующих вычислений потребуется использовать не сам вектор чувствительности

а только его длину

Having finished balancing, the microcontroller starts processing the received data. This process is illustrated by the vector diagram in FIG. 3. Since when balancing the paper web in the path is absent and the transport mechanism is turned off, there are no vibrations of the sensor plates and imbalance due to inhomogeneities of the web. During balancing, it is necessary to find the initial values of spurious imbalance and the sensitivity vector of the sensor, which will then be used to make corrections to the measurement results. For this, the microcontroller calculates the sensitivity vector when balancing

The length of this vector characterizes, at the time of balancing, the transmission coefficient on the way from the generator 1 to the input of the ADC of the microcontroller 17, and the sensitivity of the sensor to the measured value. The direction of the vector corresponds to the significant component of the output vector

sensor. Note that during balancing, a significant component of the output vector is a parasitic imbalance. To find it, the microcontroller calculates the projection of the output vector on the sensitivity vector for the balancing mode

Note that the scalar value obtained with such a projection has a sign, that is, it can be both positive and negative. For subsequent calculations, it is required to use not the sensitivity vector itself

but only its length

из второй - вектора

из третьей - вектора

Из-за появления полотна в зазоре датчика 5 и из-за вибрации его пластин изменяется емкость измерительных конденсаторов 6-8 и 7-8. В свою очередь, это приводит к сдвигу резонансной частоты цепи 10, куда подключены измерительные конденсаторы. Сдвиг резонансной частоты изменяет коэффициент передачи приемной схемы 9 и набег фазы на пути от генератора до входа фазового детектора. Дополнительно, изменение емкостей измерительных конденсаторов 6-8 и 7-8 изменяет токи, втекающие в приемную схему 9, что также изменяет коэффициент передачи. Поэтому, в сравнении с моментом балансировки, векторная диаграмма сигналов OA'В' поворачивается в положение ОАВ на угол β, на который изменился набег фазы, и масштабируется в соответствии с изменением коэффициента передачи. Дополнительно, из-за появления неоднородности бумажного полотна в зазоре между пластинами 7 и 8, изменяется емкость образованного ими измерительного конденсатора, и возникает дополнительный малый полезный сигнал

прибавляющийся к сигналу

и сонаправленный ему.To measure heterogeneity, the microcontroller performs the acquisition of the first, second and third pairs of values in exactly the same way as balancing is done. The coordinates of the output vector are formed from the first pair

from the second - vector

from the third - vector

Due to the appearance of the web in the gap of the sensor 5 and due to the vibration of its plates, the capacitance of the measuring capacitors 6-8 and 7-8 changes. In turn, this leads to a shift in the resonant frequency of circuit 10, where the measuring capacitors are connected. The shift of the resonant frequency changes the gain of the receiving circuit 9 and the phase incursion from the generator to the input of the phase detector. Additionally, changing the capacitances of the measuring capacitors 6-8 and 7-8 changes the currents flowing into the receiving circuit 9, which also changes the transmission coefficient. Therefore, in comparison with the moment of balancing, the vector diagram of the OA'B 'signals is rotated to the OAV position by an angle β, by which the phase shift has changed, and is scaled in accordance with the change in the transmission coefficient. Additionally, due to the heterogeneity of the paper web in the gap between the plates 7 and 8, the capacitance of the measuring capacitor formed by them changes, and an additional small useful signal appears

added to the signal

and co-directed to him.

Actually, finding a useful imbalance, that is, the length of this signal characterizing heterogeneity, is the ultimate goal of measurement.

на измеряемую величину невозможно непосредственно измерить, но его можно получить методом вычисления из непосредственно измеряемых векторов

а также данных балансировки. Для этого, данные измерения необходимо обработать с учетом описанных изменений коэффициента передачи и набега фазы, а также результатов балансировки.Converter response increment

it is impossible to directly measure by the measured value, but it can be obtained by calculating from directly measured vectors

as well as balancing data. For this, the measurement data must be processed taking into account the described changes in the transmission coefficient and phase incursion, as well as the results of balancing.

Отметим, что в силу малости отклика

его вкладом в длину вектора чувствительности можно пренебречь. Поворот и изменение вектора

по отношению к ранее найденному

характеризуют изменение набега фазы и коэффициента передачи при измерении, в сравнении с моментом проведения балансировки.The microcontroller calculates the sensitivity vector when measuring

Note that due to the small response

its contribution to the length of the sensitivity vector can be neglected. Rotate and change the vector

in relation to previously found

characterize the change in phase incursion and transmission coefficient during measurement, in comparison with the moment of balancing.

который характеризует суммарное изменение коэффициента передачи. После этого, он вычисляет проекцию выходного вектора на вектор чувствительности для режима измерения

Как видно из векторной диаграммы, значение MP с высокой степенью точности представляет собой сумму компонента MN паразитного дисбаланса в направлении полезного сигнала плюс скалярное значение полезного дисбаланса NP. На самом деле, NP не совсем точно передает полезную разность |АС| между амплитудами выходного сигнала преобразователей, поскольку |NP|=|AC|cosδ, где δ=α/2 и α есть угол между выходными сигналами преобразователей 6-8 и 7-8. Относительная погрешность такого представления полезного дисбаланса равна 1-cosδ. Для угла α, равного 20 градусам, что соответствует полученным нами наихудшим экспериментальным результатам, эта погрешность примерно равна 0,015. Это значение погрешности мало подвержено влиянию возмущающих факторов, так как угол зависит от мало изменяющихся параметров, таких как геометрическая разность хода возбуждающих сигналов по проводникам к пластинам 6 и 7, и разность задержек распространения в логических элементах 3 и 4. Поэтому, значимость данной погрешности невелика и она может рассматриваться как небольшое и постоянное по величине снижение чувствительности датчика.Microcontroller calculates sensitivity measurement coefficient

which characterizes the total change in transmission coefficient. After that, it calculates the projection of the output vector on the sensitivity vector for the measurement mode

As can be seen from the vector diagram, the MP value with a high degree of accuracy is the sum of the component MN of the parasitic imbalance in the direction of the useful signal plus the scalar value of the useful unbalance NP. In fact, NP does not accurately convey the useful difference | AC | between the amplitudes of the output signal of the converters, since | NP | = | AC | cosδ, where δ = α / 2 and α is the angle between the output signals of the converters 6-8 and 7-8. The relative error of such a representation of the useful imbalance is 1-cosδ. For an angle α equal to 20 degrees, which corresponds to the worst experimental results we obtained, this error is approximately 0.015. This error value is little affected by disturbing factors, since the angle depends on slightly changing parameters, such as the geometric difference in the path of the exciting signals along the conductors to plates 6 and 7, and the difference in propagation delays in logic elements 3 and 4. Therefore, the significance of this error is small and it can be considered as a small and constant in magnitude decrease in the sensitivity of the sensor.

Spurious imbalance, as already considered, when exposed to a disturbing factor is scaled along with the entire vector diagram. Based on this, the controller finds the unnormalized value of the useful imbalance by performing a scalar subtraction of the initial parasitic imbalance adjusted for sensitivity changes:

And finally, the controller calculates the value of the useful imbalance, normalized to the sensitivity of the sensor at the time of balancing:

The non-normalized NP and normalized D values of useful unbalance obtained by us are relative values. The value of the useful signal can be reduced to absolute physical quantities, for example, to the value of the difference in the thickness of the paper, measured in micrometers. To do this, pass through the device a specially made calibration target with a known thickness difference, measured with a micrometer. The value of the useful imbalance, correlated with the difference in thickness on the target, allows you to obtain a calibration factor for subsequent conversion from the relative value of the useful signal in micrometers. However, if an unnormalized value of NP is used for this purpose, then it will be different for successive measurements for the same target, since the sensitivity of the sensor changes due to external disturbances. Therefore, the calibration conversion factor will be inconsistent and therefore inaccurate. The value of the normalized value of D is that when calculating it, the sensitivity change in each measurement is taken into account. Therefore, its application gives much higher accuracy, since the calibration coefficient was initially measured without the influence of external perturbations, and the subsequent use of this coefficient to convert the measurement results to micrometers also minimizes the effect of external perturbations.

Note that the vast majority of calculations performed during balancing and measurement are performed in scalar form. In vector form, only scalar products are calculated when projections are obtained according to formulas (2) and (3). This determines the small computational complexity of the claimed method.

After a single balancing, measurements can be made repeatedly. For example, balancing can be done once at the beginning of a work shift, and then during this shift measurements can be taken at a frequency of several hundred times per second.

If the disturbing effect is rather slow in comparison with the frequency of measurements, it is permissible to register the second and third pair of values of the quadrature signal not in each measurement cycle. That is, the first pair must be recorded in each measurement cycle, and the second and third pair can be recorded less frequently, for example, in every fifth measurement cycle. To calculate in the measurement cycle when only the first pair is recorded, the values of the second and third pairs for measurements should be taken from the previous measurement cycle. Such a decrease in the measurement frequency of the second and third pairs is acceptable if these values characterizing the disturbing effect change little between adjacent measurements. It is also possible to register a second pair in one measurement cycle and a third pair in another.

The described options for a relative decrease in the recording frequency of the second and third pairs allows increasing the registration frequency of the first pair of values, which carries a significant component of the output vector. This can be important in case of a rapid change in the measured value in comparison with the characteristic time of change of the disturbing effect.

When registering the second and third pairs of values of quadrature signals, the excitation signal is supplied only to one of the exciting plates 6 and 7, and the potential of the common wire is supplied to the other. Due to the fact that logic elements 3 and 4 have an almost constant output impedance at any value of the output signal, plates 6 and 7 are connected to a common wire by a constant impedance, which does not differ from the impedance when registering the first pair of values. The specified impedance is included in the resonant circuit 10 through the capacitance of the measuring capacitors 6-8 and 7-8. Since the impedance does not change, the transfer characteristic in the area from the generator 2 to the input of the synchronous detector 12 remains unchanged during registration of all three pairs of quadrature signals. Due to this, the sensitivity vector with high accuracy reflects the specified transfer characteristic at the time of registration of the first pair of quadrature values, which increases the accuracy of the measurements.

In principle, it would be possible to use buffers with the third state instead of logic elements 3 and 4. Then, to stop the supply of the exciting signal, it would be necessary to transfer the corresponding output of the buffer to the high-impedance state. However, this would lead to a change in the impedance included in the resonant circuit 10, and a change in the transfer characteristic. As a result, the transfer characteristic when registering the first pair of values would be slightly different from the characteristic when registering the second and third pairs. This would result in a slight decrease in measurement accuracy. In general, when implementing the inventive method, it is desirable to use excitation signal sources with stable impedance. However, a change in the source impedance leads to a slight decrease in accuracy, but not to a loss in the overall performance of the method and its technical result.

To further improve the accuracy of the sensor, it is advisable to compensate for static error. The synchronous detector 12, the operational amplifier 14, and the ADC of the microcontroller 17 are sources of static errors arising as adding a constant non-zero value to any real value obtained as a result of synchronous detection. Static errors are common to all analog circuits that process DC signals. In addition, the source of the static error can be a constant amplitude pickup of the signals of the generator 2 passing to the elements of the receiving circuit 9 bypassing the sensor 5.

To compensate for the static error, the microcontroller 17 must completely turn off the supply of the high-frequency excitation signal to the sensor 5, issue a control level PIO3 to close the switch 19, and register the first static pair of quadrature signals. Then, it must issue the control level PIO3 to open the switch 19 and register the second static pair of quadrature signals. During the subsequent registration, the values of the first static pair must be subtracted from the values of the first pair of quadrature signals, and the values of the second static pair must be subtracted from the values of both the second and third pairs of quadrature signals to obtain the true values of these three pairs. The use of two static pairs is due to the fact that static errors change when switching the gain of amplifier 14. It is advisable to register the first and second static pairs every few seconds in order to monitor the temperature drift of the static error of analog elements 12, 14, and 17 that process DC signals .

To control the heterogeneity of the paper web over the entire width, a plurality of capacitive sensors 5 can be installed in the tract, equipped with a measuring circuit constructed according to the principle shown in FIG. 2. For each of the sensors, it is necessary to ensure that the measurement method described above is performed. Each of the sensors, when the web is moving by the transport mechanism, controls an extended track on the web. Sensors must be positioned so that these tracks overlap to a small degree. In order to avoid mutual interference in the form of beats, the sensors must be synchronized in frequency and phase of the exciting signal. To do this, their measuring circuits must have a common generator.

A simpler solution for controlling the web across the entire width, requiring fewer components, can be implemented using the multi-element sensor shown in FIG. 4. A similar sensor is described, for example, in the prototype of the claimed invention. The multi-element capacitive sensor 21 contains sixteen exciting plates 22, designated E0-E15, and one common receiving plate 23. Measuring capacitors are formed between each of the plates 22 and the receiving plate 23. The sensor 21 allows you to implement 15 modes, numbered from 0 to 14. In the mode 0 antiphase excitation signals are fed to the plates E0, E1; in mode 1 - on plates El, E2; and so on to mode 14, in which the signals are applied to the plates E14, E15.

Each of the modes corresponds to the activation of a pair of adjacent plates 22, and provides a measurement of the heterogeneity of the material in the region covered by these plates. Thus, a pair of adjacent plates forms a measuring element of the sensor. The consecutive measurement of the heterogeneity for each element of the sensor allows you to measure the heterogeneity of the material located in the gap between the plates 22 and 23, across the entire width of the sensor.

Plate E0-E15 connected to the outputs of the multiplexer 24, made by CMOS technology. The multiplexer 24 provides switching of the input signals A and B received from the output of the elements OR 3 and 4. On the control bus GN0 ... 3 of the multiplexer 24 from the outputs PI04 ... 7 of the microcontroller 17, the code of the mode number T is supplied. The multiplexer 24 transmits a signal from input A to output E [T], and from input B to output E [T + l], and connects the remaining outputs to a common wire. That is, the mode number M, set by the microprocessor on the line PIO4 ... 7, provides the supply of excitation signals to those plates 22 of the sensor 21 that correspond to this mode.

Plates to which excitation signals are not supplied in a certain T mode turn out to be connected to a common wire through the same impedance of the output CMOS logic stage as the excited plates. This ensures the invariance of the total impedance connecting the receiving plate 23 through the measuring capacitors of the sensor 21 with a common wire. The invariance of the impedance can reduce the influence of switching modes on the transfer coefficient of the resonant circuit 10 and the aperiodic amplifier 11.

The described multiplexer 24 performs a logical function that is not typical for standard logic circuits. Therefore, it is most conveniently performed on the basis of a programmable logic chip (CPLD or FPGA).

для каждого режима T с номером от 0 до 14. После запуска процесса измерения, микроконтроллер последовательно перебирает все режимы и проводит измерения трех пар значений и вычисления для каждого отдельного режима Т с номером от 0 до 14, получая нормированное значение полезного сигнала D_T. При вычислениях используются результаты балансировки, ранее запомненные для соответствующего режима.In each of the modes, balancing, measuring and calculating the result are performed under the control of the microcontroller, exactly as described previously for a single differential sensor. During balancing, the microcontroller sequentially goes through all the modes and remembers the length of the sensitivity vector | OS ' _T | and the projection of the output vector M'N ' _T on the direction of the sensitivity vector

for each mode T with a number from 0 to 14. After starting the measurement process, the microcontroller sequentially goes through all the modes and measures three pairs of values and calculates for each individual mode T with a number from 0 to 14, obtaining a normalized value of the useful signal D _T. In the calculations, the balancing results previously stored for the corresponding mode are used.

Thus, each of the obtained values of D _T corresponds to a normalized inhomogeneity located on the element of the sensor formed by adjacent plates 22, designated EM and E (M + 1). The influence of a disturbing factor on each pair of plates is processed separately, which allows you to neutralize local disturbances that act differently on the sensor elements. Consider the case when the plate 23 was bent so that the distance from the plates E0, E1, E14 and E15 remained practically unchanged, and the distance to the plates E7 and E8 increased by almost the same amount, since these plates are located directly next to each other. This is a realistic example of how the vibrations of the sensor elements appear, fixed along the edges of the path and not having support in the middle of the path.

в режимах 0 и 14 останется неизменным, а в режиме 7 уменьшится. Соответственно, уменьшится коэффициент изменения чувствительности K_S7 для режима 7, что приведет к соответствующей поправке при вычитании начального паразитного дисбаланса и практически полной его компенсации. И, наконец, полезный сигнал D₇ будет отнормирован к чувствительности на момент балансировки, когда вибрация и деформация датчика отсутствовали. Такое же вычитание начального паразитного дисбаланса и нормировка произойдет и для краевых элементов датчика, с получением полезных сигналов D₀ и D₁₄. В каждом из этих случаев, для каждого из режимов Т, используются индивидуальные величины коэффициента K_ST изменения чувствительности. Таким образом, влияние вибрации на элементы датчика, несмотря на локальные различия в величине этой вибрации, оказывается нейтрализованным.As a result of measurement and subsequent calculations, the sensitivity vector

in modes 0 and 14 it will remain unchanged, while in mode 7 it will decrease. Accordingly, the coefficient of sensitivity change K _S7 for mode 7 will decrease, which will lead to a corresponding correction when subtracting the initial spurious imbalance and its almost complete compensation. And finally, the useful signal D ₇ will be normalized to sensitivity at the time of balancing, when vibration and deformation of the sensor were absent. The same subtraction of the initial spurious imbalance and normalization will occur for the edge elements of the sensor, with the receipt of useful signals D ₀ and D ₁₄ . In each of these cases, for each of the T modes, individual values of the coefficient K _{ST of} sensitivity change are used. Thus, the effect of vibration on the sensor elements, despite local differences in the magnitude of this vibration, is neutralized.

The static error compensation described earlier for a differential sensor can be applied in exactly the same way to a multi-element differential sensor. Note that the zero pair is common to all modes 0-14, since the static errors are the same for these modes.

The measurement method described here using a multi-element differential sensor can be applied not only to control the paper web, but also to control the presence of heterogeneity on banknotes or other sheet documents. This allows you to detect defects and signs of wear of the document, such as through holes, thinning at the folds, as well as gluing made with a layer of adhesive polymer tape.

Claims (10)

1. A measurement method using a differential sensor designed to excite high-frequency antiphase signals, in which the differential sensor contains at least two transducers that are connected to a common output of the sensor and configured to supply excitation signals to them; to rectify the output of the sensor, a synchronous detector is used, synchronized with the excitation signal and configured to generate a pair of quadrature signals, where the first and second excitation signals, which are out of phase, are fed to the sensor at the same time, and the values of the first pairs of quadrature signals, then turn off the supply of the first excitation signal while maintaining the supply of the second excitation signal and register on the output value of the second pair of quadrature signals, after which, based on the obtained values of the first pair, find the scalar value of the sensor imbalance corresponding to the measured value, and the values of the second pair are used to compensate for errors in finding the scalar value of the imbalance of the sensor, and carry out the measurement cycle at least once . 2. The measurement method according to claim 1, in which, on the basis of the first pair of quadrature signals, an output vector is formed, and on the basis of the second pair, a sensor sensitivity vector is formed, to find the scalar value of the sensor imbalance, the projection length of the output vector on the direction of the sensitivity vector is found, and the value of the length of the sensitivity vector is used to compensate for the time variation of the sensitivity of the sensor transducers. 3. The measurement method according to claim 1, in which, in the measurement cycle, the second excitation signal is cut off while maintaining the first excitation signal and the values of the third pair of quadrature signals are recorded, and the values of the third pair are used to compensate for errors when finding the scalar value of the sensor unbalance. 4. The measurement method according to claim 3, wherein in the measurement cycle, an output vector is formed on the basis of the first pair of quadrature signals, and the corresponding sensitivity vectors are formed on the basis of the second and third pairs and the sensitivity vector of the sensor is calculated as the difference of these sensitivity vectors, then to find the scalar value of the sensor imbalance is the value of the projection length of the output vector on the direction of the sensor sensitivity vector, and the value of the sensitivity vector length is used to compensate and Changes in time of the sensitivity of the sensor transducers. 5. The measurement method according to any one of paragraphs. 3, 4, in which, when registering the values of the second and third pairs of quadrature signals, the transmission coefficient is reduced in the signal conversion section from the sensor output to the value recording device, in order to prevent signal limitation in this section and in the registration device. 6. The measurement method according to claim 5, in which the reduction of the transmission coefficient is implemented in the signal conversion section from the input of the synchronous detector to the registration device. 7. The measurement method according to claim 6, in which a sensor is used containing two transducers, and before the measurement, in the absence of the measurement object, the initial sensitivity vector and the initial sensor unbalance are determined, and during the measurement cycle, the measured sensitivity vector and the measured sensor unbalance are determined and calculating a coefficient of sensitivity change equal to the ratio of the module of the measured sensitivity vector to the module of the initial sensitivity vector, and to find the imbalance of the sensor, respectively According to the measured value, the measured sensor imbalance is divided by the coefficient of sensitivity change and the initial sensor imbalance is subtracted. 8. The measurement method according to claim 5, in which the reduction of the transmission coefficient is implemented in the signal conversion section from the sensor output to the input of the synchronous detector, and when the imbalance of the sensor is found, the previously known change in the phase shift in the named section, which occurs when the transmission coefficient is reduced, is taken into account. 9. The measurement method according to any one of paragraphs. 1-6, 8, in which a sensor containing more than two transducers is used, and for carrying out measurements, a set of excitation signal supply modes is preliminarily determined, so that for each of the modes at least one converter is specified in advance for supplying a first excitation signal to it and at least one transducer is indicated in advance for supplying a second excitation signal to it, and no excitation signal is supplied to the remaining transducers, and a measurement cycle is performed in each of the defined modes. 10. The measurement method according to claim 9, in which, for each of the modes, before the measurement, in the absence of the measurement object, the initial sensitivity vector and the initial sensor imbalance corresponding to the selected mode are determined, and during the measurement cycle, the measured sensitivity vector and the measured the imbalance of the sensor corresponding to the selected mode and calculate the coefficient of sensitivity change corresponding to the selected mode and equal to the ratio of the module of the measured sensitivity vector, respectively the selected mode, to the module of the initial sensitivity vector corresponding to the selected mode, and to find the imbalance of the sensor corresponding to the measured value in the selected mode, the measured sensor imbalance corresponding to the selected mode is divided by the sensitivity change coefficient corresponding to the selected mode, and the initial sensor unbalance is subtracted corresponding to the selected mode.

Images