Eddy current sensor sensitivity detection circuit based on four-arm bridge

Abstract

The invention provides an eddy current sensor sensitivity detection circuit based on a four-arm bridge, belongs to the technical field of sensor performance detection, and can at least partially solve the problem that an existing eddy current sensor is insufficient in sensitivity. The eddy current sensor sensitivity detection circuit based on the four-arm bridge comprises a four-arm bridge topological structure which comprises a symmetrical resistance arm formed by a first fixed resistor and a second fixed resistor and a variable impedance arm formed by probe equivalent impedance and reference impedance; the probe equivalent impedance and the reference impedance are connected in parallel with the LC resonant network; the differential amplifier module is electrically connected to the four-arm bridge topological structure and is used for amplifying a differential voltage signal output by the bridge; and the dynamic balance calibration unit is electrically connected between the four-arm bridge topological structure and the differential amplifier module and comprises an adjustable impedance module and a feedback control module so as to realize power-on self-test calibration and periodic dynamic calibration.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of sensor performance detection circuit, in particular to a kind of eddy current sensor sensitivity detection circuit based on four-arm bridge, it is suitable for precision displacement measurement, rotating machinery vibration monitoring etc. Industrial scene. BACKGROUND

[0002] Eddy current sensor detects metal conductor displacement using electromagnetic induction principle, and is widely used in equipment condition monitoring. Eddy current sensor is based on electromagnetic induction principle, and realizes non-contact measurement by detecting the coil impedance change caused by the eddy current effect on the surface of metal conductor.

[0003] Traditional analog sensor mostly uses single-arm or half-bridge structure, and its sensitivity is limited by the topology characteristics of the bridge. The typical output sensitivity is less than 1V / μm, and there is more than 2% nonlinear error. The response of such structure to small displacement (<0.1μm) or high-speed vibration signal is easily overwhelmed by noise, and the signal-to-noise ratio is generally less than 40dB, which leads to false detection or missed detection in rotating machinery shaft vibration monitoring and other scenes.

[0004] An improved scheme in the prior art introduces a differential bridge design to improve sensitivity, but does not fully solve the problem of zero drift in dynamic detection: the baseline drift caused by temperature change or electromagnetic interference can reach 0.1% / ℃ of full scale, which significantly affects long-term stability. In addition, the impedance matching network of the traditional scheme mostly uses fixed parameter design, which is difficult to balance high frequency (>100kHz) dynamic response and low frequency (<10kHz) measurement accuracy, which restricts its applicability in complex working conditions.

[0005] That is, the prior art mostly uses single-arm / half-bridge circuit, which has three major defects: first, low sensitivity: output voltage sensitivity ≤1V / μm, it is difficult to detect <0.1μm small displacement; second, weak anti-interference ability: signal-to-noise ratio (SNR) ≤40dB, power frequency noise easily overwhelms the effective signal; third, poor stability: temperature drift reaches 0.1%FS / ℃, high frequency (>100kHz) amplitude attenuation >3dB. The improved scheme in the prior art, such as using differential bridge, still has the problem of dynamic zero drift, and the fixed parameter LC network cannot balance the wide frequency band response.

[0006] Therefore, there is an urgent need for a detection circuit that balances high sensitivity, low drift and wide dynamic range to meet the rigid demand for high-precision displacement detection in precision manufacturing and high-end equipment monitoring fields. SUMMARY

[0007] The present application aims to at least solve one of the technical problems existing in the prior art, and provides an eddy current sensor sensitivity detection circuit based on four-arm bridge.

[0008] One aspect of the present application provides a four-arm bridge-based eddy current sensor sensitivity detection circuit, comprising: a four-arm bridge topology, which comprises symmetric resistance arms formed by a first fixed resistance and a second fixed resistance, and a variable impedance arm formed by a probe equivalent impedance and a reference impedance; a probe equivalent impedance and a reference impedance in parallel LC resonance network; a differential amplifier module electrically connected to the four-arm bridge topology for amplifying the differential voltage signal of the bridge output; and a dynamic balance calibration unit electrically connected between the four-arm bridge topology and the differential amplifier module, and comprising an adjustable impedance module and a feedback control module to realize power-on self-test calibration and periodic dynamic calibration.

[0009] Preferably, the resonance frequency of the LC resonance network is consistent with the probe operating frequency, and the error is ≤±1%.

[0010] Optionally, the common-mode rejection ratio of the differential amplifier module is ≥90dB, and the gain bandwidth product is ≥10MHz.

[0011] Specifically, the differential amplifier module uses AD8421 amplifier, the input impedance is ≥1GΩ, and the output signal SNR is ≥60dB.

[0012] Optionally, the adjustable impedance module is a parallel structure of balanced impedance and balanced capacitor, wherein the dynamic balance calibration unit adjusts the resistance value by using a digital potentiometer as the balanced impedance, and performs PID feedback adjustment.

[0013] Optionally, in the LC resonance network in parallel with the probe equivalent impedance and the reference impedance, the inductance and capacitance parameters of the LC resonance network satisfy the relationship: ; Where f is the probe operating frequency, L is the inductance of the LC resonance network, and C is the capacitance of the LC resonance network.

[0014] Preferably, the four-arm bridge topology uses constant voltage source power supply, the first fixed resistance and the second fixed resistance have consistent resistance values, and the error is ≤0.1Ω, and the output sensitivity is ≥2V / μm.

[0015] Optionally, the capacitor in the adjustable impedance module selects ceramic capacitor 100pF±0.5pF, and the equivalent series resistance ESR is ≤0.1Ω.

[0016] Preferably, the feedback control module has a temperature compensation unit to trigger a pre-calibration program to adjust the resistance value of the adjustable impedance module when the ambient temperature change ΔT is ≥0.5℃.

[0017] Especially preferably, the circuit wiring of the eddy current sensor sensitivity detection circuit satisfies: the wiring length symmetry error of the four-arm bridge topology is ≤0.1 mm, and there is no digital signal wiring within 20 mm around the differential amplifier module.

[0018] The application provides an eddy current sensor high-sensitivity detection circuit based on a four-arm bridge, which realizes high-sensitivity, high-resolution and strong anti-interference capability by a full-symmetry topology structure (two fixed resistors and two impedance arms form a balanced bridge circuit) of the four-arm bridge and a high-performance signal processing circuit, and specific indexes include: sensitivity ≥ 2 V / μm, zero drift < 0.008% FS / ℃, and signal-to-noise ratio (SNR) of 63 dB in a high-frequency band (10 kHz-1 MHz).

[0019] In the above embodiment, the application provides an eddy current sensor high-sensitivity detection circuit based on a four-arm bridge, which breaks through the technical bottleneck of traditional eddy current sensor sensitivity and stability by the synergistic design of a symmetric bridge topology structure and a dynamic balance calibration technology, adopts the four-arm bridge as a core detection unit, adopts two fixed resistors and two impedance arms (probe equivalent impedance and reference impedance) to form a full-symmetry bridge, and amplifies the small change amount ΔZ of the probe impedance through differential output, and the theoretical basis lies in the bridge balance principle: in the application process, when the distance between the probe and the metal conductor changes, the impedance change of the probe equivalent impedance will break the initial balance state of the bridge, the differential voltage output end generates a signal ΔV proportional to ΔZ, and the sensitivity expression is Where ΔZ is the probe impedance change amount, and Z0 is the initial impedance. Compared with the traditional half-bridge circuit, the voltage sensitivity of the four-arm bridge can be theoretically improved by 2 times, the experimental verification output sensitivity is 2.1 V / μm, and the sensitivity is improved by 52% compared with the prior art. In the signal processing link, the differential amplifier module is innovatively introduced, which can effectively amplify the differential signal and suppress the environmental noise, through the differential output mode, the signal-to-noise ratio (SNR) of the system is improved to 63 dB, which is significantly better than the 40 dB level of the traditional single-end output scheme, especially for the high-speed rotating equipment detection scene, the LC resonant network is designed to ensure the signal transmission efficiency in the 10 kHz-1 MHz wide frequency band, and the problem of response attenuation of the traditional fixed parameter matching circuit in the high-frequency band is solved. Especially, in order to overcome the problem of zero drift caused by temperature drift and electromagnetic interference, the dynamic balance calibration unit is adopted, which realizes real-time compensation of the initial unbalanced voltage of the bridge by integrating the adjustable impedance and the feedback control module. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The application provides a topology structure diagram of an eddy current sensor sensitivity detection circuit based on a four-arm bridge.

[0021] Figure 2The flowchart below shows the dynamic balance calibration process of the eddy current sensor sensitivity detection circuit according to a specific embodiment of the present invention, which includes feedback control algorithm logic. Detailed Implementation

[0022] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] like Figure 1 As shown, the sensitivity detection circuit of the eddy current sensor based on a four-arm bridge in the basic embodiment of the present invention includes: a four-arm bridge topology, which includes a symmetrical resistance arm formed by a first fixed resistor R1 and a second fixed resistor R2, and a variable impedance arm formed by a probe equivalent impedance Z1 and a reference impedance Z2, wherein the probe equivalent impedance Z1 and the reference impedance Z2 are connected in parallel to an LC resonant network; a differential amplifier module, which is electrically connected to the above-mentioned four-arm bridge topology to amplify the differential voltage signal output by the bridge; and a dynamic balance calibration unit, which is electrically connected between the four-arm bridge topology and the differential amplifier module, and includes an adjustable impedance module and a feedback control module to realize power-on self-test calibration and periodic dynamic calibration.

[0024] In the above basic implementation, the high-sensitivity detection circuit of the eddy current sensor based on the four-arm bridge of the present invention employs: (1) A four-arm bridge topology, in which the first fixed resistor R1 and the second fixed resistor R2 form symmetrical resistor arms, the equivalent impedance of the probe Z1 and the reference impedance Z2 form a variable impedance arm, and the equivalent impedance of the probe Z1 and the reference impedance Z2 are connected in parallel resonant network. Compared with the traditional half-bridge / single-arm scheme, the voltage sensitivity is theoretically improved by more than 2 times, and the measured value reaches 2.1V / μm (+52% compared with the traditional scheme). In the specific implementation of the above four-arm bridge topology, the equivalent impedance of the probe Z1 and the reference impedance Z2 are connected in parallel resonant network. The equivalent impedance of the probe Z1 and the reference impedance Z2 can be built into an LC resonant network. The resonant frequency can generally be kept consistent with the probe operating frequency, with an error ≤±1%, ensuring the signal transmission efficiency of the 10kHz-1MHz wideband (amplitude attenuation of the high-frequency band <0.5dB). In terms of initial matching accuracy, the initial error of the four-arm impedance is ≤0.1Ω, and the power supply can use a 5V±0.1% constant voltage source to suppress zero-point offset from the hardware level. This fully symmetrical structure can use differential output to cancel common-mode interference (such as temperature drift and electromagnetic noise), resulting in a nonlinear error ≤0.2%FS, which meets the stringent linearity requirements of precision measurement.

[0025] (2) A differential amplifier module is used to amplify the differential voltage signal of the bridge output. In some embodiments, the common-mode rejection ratio (CMRR) is ≥ 90 dB, which can effectively suppress environmental noise (such as power frequency interference, radio frequency coupling), and the measured signal-to-noise ratio (SNR) reaches 63 dB, which is significantly improved compared to the traditional single-ended scheme (40 dB), and the high-frequency distortion is ≤ 0.0087%. In addition, the gain-bandwidth product is ≥ 10 MHz, which can match the high-frequency dynamic response requirement (10 kHz-1 MHz) of the eddy current sensor, cover the target frequency band, and ensure that the signal is amplified without distortion. Referring to Figure 1 As shown in the circuit configuration, the gain G = 100 V / V can be achieved by laser trimming resistors Rf = 99 kΩ and Rg = 1 kΩ, and the output swing 0-5 V matches the subsequent ADC (analog-to-digital converter).

[0026] (3) A dynamic balance calibration unit solves the problem of baseline fluctuation caused by temperature drift and electromagnetic interference in traditional schemes, and realizes long-term stability optimization. In specific implementation, it can include an adjustable impedance module (R3C3 parallel network), so that through self-checking and calibration at startup, for example, referring to Figure 1 As shown, the initial unbalanced voltage Vout is ≤ 10 μV by adjusting the R3C3 parallel network through the digital potentiometer AD5207. At the same time, periodic dynamic calibration is performed: for example, PID feedback adjustment is triggered once every 5 minutes to compensate for the temperature coefficient (≤ 0.01% FS / ℃), and the measured zero-point drift is suppressed to 0.008% FS / ℃ (≤ 0.01% FS / ℃), which is at least two orders of magnitude higher than static calibration.

[0027] On the basis of the above basic embodiments of the present application, further, in some embodiments, as described above, the probe equivalent impedance Z1 and the reference impedance Z2 can include a built-in LC resonance matching network, the resonance frequency of which is consistent with the working frequency of the probe, with an error of at least ≤ ± 1%, to improve the signal transmission efficiency in the high frequency band of 10 kHz-1 MHz.

[0028] In some specific implementation embodiments, the common-mode rejection ratio of the above-mentioned differential amplifier module is ≥ 90 dB, and the gain-bandwidth product is ≥ 10 MHz. For example, the differential amplifier module can be selected from AD8421 amplifier or devices with equivalent performance, with an input impedance ≥ 1 GΩ and an output signal-to-noise ratio (SNR) ≥ 60 dB. Alternatively, the differential amplifier output is connected to a second-order Butterworth low-pass filter with a cutoff frequency of 200 kHz.

[0029] In some embodiments, the four-arm bridge power supply uses a constant voltage source, for example 5V±0.1%, and the initial matching error of the four-arm impedance is ≤ 0.1Ω, the measurement range is 0-2mm, and the output sensitivity is ≥ 2V / μm.

[0030] As submitted above, in some embodiments, the adjustable impedance module can be a parallel structure of balancing impedance R3 and balancing capacitor C3. The dynamic balance calibration unit adjusts the resistance value of R3 using a digital potentiometer, such as an AD5207 digital potentiometer, with PID feedback to ensure the initial imbalance voltage is ≤10μV. PID feedback adjustment is triggered every 5±0.5 minutes, resulting in zero-point drift ≤0.008%FS / ℃. Specifically, during the power-on self-test phase, the resistance value of R3 is adjusted via the digital potentiometer to ensure the output voltage Vout ≤10μV. During dynamic detection, a calibration procedure is triggered every 5 minutes, using a feedback loop to correct for temperature coefficient (<0.01%FS / ℃) and baseline offset caused by electromagnetic interference. This method improves long-term stability to 0.008%FS / ℃ through periodic dynamic adjustment, solving the problem of measurement error accumulation caused by environmental changes in traditional solutions.

[0031] In an LC resonant network where the equivalent impedance Z1 of the probe is connected in parallel with the reference impedance Z2, the inductance and capacitance parameters of the LC resonant network satisfy the following relationship: ; Where f is the probe's operating frequency, L is the inductance of the LC resonant network, and C is the capacitance of the LC resonant network.

[0032] In some specific implementations, the capacitor C3 in the adjustable impedance module is a ceramic capacitor with a capacitance of 100pF ± 0.5pF and an equivalent series resistance ESR ≤ 0.1Ω.

[0033] In some preferred embodiments, the feedback control module may have a temperature compensation unit to trigger a pre-calibration procedure to adjust the resistance value of the adjustable impedance module when the ambient temperature changes by ΔT ≥ 0.5℃.

[0034] Particularly preferred is that the circuit wiring of the eddy current sensor sensitivity detection circuit of the present invention satisfies the following: the symmetry error of the trace length of the four-arm bridge topology is ≤0.1mm, and there are no digital signal traces within 20mm around the differential amplifier module.

[0035] To help those skilled in the art to more deeply understand and specifically implement this invention, the following is combined with... Figure 1 and Figure 2 This description provides a more comprehensive and specific embodiment of the present invention, including detailed technical specifications. It should be noted that the following embodiments containing technical details are for illustrative purposes only and should not be construed as limiting the scope of protection of the invention to the specific parameters or parameter ranges described in the embodiments.

[0036] Regarding hardware circuit setup and initial configuration: (1) PCB design and impedance matching Four-layer printed circuit board (PCB) design is adopted, and the layering structure is divided according to the signal integrity rules: the top layer is used for high-frequency signal wiring (bridge and differential amplifier), the bottom layer is arranged with power supply and ground plane, the inner layer one is digital control signal layer, and the inner layer two is analog reference signal layer. The wiring rules need to meet: first, the wiring length of the four arms of the bridge is strictly symmetrical (error ≤0.1 mm), and 50Ω characteristic impedance control is adopted to reduce high-frequency signal reflection; second, the pad spacing between the probe coil and the reference impedance Z2 needs to reserve thermal expansion compensation (design according to the temperature expansion coefficient CTE=17ppm / ℃); third, digital signal wiring is prohibited within a range of 20 mm around the differential amplifier AD8421 to prevent crosstalk.

[0037] (2) Fixed resistance and probe coil assembly The first and second fixed resistances R1 / R2 are metal film resistors with a temperature coefficient ≤5ppm / ℃, and the four-wire measurement method is used to ensure ΔR / R<0.05%. The probe coil is made of single-crystal copper wire with a diameter of 0.2mm, the number of turns N=500±1, and the inductance L=10μH±2%. A constant tension winding machine can be used during winding, the paint film thickness is controlled within 0.02mm±0.005mm, and polyimide film (0.05mm) is used for interlayer insulation. Low-impedance solder (such as Sn96.5Ag3.0Cu0.5) is used for the connection between the coil and the PCB, and after soldering, the cavity rate is ≤1% through X-ray detection.

[0038] The four-wire measurement method separates the current loop and the voltage detection loop, eliminates the interference of lead resistance and contact resistance on the measurement results, and is suitable for high-precision resistance value calibration. The specific implementation steps can be as follows: first, hardware connection and device configuration: select a high-precision four-wire resistance meter such as Keithley 2000 digital multimeter which has independent current source (I+, I-) and voltage detection terminal (V+, V-). The measured resistance uses a metal film resistance with a nominal value of 1kΩ and a temperature coefficient of ≤5ppm / ℃. When connected, a constant current of 1mA is output from the I+ terminal, returned from the I- terminal after passing through the resistance; the V+ and V- terminals directly contact the two ends of the resistance to collect the voltage, avoiding the voltage division of the lead resistance. Second, measurement process and error suppression: the resistance meter has a built-in constant current source with a current fluctuation of ≤0.01%, and an amplifier with an input impedance of ≥10GΩ is used to collect the voltage, and 10 times of sampling are taken to suppress noise. According to Ohm's law, the resistance value is calculated by the constant current and the measured voltage, and the influence of lead resistance is eliminated. Compared with two-wire, four-wire directly connects the voltage detection terminal to the two ends of the resistance, so that the measured resistance value is equal to the true resistance value. Third, precision guarantee measures: in a constant temperature environment of 25℃±0.1℃, the PCB board is preheated for 2 hours to stabilize the resistance value, gold-plated probes and other low-oxidation contacts are used, the contact resistance is ≤0.1mΩ, the pins are cleaned before measurement, and the existing built-in algorithm is used to compensate for temperature drift and correct resistance value. Use the 3σ principle to eliminate abnormal points and ensure ΔR / R<0.05%.

[0039] (3) Reference impedance Z2 construction and resonant frequency debugging The reference impedance Z2 can be connected in parallel with an adjustable inductor (Coilcraft 1010-510J) and an NP0 ceramic capacitor (100nF±1%). The resonant frequency The specific debugging steps are as follows: first, use a vector network analyzer (Keysight E5061B) to sweep the frequency to determine the self-resonant frequency f0=105kHz of the probe. Second, adjust the C3 capacitor value to make it pure resistance (phase angle θ=0°±0.5°); ③ measure the insertion loss, and after optimization, the amplitude attenuation in the high frequency band (100kHz-1MHz) is <0.4dB.

[0040] (4) Differential amplifier and ADC interface design The AD8421 is configured in G = 100V / V gain mode, and the feedback resistor network uses a laser trimmed thin-film resistor (Rf = 100kΩ ± 0.1%), with a matching error < 0.02%. When the output is connected to a 16-bit ADC (AD7621), note the following: first, the differential amplifier output swing needs to match the ADC input range (0-5V); second, a second-order Butterworth low-pass filter (cutoff frequency 200kHz) is added between the AD8421 and the AD7621 to suppress out-of-band noise; fourth, shielded cables are used to connect the analog front end and digital circuit, and the shield layer is single-point grounded to the PCB bottom layer.

[0041] (5) Dynamic balance calibration unit integration The dynamic balance calibration unit is composed of an R3C3 parallel network and a feedback algorithm: first, R3 uses an AD5207 digital potentiometer, with a non-linearity < 0.1% INL and a temperature drift coefficient ≤ 35ppm / ℃; second, C3 selects a C0G ceramic capacitor (100pF ± 0.5pF) with an ESR < 0.1Ω; third, the microcontroller (STM32F407) has a built-in PID algorithm that collects Vout every 5 minutes and compares it with the threshold value to calculate the adjustment amount where Vref = 5V reference voltage. During debugging, the single adjustment step is ≤ 0.05%, and the maximum cumulative adjustment amount does not exceed 2% of the initial value of R3.

[0042] As for the initial balance debugging of the bridge: The initial balance debugging of the bridge can generally be completed in a constant temperature environment to eliminate the influence of temperature gradient on impedance matching. Before debugging, use a four-wire resistance meter to calibrate the first and second fixed resistors R1 and R2 to ensure that their resistance deviation ΔR / R < 0.1%. This measurement method separates the current lead and voltage detection lead to eliminate contact resistance interference, meeting the requirements of Wheatstone bridge for bridge arm resistance accuracy. After calibration, place the PCB board in a constant temperature oven, set the temperature to 25℃ ± 0.1℃, and preheat for 2 hours to eliminate impedance drift caused by material thermal stress.

[0043] After power supply startup, use a Tektronix MDO3024 oscilloscope to monitor the bridge output voltage Vout. If the initial Vout > 10μV, it indicates that the bridge is in an unbalanced state. At this time, start the microcontroller (STM32F407) program to dynamically adjust the balance impedance R3 resistance value through the digital potentiometer AD5207. The specific adjustment logic is as follows: first, the program increases / decreases the balance impedance R3 resistance value by 0.1% steps, waits for 5 seconds to stabilize the system after each adjustment; second, real-time acquisition of Vout and calculation of common mode rejection ratio (CMRR) are performed, and when CMRR ≥ 90dB and Vout ≤ 10μV, the adjustment is terminated; third, the final balance impedance R3 resistance value is recorded to EEPROM as the initial reference value for dynamic balance calibration. It should be noted here that the common-mode rejection ratio CMRR is a key indicator to measure the ability of differential amplifier to suppress common-mode signals, defined as the ratio of differential gain Ad to common-mode gain Ac, usually expressed in decibels dB. The calculation formula is: , where A d is the differential gain, i.e. the amplification factor of the amplifier to the differential signal ; Ac is the common-mode gain, i.e. the amplification factor of the amplifier to the common-mode signal ).

[0044] Special attention should be paid to the bridge balance condition during debugging: when R1 / R2=Z1 / Z2, the bridge output is zero. Due to the presence of capacitive impedance components in the probe equivalent impedance Z1 and the reference impedance Z2, the actual adjustment needs to take into account the amplitude-frequency characteristics. Verify the consistency of the Z1 / Z2 amplitude angle through a vector network analyzer, ensure that the phase difference Δθ<0.5°, and avoid false balance points caused by capacitive impedance mismatch.

[0045] After debugging is completed, the Vout stability needs to be observed in the thermostat for 24 hours. If the long-term drift is >5μV, then the temperature coefficient of R1 / R2 needs to be recalibrated and replaced with a metal film resistor with a temperature coefficient ≤5ppm / ℃. After the final confirmation of the balance impedance R3 resistance value, the digital potentiometer adjustment knob is encapsulated with epoxy resin to prevent resistance value drift caused by mechanical vibration. This process ensures that the initial unbalance voltage of the bridge is ≤10μV, providing a stable reference for subsequent dynamic calibration.

[0046] In terms of dynamic balance calibration algorithm deployment: The core of the dynamic balance calibration algorithm is to achieve long-term stability of the bridge system through periodic feedback adjustment. The feedback control algorithm embedded in the microcontroller (STM32F407) triggers the calibration process every 5 minutes: at the beginning of each calibration, the program acquires the bridge output voltage Vout and compares it with the threshold value 10μV. If |Vout|>10μV, then the PID control algorithm is started to calculate the adjustment amount , where Vref is the reference voltage (5V). The design of this algorithm is based on the bridge balance equation R1 / R2=Z1 / Z2, and through dynamic adjustment of the balance impedance R3 resistance value in the reference impedance Z2, it compensates for the imbalance caused by temperature drift or electromagnetic interference.

[0047] (1) Algorithm implementation details First, PID parameter tuning: the proportional gain Kp is set to 0.05%, the integral time Ti=10s, and the derivative time Td=0.5s. This parameter configuration can ensure that the system overshoot is <2% and the regulation time is <30s under step response.

[0048] Second, adjust the step control: the amount of each adjustment step is set to 0.05% (corresponding to the minimum step resolution of AD5207 digital potentiometer), and the maximum adjustment amount is not more than 1% of the current value of R3. This limit is based on the non-linear analysis of the bridge balance characteristics, which may cause the system to enter the non-linear region when ΔR / R3>1%, triggering oscillation.

[0049] Third, temperature compensation mechanism: the algorithm has a built-in temperature sensor (DS18B20) reading compensation module, which automatically triggers the pre-calibration program when the ambient temperature changes ΔT≥0.5℃, and adjusts the balance impedance R3 value in advance to offset the thermal expansion effect.

[0050] (2) Experimental verification and performance comparison After continuous operation in a thermostat (25℃±0.1℃) for 72 hours, the system zero drift is 0.0078%FS / ℃, which is two orders of magnitude higher than the traditional static calibration (single adjustment without updating) of 0.8%FS / ℃. Frequency domain analysis shows that in the 1Hz-1kHz interference frequency range, the calibration algorithm keeps the common mode rejection ratio (CMRR) above 95dB.

[0051] As for the optimization of differential amplifier parameters: The gain configuration of differential amplifier AD8421 needs to consider both signal amplification capability and stability. According to the data manual, the gain of AD8421 is set by external resistance (G=1 to 10000), and this scheme uses a gain mode of G=100V / V. In specific implementation, the ratio of feedback resistance Rf to gain setting resistance Rg is set to 99:1 (Rf=99kΩ±0.1%, Rg=1kΩ±0.1%), and the matching error is realized to be less than 0.02% through laser trimming film resistance. Under this configuration, the -3dB bandwidth of the amplifier needs to cover 10kHz-1MHz to meet the high-frequency dynamic response requirements of eddy current sensors.

[0052] (1) Bandwidth adjustment and signal injection A 100kHz sine wave is injected as a test signal using a Keysight 33500B function generator, and the specific steps are as follows: first, set the function generator output amplitude to 50mVpp and frequency to 100kHz, ensuring that the signal source resistance matches the amplifier input impedance (50Ω); second, monitor the amplifier output end through an oscilloscope (Tektronix MDO3024), and adjust the Rf resistance until the output waveform has no obvious clipping (Vout=5Vpp); third, sweep the bandwidth: gradually increase the signal source frequency from 10kHz to 1MHz, and record the critical point where the output amplitude decays to -3dB.

[0053] (2) Noise and distortion measurement Noise density and total harmonic distortion (THD) are key indicators for measuring the performance of an amplifier. The following tests were conducted using a Rohde & Schwarz FPC1500 spectrum analyzer: First, noise density measurement: disconnect the signal source input and connect a 50Ω termination resistor to simulate a pure noise environment. Set the spectrum analyzer resolution bandwidth (RBW) to 1 kHz, the scan range to 10 kHz-1 MHz, and record the noise power spectral density at each frequency band. Second, THD test: reconnect the 100 kHz sine wave and adjust the signal amplitude to make the output reach full scale (5Vpp). Enable the spectrum analyzer harmonic analysis function, extract the 2nd to 10th harmonic components, and calculate The actual THD is 0.0087%, which is better than the design target of 0.01%.

[0054] (3) SNR calculation and verification SNR is calculated by the formula , where Vrms is the effective signal root mean square value, and Nrms is the noise root mean square value. The specific implementation steps are as follows: First, measure the effective value of the signal: under the condition of 100 kHz input, use a true RMS multimeter (Fluke 45) to measure the output Vrms=3.535V; Second, measure the effective value of the noise: after disconnecting the input signal, repeat the measurement of the noise Nrms=50μV; Third, calculate , this result is cross-verified by the integral function of the spectrum analyzer, with an error <0.5dB.

[0055] Experimental data show that this parameter optimization scheme makes the differential amplifier have high gain, ultra-low distortion, and high SNR characteristics, which can provide high-quality signal sources for subsequent ADC digital conversion. The spectrum analyzer measurement results are highly consistent with the theoretical calculations, verifying the effectiveness of the optimization process.

[0056] In terms of LC resonant network debugging The core goal of resonant network debugging is to maximize signal transmission efficiency by precisely controlling the resonant characteristics of the probe coil and reference impedance. The specific implementation process is as follows: (1) Self-resonant frequency calibration Use a Keysight E5061B vector network analyzer (equipped with a gain phase test port) to perform frequency sweep testing on the probe coil to determine its self-resonant frequency f0. The testing method is as follows: First, connect the probe coil to the test fixture and set the frequency scan range to 10 kHz-1 MHz with a resolution bandwidth (RBW) of 1 kHz; Second, observe the impedance trajectory through the Smith circle, and when the phase angle θ changes from +90° to -90°, the corresponding frequency is f0. Experimental data show that the self-resonant frequency of the probe coil is stable at 105 kHz±0.5%.

[0057] (2) LC resonance parameter design LC parallel network of reference impedance Z2 needs to meet the resonant frequency , the error is controlled within ±1%. The specific parameter selection: fixed inductance L=10μH; initial capacitance C=100nF.

[0058] Theoretical calculation: , the deviation from the target value is +0.65%, and the C value needs to be fine-tuned. Through the time domain / fault location analysis option of E5061B, the mismatch point caused by the parasitic parameters of the capacitor is located, and finally the C is adjusted to 99.3nF to achieve f=100kHz±0.5%.

[0059] (3) Phase angle compensation and impedance matching At f0=105kHz, it is necessary to ensure that Z1(probe coil) and Z2(reference impedance) exhibit pure resistance(θ=0°). The debugging steps are as follows: first, use the Smith chart function of E5061B to monitor the impedance trajectory of Z1 / Z2. Second, if the phase angle θ≠0°, it indicates that there is a capacitive / reactive mismatch. Phase compensation is performed through the parallel adjustable capacitor C3(100pF±0.5pF): when θ>0°, increase the C3 capacitance to offset the inductance; when θ<0°, reduce the C3 capacitance to balance the capacitance. Third, finally make the phase angle of Z1 / Z2 at f0 stable at 0°±0.3°, corresponding to the reflection coefficient |Γ|<0.05, the signal transmission efficiency .

[0060] (4) High frequency band performance verification After matching, the amplitude-frequency characteristic of 100kHz-1MHz frequency band needs to be verified: ① amplitude attenuation test: inject a sweep signal(100kHz-1MHz), record the amplitude difference between input / output signals. The measured data shows that the attenuation is 0dB at 100kHz, and ≤0.4dB at 1MHz, which is better than the design index; ② phase distortion analysis: through the phase measurement function of E5061B, record the phase shift in the whole frequency band. The experimental results show that the phase distortion is ≤1.8°, which meets the requirements of signal fidelity for high-speed vibration detection.

[0061] Experimental verification shows that the resonant matching network improves the high frequency response capability of the eddy current sensor by 3 times(100kHz→300kHz), and the THD is reduced to 0.0087%. Through the time domain option of E5061B, the mismatch point of PCB wiring can also be located, and the characteristic impedance of the transmission line is further optimized to achieve return loss>25dB.

[0062] In terms of system-level performance verification To verify the performance of the circuit in actual detection scenarios, a test platform based on an axial vibration table is built. The platform uses electromagnetic excitation principle, and through a power amplifier to drive the vibration table to realize controllable vibration output with amplitude range of 50-500 pm and frequency of 10-1 kHz. During the test, Keyence LKG5000 laser displacement sensor is used as a reference standard, with resolution of 0.1 pm and sampling frequency of 100 kHz, meeting the synchronous acquisition requirement.

[0063] (1) Nonlinear error and repeatability error test The calculation of nonlinear error ε is based on least square fitting curve, and the formula is where Δy is the maximum deviation of the measured value from the fitted straight line, and yFS is the full-scale output (5 V corresponds to 500 pm). The specific implementation steps are as follows: first, set the vibration table to perform stepwise amplitude scanning (50 pm to 500 pm) at a frequency of 10 Hz; second, record the output voltage Vout of the circuit and the displacement value of the laser sensor synchronously, and establish the input-output relationship curve; third, perform linear fitting on the data points by using MATLAB toolbox, extract Δy=0.98 pm, and calculate ε=0.196%FS.

[0064] The repeatability error test adopts multiple cycle loading-unloading experiments: under the same amplitude (such as 300 pm), 50 consecutive tests are performed, the standard deviation σ of the output value is calculated as 0.022%FS, and the repeatability error is calculated as 0.066%FS according to the 3σ principle. The experimental data show that the error distribution conforms to the normal distribution (K-S test p=0.21>0.05), indicating that the system has high consistency.

[0065] (2) Dynamic response time test For the monitoring needs of high-speed rotating equipment, a transient vibration detection experiment is designed: first, a function generator (Keysight 33500B) is used to trigger the vibration table to generate a step signal (rise time <1 ns); second, an oscilloscope (Tektronix MDO3024) is used to capture the laser sensor signal and the output signal of the circuit simultaneously; third, the time difference Δt between the two is measured as 10 ns, which is significantly better than the 1 ps level of the traditional scheme. This performance improvement is due to the synergistic optimization of the four-arm bridge and the LC resonance matching network, which makes the signal transmission efficiency ≥95% and the phase distortion ≤1.8°.

[0066] The experimental data show that the phase distortion of the circuit in the low frequency band (<100 Hz) is ≤0.5°, and the amplitude attenuation in the high frequency band (>10 kHz) is <0.4 dB. This result verifies the feasibility of the four-arm bridge topology structure and dynamic balance calibration technology in engineering applications.

[0067] For environmental adaptability test The environmental adaptability test is divided into temperature cycle stability verification and electromagnetic compatibility (EMC) evaluation, aiming to ensure the reliability of the circuit of the application in long-term operation in complex industrial environments.

[0068] (1) Temperature cycle stability test Design temperature cycle test scheme: first, test parameter setting: place the circuit board in a temperature cycle box (ESPEC PL-3KPH), set the temperature change rate to 15℃ / min, and the cycle range to -20℃ (low temperature for 2 hours) → 85℃ (high temperature for 2 hours) → -20℃ (a total of 36 cycle periods). Second, data acquisition and analysis: record the bridge output zero drift every hour, and use a high-precision data acquisition system (Keysight 34972A) to monitor the environmental temperature and Vout voltage simultaneously. The experimental data show that the total drift is 0.18%FS after 72 hours of continuous operation, which is significantly better than the 0.8%FS level of the traditional scheme.

[0069] (2) Electromagnetic compatibility (EMC) evaluation According to the IEC 61000-6-2 standard, build an anti-interference test environment: first, electrostatic discharge (ESD) test: use Teseq IEC 61000-4-2 generator to apply ±8kV pulse in contact discharge mode (4th level anti-interference requirement). During the test, through the common mode rejection ratio (CMRR≥90dB) of the differential amplifier AD8421 and the PCB shielding design (no digital signal wiring within 20mm), the output signal fluctuation is <2μV, meeting the stringent requirements of ISO 10605 on industrial equipment. Second, conducted / radiated immunity test: inject interference signals in the frequency range of 30MHz-1GHz, with a power density of 3V / m, and use Rohde&Schwarz TSMA6B sweep instrument to monitor the abnormal function of the circuit. The results show that the differential output mode attenuates the common mode interference to below -65dBc, which is 20dB higher than the single-ended output scheme.

[0070] The experimental data show that the circuit can still maintain a detection accuracy of 0.2μm under -20℃ cold start conditions, and has no false touch phenomenon under IEC61000-4-2 level 4 ESD impact, fully meeting the reliability requirements of GB / T 17626.2-2018 on industrial measurement equipment.

[0071] On the basis of the above detailed and specific implementation, the following briefly describes related specific embodiments in the research and development process of the application.

[0072] Example 1: Four-arm bridge topology construction Hardware configuration: R1, R2 adopt metal film resistance (ΔR / R <0.05%, temperature drift ≤5ppm / ℃), four-wire calibration method to eliminate contact resistance; Probe coil: 0.2mm single crystal copper wire is wound 500±1 turns, inductance 10μH±2%; Technical effect: initial unbalance voltage ≤10μV, output sensitivity ≥2V / μm.

[0073] Example 2: Dynamic balance calibration Calibration process ( Figure 2 ): ① Self-checking: AD5207 adjusts R3 to make V_out ≤10μV; ② Periodic calibration: PID adjusts ΔR=R3×(V_out / V_ref) every 5 minutes; Temperature compensation: DS18B20 sensor triggers pre-calibration when ΔT ≥0.5℃; Technical effect: 72-hour drift 0.0078%FS / ℃.

[0074] Example 3: Differential amplifier optimization Parameter adjustment: Inject 100kHz sine wave (50mV_pp), verify -3dB bandwidth ≥1MHz by sweeping frequency; Noise density test: disconnect the signal source, RBW=1kHz, N_rms=50μV is measured; Technical effect: SNR=63dB.

[0075] Example 4: LC resonance matching debugging Network analyzer calibration (Keysight E5061B): ① Smith circle positioning self-resonant frequency f0=105kHz±0.5%; ② Adjust C3 to make Z1 / Z2 phase angle θ=0°±0.3°; Technical effect: amplitude attenuation ≤0.4dB at 1MHz, THD ≤0.0087%.

[0076] Example 5: System level verification Axial vibration table test: Nonlinear error: 0.196%FS (least squares fitting); Dynamic response: step signal response time 10ns (traditional scheme needs 1μs); Environmental adaptability: Temperature cycle (-20℃ ↔ 85℃): 72-hour drift 0.18%FS; ESD immunity (±8kV): output fluctuation <2μV.

[0077] The innovation of the technical solution is reflected in the synergistic optimization of three technical elements: (1) The full-symmetry structure of the four-arm bridge forms a sensitive response to small ΔZ through the matching design of fixed resistors and impedance arms; (2) The high CMRR characteristics of the differential amplifier and the frequency selectivity of the LC resonance matching network jointly build a high signal-to-noise ratio signal channel; (3) The dynamic balance calibration technology realizes the long-term stability of the bridge system through real-time feedback control. Experimental data show that the circuit has a non-linear error of ≤0.2% FS in the 0-2mm measurement range, a repeatability error of ≤0.05% FS, and can stably detect displacement changes of 0.1μm, meeting the precision requirements of ISO 7919-1 standard for rotating machinery vibration monitoring.

[0078] The technology has been verified by axial vibration table (10Hz-1kHz frequency range), and the output signal has a phase distortion of ≤2° in the low frequency band (<100Hz) and an amplitude attenuation of <0.5dB in the high frequency band (>10kHz), showing excellent dynamic response characteristics. Especially in high-speed rotating equipment applications, the circuit can capture transient vibration signals above 100kHz, with a response speed of 10ns, which is significantly better than the 1μs level of traditional solutions. This performance improvement is due to the broadband matching design of the four-arm bridge and the differential amplifier, as well as the effective suppression of signal baseline fluctuations by the dynamic calibration algorithm. In summary, the technical solution has made breakthrough progress in sensitivity, stability and dynamic response by integrating topology innovation, signal processing optimization and dynamic calibration technology, providing a new solution for precision displacement measurement and equipment state monitoring.

[0079] It can be seen that the eddy current sensor high sensitivity detection circuit based on the four-arm bridge has significant technical advantages in actual application. Through the full-symmetry topology design of the four-arm bridge (two fixed resistors R1=R2=1kΩ and the probe Z1, the reference Z2 impedance arm constitute a balanced bridge circuit), combined with the high common mode rejection characteristic (CMRR≥90dB) of the differential amplifier (AD8421), the breakthrough improvement of the output sensitivity ≥2V / μm is realized. Experimental data shows that the scheme improves the sensitivity of the traditional half-bridge circuit by 52%, and can stably detect 0.1μm level displacement change, solves the technical problem that the micro vibration signal is easy to be submerged by noise. In terms of anti-interference ability, the signal-to-noise ratio (SNR) is improved to 63dB in the differential output mode, which is significantly better than the 40dB level of the traditional single-ended output. Through the integration of the LC resonant matching network (resonant frequency error ≤±1%), the signal transmission efficiency is optimized in the 10kHz-1MHz wide frequency band, the high frequency band amplitude attenuation is <0.5dB, and the phase distortion is ≤2°, which meets the dynamic detection requirements of high-speed rotating equipment. The dynamic balance calibration technology cooperates with the adjustable impedance module (R3C3 parallel) and the feedback control algorithm, and the zero drift is suppressed to below 0.008%FS / ℃.

[0080] Experimental verification shows that the non-linear error of the present application is ≤0.2%FS in the 0-2mm measurement range, and the repeatability error is ≤0.05%FS, which meets the precision requirements of ISO 7919-1 standard for rotating machinery vibration monitoring. In the axial vibration table test, its 10ns level response speed is significantly better than the 1μs level of the traditional scheme, the phase distortion is ≤2° in the low frequency band (<100Hz), the amplitude attenuation is <0.5dB in the high frequency band (>10kHz), and excellent dynamic response characteristics are exhibited. The present application is suitable for precision manufacturing, high-end equipment state monitoring and other industrial scenes. Compared with the prior art, the present application focuses on the cooperative innovation of four-arm bridge differential output and impedance matching optimization, and solves the problem of zero drift in high-speed dynamic detection.

[0081] It can be understood that the above embodiments are only exemplary embodiments adopted for illustrating the principles of the present application, and the present application is not limited thereto. Various modifications and improvements can be made by those skilled in the art without departing from the spirit and essence of the present application, and these modifications and improvements are also regarded as the protection scope of the present application.

Claims

1. A sensitivity detection circuit for an eddy current sensor based on a four-arm bridge, characterized in that, include: The four-arm bridge topology includes a symmetrical resistance arm formed by a first fixed resistor (R1) and a second fixed resistor (R2), and a variable impedance arm formed by a probe equivalent impedance (Z1) and a reference impedance (Z2); the probe equivalent impedance (Z1) and the reference impedance (Z2) are connected in parallel to an LC resonant network. A differential amplifier module, electrically connected to the four-arm bridge topology, is used to amplify the differential voltage signal output by the bridge; and A dynamic balance calibration unit is electrically connected between the four-arm bridge topology and the differential amplifier module, and includes an adjustable impedance module and a feedback control module to achieve power-on self-test calibration and periodic dynamic calibration.

2. The eddy current sensor sensitivity detection circuit according to claim 1, characterized in that, The resonant frequency of the LC resonant network is consistent with the operating frequency of the probe, and the error is ≤ ±1%.

3. The eddy current sensor sensitivity detection circuit according to claim 1, characterized in that, The differential amplifier module has a common-mode rejection ratio of ≥90dB and a gain-bandwidth product of ≥10MHz.

4. The eddy current sensor sensitivity detection circuit according to claim 3, characterized in that, The differential amplifier module uses an AD8421 amplifier with an input impedance ≥1GΩ and an output signal-to-noise ratio ≥60dB.

5. The eddy current sensor sensitivity detection circuit according to claim 1, characterized in that, The adjustable impedance module is a parallel structure of a balanced impedance (R3) and a balanced capacitor (C3), wherein the dynamic balance calibration unit adjusts the resistance value by using a digital potentiometer, which serves as the balanced impedance (R3), and performs PID feedback regulation.

6. The eddy current sensor sensitivity detection circuit according to claim 1, characterized in that, In the LC resonant network where the equivalent impedance (Z1) of the probe is connected in parallel with the reference impedance (Z2), the inductance and capacitance parameters of the LC resonant network satisfy the following relationship: ； Where f is the probe's operating frequency, L is the inductance of the LC resonant network, and C is the capacitance of the LC resonant network.

7. The eddy current sensor sensitivity detection circuit according to claim 1, characterized in that, The four-arm bridge topology is powered by a constant voltage source. The first fixed resistor (R1) and the second fixed resistor (R2) have the same resistance value and the error is ≤0.1Ω. The output sensitivity is ≥2V / μm.

8. The eddy current sensor sensitivity detection circuit according to claim 1, characterized in that, The capacitor (C3) in the adjustable impedance module is a ceramic capacitor with a value of 100pF ± 0.5pF and an equivalent series resistance (ESR) ≤ 0.1Ω.

9. The eddy current sensor sensitivity detection circuit according to claim 1, characterized in that, The feedback control module has a temperature compensation unit to trigger a pre-calibration program to adjust the resistance value of the adjustable impedance module when the ambient temperature changes by ΔT ≥ 0.5℃.

10. The eddy current sensor sensitivity detection circuit according to any one of claims 1 to 9, characterized in that, The circuit wiring of the eddy current sensor sensitivity detection circuit satisfies the following conditions: the symmetry error of the trace length of the four-arm bridge topology is ≤0.1mm, and there are no digital signal traces within 20mm around the differential amplifier module.

Images