Position detection with inductive position sensor

Abstract

For an easy-to-implement method for position determination with inductive position sensors having an improved accuracy of the position information, it is provided that the position sensor generates a measurement signal from which a frequency functional (F f ) depending on an excitation frequency (ω) is formed, which frequency functional represents a measure for a noise signal (RS), and the excitation frequency (ω) of the excitation signal (ES) is varied such that the frequency functional (F f ) is minimized or maximized, and the excitation frequency (ω) at which the frequency functional (F f ) is minimized or maximized is used for the excitation signal (ES).

Description

Technical Field

[0001] This invention relates to a method for determining the position of a moving part relative to a stationary part using an inductive position sensor, wherein an excitation winding is arranged on the moving part and at least one secondary winding is arranged on the stationary part, or vice versa, and an excitation signal having an excitation frequency and an excitation amplitude is fed into the excitation winding, the excitation signal generating an electromagnetic excitation field that induces a measurement signal in the secondary winding, the measurement signal being detected, and a noise signal being superimposed on the measurement signal. The invention also relates to an evaluation unit for utilizing such an inductive position sensor for determining position.

[0002] Background Technology: Resolvers are robust and inexpensive rotary position encoders that operate inductively, and therefore are used in a variety of ways. A resolver enables the detection of the angular position of its rotor. If the resolver's rotor is connected to a machine's shaft, the shaft's angular position can be detected. A typical application example of a resolver is obtaining the position information of a motor shaft in the case of an electric motor. The combination of a motor and a resolver is often referred to as a servo motor. The position information can then be used to regulate (e.g., position adjustment or speed adjustment) a servo drive consisting of a servo motor and the load it drives.

[0003] In a resolver, there exists an excitation winding that rotates with the resolver's rotor and at least two secondary windings that are stationary relative to the excitation winding. A high-frequency excitation signal (typically in the frequency range of 1 kHz to 10 kHz) is applied to the excitation winding. The voltage induced in the secondary windings due to the rotating magnetic field of the excitation winding pulsates at the same frequency as the excitation signal; however, its amplitude depends on the position of the excitation winding relative to the corresponding secondary winding. These voltages are output as measurement signals by the resolver and evaluated in a downstream evaluation unit to determine the angular position. This involves amplitude modulation of the measurement signal. The secondary windings are typically arranged 90° off from each other so that the measured voltages have a 90° phase shift. Of course, other angles are also possible. Therefore, the amplitude distribution of the measured voltages is either a sine function depending on the position of the excitation winding relative to the secondary windings, or a cosine function in the case of secondary windings arranged at a 90° offset, or generally a sine function with a specific phase shift. The evaluation of the two measured signals (e.g., by means of the arctangent function (usually the arctan2 function)) enables an explicit calculation of the current angular position of the resolver rotor.

[0004] Therefore, the actual position information lies within the envelope of the measurement signal extracted from the measurement signal. The envelope corresponds to a sine (or cosine) function of the measured voltage, where the period duration corresponds to the rotor's rotation, and the period duration therefore depends on the rotor's angular velocity. A common approach for evaluation is to always sample the measurement signal by an electronic evaluation system with an analog-to-digital converter when the excitation signal reaches its maximum value. Because the excitation signal is typically also generated in the electronic evaluation system, the required sampling time point is precisely known, or alternatively, can be easily determined. Therefore, the electronic evaluation system only detects the peak value of the measurement signal, thus eliminating the high-frequency excitation signal. What remains is a sinusoidal or cosine signal with a period duration corresponding to the rotor's rotation, which contains the actual position information and is evaluated (e.g., by means of arctangent).

[0005] The excitation signal of the resolver, especially its amplitude and frequency, is usually predetermined (e.g., due to the resolver's datasheet). The excitation signal typically does not change after being set during the resolver's operation.

[0006] However, in practical implementations, noise is always present in the measurement signal output by the resolver (e.g., due to random component noise or system noise). Random noise is typically caused by thermal noise and shot noise from components in the electronic evaluation system and can be modeled as white noise. System noise can be caused by the application of the resolver. In the case of servo drives, converters or inverters are typically used, where semiconductor switches are switched at high frequencies. This can lead to high-frequency coupling with the measurement signal. Inductive crosstalk from motor current in the servo drive or other currents in the resolver environment is also conceivable. Such sources of interference can significantly limit the accuracy of the position information obtained from the resolver. Here, accuracy is also understood as the achievable time resolution of the obtained position information. This is particularly understood as the inability to obtain position information at the desired time resolution during the rotation of the resolver rotor due to interference with the measurement signal, because the position information cannot be determined at certain points in time due to noise.

[0007] The resulting problem is the limited accuracy of position information. For example, in a servo drive, position information is used to regulate the servo drive. Therefore, the position information is used as the actual value for regulation. The less accurate the actual value, the more robustly the regulation must be implemented, which usually results in reduced regulation dynamism. Consequently, the servo drive may follow the dynamic setpoint poorly or slowly (in the sense of the setpoint's trend over time with a large rate of change, i.e., large changes in the setpoint within short time intervals). Thus, the servo drive's regulation behavior deteriorates and its regulation dynamism is limited.

[0008] As explained, the useful signal in the measurement signal (i.e., the amplitude-modulated sinusoidal oscillator) is inevitably embedded in the noise signal. To obtain position information, it is essential to extract the useful signal from the measurement signal. For this, it is necessary to clearly reveal the useful signal from the background noise (noise signal). In this case, the signal-to-noise ratio (also known as signal-to-noise distance) (abbreviated as SNR) is introduced. SNR is a measure of the quality of the useful signal embedded in the noise signal. SNR is generally defined as the ratio of the average power of the useful signal to the average noise power of the noise signal. A higher SNR allows for better extraction of the useful signal from a noisy measurement signal. If the SNR is too low, the useful signal cannot be extracted, or only a insufficient or low-time-resolution useful signal is extracted.

[0009] To ensure a sufficiently high SNR, the amplitude of the excitation signal has been chosen to be sufficiently large to date. However, choosing an excessively large amplitude is not recommended, as this can lead to thermal problems in the resolver and increase the requirements for the electronic evaluation system. Furthermore, a high-amplitude excitation signal also places higher demands on its generation.

[0010] Therefore, methods for noise optimization evaluation of measurement signals from resolvers are known in the prior art. An example of this is “Noise Reduction for High-Accuracy Automatic Calibration of Resolver Signals via DWT-SVD Based Filter” by Guo M. et al., Electronics 2019, 8(5):516. In this method, the envelope of the measurement signal is subjected to discrete wavelet transform to filter out noise. This method is computationally very complex, which is disadvantageous for many applications. Although noise optimization evaluation is performed, this method must of course still have a certain SNR for the evaluation to be successful.

[0011] Practice has shown that problems can arise, especially from narrowband (frequency content-dependent) interference sources caused by inductive or capacitive coupling, because the SNR can be very low in these frequency ranges.

[0012] However, such sources of interference are not known in advance. Consequently, the achievable accuracy of the determined location information may be reduced. This is especially true in systems where the resolver and electronic evaluation system are geographically separated and interconnected, for example, by cables. Such systems are particularly vulnerable to interference because the measurement signals are transmitted via cables, which can lead to the coupling of interference signals from different sources along the cables.

[0013] However, the aforementioned problem not only occurs in resolvers, but in principle can occur in all inductive position sensors having an excitation winding and at least one secondary winding in which a measurement signal is induced due to the excitation field of the excitation winding. The excitation winding and the secondary winding can move relative to each other such that the measurement signal depends on the position. Summary of the Invention

[0014] The objective of this invention is to provide an easy-to-implement method that improves the accuracy of position information from an inductive position sensor.

[0015] This task is solved by: forming a frequency functional dependent on the excitation frequency from the measured signal, which represents a measure of the noise signal; changing the excitation frequency of the excitation signal to minimize or maximize the frequency functional; and using the excitation frequency that minimizes or maximizes the frequency functional for the excitation signal. By changing the excitation frequency, it is particularly possible to ensure that the excitation occurs within a frequency range where the noise signal is as weak as possible. Thus, the useful signal in this frequency range is minimally affected by noise, and the useful signal can be optimally extracted due to the high signal-to-noise ratio in this frequency range. Therefore, the excitation signal of the inductive position sensor can be optimally adapted to a specific application and the accuracy of the position information can be improved.

[0016] Advantageously, the amplitude of the measured signal can be examined using a frequency functional and its deviation from the expected signal trajectory, since noise signals are directly perceptible in terms of amplitude. This can be achieved by deriving amplitude information of the measured signal from the measured signal, and the frequency functional being a function of the amplitude information.

[0017] Of particular advantage, the measured signal is demodulated to determine amplitude information. Since the measured signal is an amplitude-modulated signal, amplitude information can be obtained technically in a simple manner by demodulation. For this purpose, known methods (e.g., the IQ method) can be used.

[0018] In the case of the IQ method, the I component representing the amplitude and the Q component representing the phase of the measured signal are determined. It is advantageous to set the phase shift of the excitation signal such that the I component of the measured signal is maximized and the Q component disappears.

[0019] For the implementation of position detection, it is advantageous to use the statistical variance of the amplitude information as a frequency functional. This can be particularly advantageously implemented in a digital implementation if the statistical variance is determined to be the expected squared deviation of the amplitude information value from the expected value of the amplitude information, wherein the arithmetic mean of the amplitude information is preferably used as the expected value. Here, the short-term variance of multiple sampled amplitude values ​​can be used as the variance. Attached Figure Description

[0020] The following will refer to Figures 1 to 7 To explain the invention in more detail, Figures 1 to 7 Advantageous design features of the invention are illustrated, by way of example and not limitation. The accompanying drawings show:

[0021] Figure 1 and Figure 2 The measurement principle of the resolver as an inductive position sensor is shown.

[0022] Figure 3 The spectrum of the noise signal is shown.

[0023] Figure 4 The frequency tuning of the excitation frequency of the excitation signal according to the present invention is shown.

[0024] Figure 5 Phase tuning according to the present invention is shown.

[0025] Figure 6 Amplitude tuning according to the present invention is shown, and

[0026] Figure 7 An advantageous implementation of an evaluation unit with frequency tuning, phase tuning, and amplitude tuning is shown. Detailed implementation

[0027] Figure 1 and Figure 2 The principle of the inductive position sensor 1 is illustrated using an example of a conventional resolver. The resolver has an excitation winding EW, which is excited by an excitation signal ES. The excitation signal ES is generated by a generation unit 3 (e.g., an electrical or electronic circuit). The excitation signal ES is an electrical alternating signal with a specific excitation amplitude R0 and an excitation frequency ω, i.e., for example, ES = R0·cos(ωt), where t represents time. If the excitation winding EW moves relative to the secondary winding SW, a voltage is induced in the secondary winding SW, which depends on the position P (angle ρ in this example) of the secondary winding SW relative to the excitation winding EW. For example, in the resolver, the two secondary windings SW are arranged to be offset from each other by 90° ( Figure 2 The excitation winding EW rotates in the resolver. However, a single secondary winding SW is sufficient. In the case of the linear inductive position sensor 1, a linear relative movement between the excitation winding EW and the secondary winding SW is provided, wherein multiple secondary windings may also be arranged sequentially in the direction of movement. The voltage induced in at least one secondary winding SW is output as a measurement signal MS of the inductive position sensor 1 and can be evaluated in the evaluation unit 2 to determine the position P of the moving part (typically the excitation winding EW) of the inductive position sensor 1.

[0028] In the case of a resolver with two secondary windings SW, the measurement signal MS comprises two measurement signal trajectories A and B. Due to the offset arrangement of the secondary windings SW, the measurement signal trajectories A and B have a specific phase offset.

[0029] The excitation winding EW is arranged on a moving part (e.g., on the motor shaft of an electric motor) in the inductive position sensor 1, and at least one secondary winding SW is arranged on a fixed part (e.g., on the housing of the position sensor 1, which in turn may be arranged on the motor housing). However, this arrangement can also be reversed. The moving part and the fixed part are arranged to be movable relative to each other.

[0030] The resolver acts as an inductive position sensor 1 and provides an excitation signal. (in Given the arrangement of the secondary winding SW with a 90° offset, the measurement signal trajectories A and B in the measurement signal MS were obtained, for example:

[0031]

[0032]

[0033] Here, u represents the known transmission ratio of the inductive position sensor 1. The delay is represented by the operating time of the inductive position sensor 1 and the processing (e.g., through a filter, etc.) of the measurement signal MS in the evaluation unit 2. Thus, the delay... follow It has a delay obtained from processing. and running time Δt. ρ represents the angular position of the excitation winding EW relative to the secondary winding SW (e.g., ...). Figure 2 (as shown), and thus represent the actual location of interest, P.

[0034] In another arrangement of the inductive position sensor 1, the measurement signal MS may of course appear different, having more or fewer measurement signal trajectories; however, the measurement signal MS always depends on the position P, i.e., MS(P). The measurement signal MS includes at least one measurement signal trajectory A, B.

[0035] The principle of the inductive position sensor 1 is well known. It is also known that the noise signal RS is typically superimposed on the measurement signal MS, such as... Figure 1 As shown. The noise signal RS is usually unknown and can include different frequency bands. The noise signal RS can be, for example, white noise WR, which also includes a narrow interference band SB within a specific frequency range. This is in Figure 3The spectrum S(f) (frequency f) of the noise signal RS is illustrated exemplarily. These are around the characteristic frequency f. P 2f P The narrow-band interference band SB can couple into the measurement signal MS from the environment (e.g., inductive ground). Of course, the noise signal RS can also have a different spectrum S(f). The noise signal RS largely determines the signal-to-noise ratio (SNR) of the measurement signal MS. In order to extract the actual useful signal from the measurement signal MS (e.g., measurement signal trajectories A, B), a maximum SNR, i.e., a large distance between the actual useful signal in the measurement signal MS and the noise signal RS, is advantageous.

[0036] The basic idea of ​​this invention is to use a frequency functional F that depends on the excitation frequency ω. f It is used as a measure of the noise signal RS in the measured signal MS, and the excitation frequency ω is changed so that the frequency functional F f Optimized, this is based on the frequency functional F f The expression corresponds to maximizing or minimizing. This makes the frequency functional F... f The minimized or maximized excitation frequency ω is then used in the excitation signal ES for the operation of the inductive position sensor 1.

[0037] Therefore, the frequency functional F, which is a measure of the noise signal RS in the measured signal MS, f This allows information about the noise signal RS in the measured signal MS to be obtained. This metric is, for example, a parameter that compares the noise signal RS to the measured signal MS or shows this to a scale. Frequency functional F f It can be mathematically represented as a mathematical function of the excitation frequency ω.

[0038] Of course, the frequency functional F f It can be represented in different ways; for example, if the noise signal RS can be separated from the measured signal MS (e.g., through filtering), it can be represented as the signal-to-noise ratio. This can be achieved using the frequency functional F. f The amplitude of the measured signal MS can also be examined by its deviation from the expected signal trajectory of the measured signal MS (specifically, at least one measured signal trajectory). This deviation is a measure of the noise signal RS. In the case of a resolver, this is, for example, a modulated sine wave, where the noise is readily apparent as a deviation from the modulated sine wave. For this purpose, for example, as the frequency functional F... f The statistical variance of the amplitude of the measured signal can be used as a measure of this deviation, and thus as a measure of the noise signal RS. For this purpose, amplitude information FA of the measured signal MS can be derived from the measured signal MS or from at least one measured signal trajectory of the measured signal MS, which is then used in the frequency functional F fThe amplitude information FA is evaluated. Here, the amplitude information FA can depend on the excitation frequency ω. Therefore, the frequency functional F... f It is a function of the amplitude information FA, i.e., F f (FA(ω)), and thus depends on the excitation frequency ω.

[0039] according to Figure 4 A particularly advantageous design scheme of the present invention is described. In this method, at least one measurement signal trajectory A, B is demodulated to obtain a useful signal, i.e., a parameter that oscillates depending on the position P, such as cos(ρ) or sin(ρ). From this demodulation, the amplitude information FA of the measurement signal MS can be obtained, and subsequently, from the amplitude information, the frequency functional F can be constructed as described above. f Due to the time delay between the excitation signal ES and the measurement signal MS... With phase shift, the known IQ method (in-phase and quadrature methods) is suitable for demodulation because it yields both amplitude and phase information. For demodulation, the measured signal MS is multiplied by the excitation oscillation cos(ωt) to obtain the I component representing the amplitude. If the measured signal MS is optionally multiplied by the excitation oscillation with a 90° phase difference, phase information (Q component) can also be obtained. Mathematically, in the case of a resolver, this can be represented as a periodic scalar product, where there is typically at least one measured signal trajectory A, B.

[0040]

[0041]

[0042] And can be chosen or required

[0043]

[0044]

[0045] I component A I B I It contains amplitude information FA and can also be squared: FA = A I 2 For further use, or alternatively, the absolute value FA = |A can be used. I In the case of multiple measurement signal trajectories A and B, the I component A can also be used. I B I The sum of squares FA = A I 2 +B I 2 Alternatively, the sum of absolute values ​​FA = |A I |+|B IIn principle, other arithmetic relations for the I component are also conceivable.

[0046] For example, the statistical variance V of the amplitude information FA, which depends on the excitation frequency ω, can then be used as the frequency functional F. f Variance V can be specified as the expected squared deviation of the magnitude information FA from its expected value E(FA) (i.e., the expected value E). The arithmetic mean can be used as the expected value E(FA). Mathematically, this can be expressed in the following form:

[0047] F f =V(FA)=E((FA-E(FA)) 2 ).

[0048] Multiple (N) values ​​of the amplitude information FA are preferably used to determine the variance V, which can be obtained by sampling the measurement signal MS at a pre-given sampling frequency (e.g., in the megahertz range). At a typical excitation frequency of 10 kHz, a sampling frequency of 1 MHz will result in a hundredfold oversampling. Therefore, the variance V can be expressed as the short-term variance in the following form:

[0049]

[0050] For example, 1,000 to 10,000 cycles of the excitation signal ES are used for determination, which will result in an observation period of 100 ms to 1 s for determining the value of the variance V.

[0051] Now, the frequency functional F f It has been optimized in terms of the excitation frequency ω, which can be mathematically achieved through... or This is represented by the excitation frequency ω, which therefore changes until the function F... f It is optimized (in the sense of minimization or maximization). Here, the excitation frequency ω varies within a pre-given frequency range.

[0052] To solve such optimization problems, numerous known solution algorithms exist, to name just a few: gradient methods, Newton's method, evolutionary methods, or successive quadratic programming. While the choice of solution algorithm is irrelevant to this invention, it is of course preferred in terms of computational effort and time. A common thread among these methods is the search for (usually iterative) possible solutions to the optimization problem until a defined termination criterion is reached. For example, the termination criterion could be the number of iterations, the difference between the solutions of two successive iterations of the optimization problem falling below a limit, or another termination criterion. The selection of the solution (i.e., the excitation frequency ω) is made in each iteration step according to pre-defined rules of the solution method, where a suitable choice of solution can be pre-defined as a starting value in the first iteration step. For example, in the gradient method, the gradient of the functional (the derivative of the functional with respect to the excitation frequency) is determined, and the set parameters for the next iteration step are selected along this gradient, where the step size from the current set parameter to the next set parameter is determined by pre-defined rules of the solution method.

[0053] In this way, the excitation frequency ω can be determined, and this excitation frequency must be used to excite the inductive position sensor 1 to minimize the influence of the noise signal RS. This, in particular, avoids narrow-band interference. If, for example, a frequency f is chosen... P Excitation frequency ω (within the range of (or integer multiples thereof) Figure 3 If the signal-to-noise ratio of the measured signal MS decreases compared to the excitation frequency ω which is far from it, then the signal-to-noise ratio of the measured signal MS will decrease.

[0054] One possible implementation of frequency tuning according to the invention as described above is in Figure 4 As shown in the figure. This example again relates to a decomposer having at least one measurement signal trajectory A in the measurement signal MS.

[0055] First, demodulate the measurement signal trajectory A and determine the I component A. I The square of this value determines the amplitude information FA. This can be done in the frequency tuner FT. Similar to the resolver, the second measured signal trajectory B is... Figure 4 The area is represented by a dashed line. In this example, the amplitude information FA will be the I component A obtained by squaring the measured signal trajectories A and B. I B I The variance V(FA) of the amplitude information FA (e.g., as the short-term variance as described above) is used as the frequency functional F. f Frequency functional F f The excitation frequency ω is minimized, and the inductive position sensor 1 subsequently operates at this excitation frequency ω. This ensures the largest possible signal-to-noise ratio in corresponding applications with predominantly noisy environments.

[0056] exist Figure 4 In determining component A of component II B I A low-pass filter (LPF) is then shown. Such a low-pass filter (LPF) can be used to suppress the double excitation frequency ω generated during demodulation.

[0057] This frequency tuning can be performed once when the inductive position sensor 1 is put into operation, and the excitation frequency ω can then remain constant under the assumption of a constant interference environment. Here, it is only necessary to ensure that the interference sources in the environment are active during the frequency tuning period.

[0058] However, frequency tuning can also be performed continuously (e.g., at predetermined time intervals) during the operation of the inductive position sensor 1.

[0059] For example, sustained frequency tuning can be performed by (e.g., using FFT (Fast Fourier Transform)) performing sustained spectral analysis on at least one measured signal trajectory A, B, and quantizing the noise power density of the noise signal RS in the spectrum as a frequency functional F. f Here, the minimum noise power density can be identified, and the excitation frequency ω can then be set at the location with the minimum noise power density.

[0060] The frequency tuning according to the present invention can also be combined with phase tuning. Thus, the I component A of the demodulated measurement signal trajectories A and B can be realized. I B I Maximum (which is beneficial for frequency tuning) and Q component A Q B Q Disappearance. This can be achieved, for example, using a phase adjuster. To execute this, the phase adjuster adjusts the phase shift of the excitation signal ES. So that the Q component A Q B Q disappear.

[0061] For example, this can be achieved by using a phase functional in the case of a resolver. This is achieved by using the scalar product again in the phase functional.

[0062] The phase functional It does not depend on the position ρ and obtains The function is in the interval There are four zeros, and at two of these zeros (at π / 2 and 3π / 2), the I component A I B I Disappears, and the Q component A Q B Q The Q component A is at its maximum at the other two zeros (at 0 and π). Q B QDisappears, and component A of I disappears. I B I maximum.

[0063] Phase adjuster Now we can use the phase functional Adjust to zero, that is Phase adjuster Ensure phase shift Locked at 0 or π. Phase adjuster For example, it can be designed as a known I-regulator (integral regulator).

[0064] Phase tuner PT can be like Figure 5 As shown. Phase setting value. It is set to zero so that the phase shift By phase adjuster Adjustment to make the phase functional It becomes zero. Then, a phase shift is set in the excitation signal ES.

[0065] The frequency tuning according to the invention can also be combined with amplitude tuning, and if necessary, with phase tuning. The idea behind amplitude tuning is to set the excitation amplitude R0 of the excitation signal ES to its maximum value. This maximum value depends on the implementation of the generation unit 3 for the excitation signal ES. For example, the amplitude R0 can be set to a value that utilizes the entire linear range of the electronic circuitry serving as the generation unit 3. This can be achieved, for example, using an amplitude modulator R that adjusts the excitation amplitude R0 to a predetermined set value. A To execute.

[0066] For example, this can be achieved by using the magnitude functional F in the case of a resolver. A =A I 2 +B I 2 +A Q 2 +B Q 2 This is achieved. If the aforementioned phase tuning is additionally used, the terms with Q components in the amplitude functional F A The magnitude functional F disappears. A It depends neither on the position ρ nor on the phase shift And based on the above definitions of the I and Q components, we obtain

[0067] Amplitude regulator R A Now we can define the magnitude functional F A Adjust to the predetermined amplitude setting value F Aset Here, the amplitude setting value F AsetThis can be obtained from the implementation of unit 3. Amplitude regulator R A For example, it can be designed as a known I-regulator (integral regulator).

[0068] Amplitude tuner AT can be like Figure 6 This is implemented as shown. Subsequently, the excitation amplitude R0 is set in the excitation signal ES.

[0069] Figure 7 The diagram illustrates advantageous position determination using a resolver, which functions as an inductive position sensor 1. The resolver consists of a phase shifter with an excitation frequency ω. The excitation signal ES with excitation amplitude R0 is used for excitation. The measurement signal trajectories A and B of the measurement signal MS output by the resolver are evaluated in evaluation unit 2. The measurement signal trajectories A and B are demodulated using the IQ method, and the I component A is determined. I B I and Q component A Q B Q Frequency functional F f Phase functional and amplitude functional F A A composed of I and Q components I B I A Q B Q Confirmed. Subsequently, the excitation frequency ω is set using a frequency tuner FT, in which the frequency tuning according to the present invention is implemented. Similarly, the phase shift of the excitation signal ES is set using a phase tuner PT. And the excitation amplitude R0 is set using the amplitude tuner AT.

[0070] Of course, the position P (in this case, the angle ρ) can also be determined by the I component A. I B I (or optionally, Q component A) Q B Q (For example, using the atan2 function) to determine.

[0071] The evaluation unit 2 is preferably implemented digitally as software on microprocessor-based hardware. Alternatively, it can be implemented on an integrated circuit, such as a field-programmable gate array (FPGA) or application-specific integrated circuit (ASIC). Of course, analog implementation is also not excluded.

[0072] In the digital implementation, the measurement signal trajectories A and B output by position sensor 1 are converted from analog to digital using a suitable analog-to-digital converter (ADC). Similarly, the excitation signal ES can be digitally generated and applied to position sensor 1 using a digital-to-analog converter (DAC). The position value P can also be output either analogly or digitally.

Claims

1. A method for determining the position (P) of a moving part relative to a fixed part by means of an inductive position sensor (1), wherein an excitation winding (EW) is arranged on the moving part, and at least one secondary winding (SW) is arranged on the fixed part, or vice versa, and an electrical excitation signal (ES) having an excitation frequency (ω) and an excitation amplitude (R0) is fed into the excitation winding (EW), the electrical excitation signal (ES) generating an electromagnetic excitation field, the electromagnetic excitation field inducing a measurement signal (MS) in the secondary winding (SW), the measurement signal (MS) being detected, and a noise signal (RS) being superimposed on the measurement signal (MS), and wherein the position of the moving part of the position sensor (1) is determined by the measurement signal (MS), characterized in that, A frequency functional (F) is formed from the measured signal (MS) that depends on the excitation frequency (ω). f The frequency functional represents a measure of the noise signal (RS), and the excitation frequency (ω) of the electrical excitation signal (ES) is changed such that the frequency functional (F) f ) is minimized or maximized, and the frequency functional (F) is minimized or maximized. f The excitation frequency (ω) that is minimized or maximized is used for the electrical excitation signal (ES).

2. The method as described in claim 1, characterized in that, Using the frequency functional (F f The amplitude of the measurement signal (MS) is checked by examining the deviation from the expected signal trajectory of the measurement signal (MS).

3. The method as described in claim 1, characterized in that, The amplitude information (AI) of the amplitude of the measured signal (MS) is derived from the measured signal (MS), and the frequency functional (F) f ) is a function of the amplitude information (AI).

4. The method as described in claim 3, characterized in that, The measurement signal (MS) is demodulated to determine the amplitude information (AI).

5. The method as described in claim 4, characterized in that, The measurement signal (MS) is demodulated using the excitation oscillation of the electrical excitation signal (ES) to determine the amplitude information (AI).

6. The method as described in claim 5, characterized in that, The measurement signal (MS) is demodulated using the excitation oscillation of the electrical excitation signal (ES) to determine the I component (A) representing the amplitude of the measurement signal (MS). I B I ).

7. The method as described in claim 5, characterized in that, The measurement signal (MS) is demodulated using an excitation oscillation of the electrical excitation signal (ES) with a 90° phase shift to determine the Q component representing the phase of the measurement signal (MS).

8. The method as described in claim 3, characterized in that, The statistical variance (V) of the amplitude information (AI) is used as the frequency functional (F). f ).

9. The method as described in claim 8, characterized in that, The statistical variance (V) is determined as the expected squared deviation between the value of the amplitude information (AI) and the expected value of the amplitude information (AI), wherein the arithmetic mean of the amplitude information (AI) is used as the expected value.

10. The method as described in claim 6, characterized in that, The phase shift (Δφ) of the electrical excitation signal (ES) is set such that the I component (A) of the measurement signal (MS) is... I B I )maximum.

11. The method according to any one of claims 1 to 10, characterized in that, The excitation amplitude (R0) is adjusted to a pre-defined amplitude setting value.

12. An evaluation unit for an inductive position sensor (1), the inductive position sensor (1) comprising a moving part having an excitation winding (EW) and a fixed part having at least one secondary winding (SW), or conversely, wherein an electrical excitation signal (ES) having an excitation frequency (ω) and an excitation amplitude (R0) is fed into the excitation winding (EW), the electrical excitation signal (ES) generating an electromagnetic excitation field, the electromagnetic excitation field inducing a measurement signal (MS) in the secondary winding (SW), the measurement signal (MS) being fed to the evaluation unit (2), and a noise signal (RS) being superimposed on the measurement signal (MS), and wherein the evaluation unit (2) determines the position (P) of the moving part of the position sensor (1) relative to the fixed part from the measurement signal (MS), characterized in that, The evaluation unit (2) forms a frequency functional (F) from the measurement signal (MS) that depends on the excitation frequency (ω). f The frequency functional represents a measure of the noise signal (RS), and the evaluation unit (2) determines the frequency functional (F) of the electrical excitation signal (ES). f Minimize or maximize the excitation frequency (ω), and make the frequency functional (F) f The excitation frequency (ω) that is minimized or maximized is used in the case of the evaluation unit (2) for the electrical excitation signal (ES).

Images