A hall displacement detection circuit and method with full stroke high signal-to-noise ratio realized by dual-ADC cooperative sampling

Abstract

This invention discloses a Hall displacement detection circuit and method with high signal-to-noise ratio across the entire stroke using dual ADC collaborative sampling. It belongs to the field of non-contact Hall precision displacement sensing and high-precision analog signal acquisition technology. The hardware relies on two symmetrical low-temperature drift following circuits, a gain conditioning circuit, a synchronous dual ADC acquisition unit with adaptive signal-to-noise ratio data fusion, and a complete adaptive weighted fusion mathematical algorithm. This achieves uniform high signal-to-noise ratio detection of Hall displacement across the entire stroke, balancing micro-displacement resolution with complete large-stroke measurement. It also reduces channel temperature drift mismatch and simplifies the hardware structure. This overcomes the shortcomings of existing technologies, such as poor signal-to-noise ratio at long distances, easy saturation at close range, temperature drift mismatch in dual-channel circuits, output jumps during range switching, and complex hardware in high-precision solutions.

Description

Technical Field

[0001] This invention belongs to the field of non-contact Hall precision displacement sensing and high-precision analog signal acquisition technology, specifically involving a Hall displacement detection circuit and method that achieves high signal-to-noise ratio throughout the entire stroke through dual ADC collaborative sampling. Background Technology

[0002] Hall effect displacement sensors achieve non-contact position measurement by relying on the correspondence between the gradient magnetic field of a permanent magnet and the air gap distance. They offer advantages such as vibration resistance, long lifespan, simple structure, and fast response speed, and are widely used in industrial closed-loop control equipment. Traditional Hall effect displacement detection circuits generally employ a single-channel buffer + single-channel ADC acquisition architecture, which has inherent and intractable defects.

[0003] (1) Fixed gain cannot adapt to the magnetic field changes throughout the entire stroke: When the permanent magnet is far away from the Hall chip, the induced magnetic field is weak, the effective signal amplitude is low, the circuit thermal noise and ADC quantization noise account for a very high proportion, the signal-to-noise ratio is severely degraded, and the micron-level small displacement is easily submerged by noise, resulting in insufficient resolution; if the amplification gain is simply increased, the analog output will be directly saturated and clipped when the permanent magnet approaches and the magnetic field becomes stronger, and the data in the large displacement range will be completely invalid.

[0004] (2) Single-channel acquisition has the problem of uneven signal-to-noise ratio throughout the entire process. The measurement accuracy is high at the near end of the displacement and poor at the far end. The consistency of detection throughout the entire stroke is poor, which cannot meet the hard requirements of precision closed-loop control for low jitter and uniform measurement accuracy throughout the entire process.

[0005] (3) Conventional dual-channel schemes lack symmetrical low temperature drift buffer pre-amplifier circuits, the input impedance of the two channels is mismatched and the temperature drift characteristics are inconsistent, introducing additional channel mismatch error; and most of them use fixed threshold to switch ranges, which are prone to output jumps and signal-to-noise ratio changes in the transition range, and do not achieve full-stroke signal-to-noise ratio balance.

[0006] (4) Existing high-precision optimization methods mostly rely on chopper modulation circuits and DAC dynamic compensation modules, which involve many peripheral devices, complex PCB wiring, and high hardware costs, which are not conducive to sensor miniaturization and mass production.

[0007] In summary, existing Hall displacement acquisition schemes cannot simultaneously meet the multiple requirements of uniform high signal-to-noise ratio throughout the entire stroke, high resolution for micro-displacement, unsaturated large stroke, and simple and low-cost circuitry, and have obvious technical shortcomings. Summary of the Invention

[0008] The technical problem solved by this invention is to address the shortcomings of existing single-ADC fixed-gain Hall displacement circuits, such as uneven signal-to-noise ratio (SNR) across the entire stroke, poor SNR at far-end micro-displacement, easy saturation at close range, temperature drift mismatch in dual-channel circuits, output jumps during range switching, and complex hardware in high-precision solutions. This invention provides a Hall displacement detection circuit and method that achieves high SNR across the entire stroke through dual-ADC collaborative sampling. Relying on two symmetrical low-temperature drift following buffers, differentiated gain conditioning, synchronous dual-ADC acquisition from the same source, and an adaptive weighted fusion algorithm for SNR, this invention achieves uniform high SNR detection of Hall displacement across the entire stroke without the need for chopper circuits or DAC compensation circuits. It balances micro-displacement resolution with complete large-stroke measurement, while reducing channel temperature drift mismatch and simplifying the hardware structure.

[0009] Technical Solution: To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0010] A Hall displacement detection method with high signal-to-noise ratio across the entire stroke is proposed by dual ADC collaborative sampling. The Hall output voltage of the Hall sensor is processed to obtain the standard Hall sensing voltage estimate. The signal-to-noise ratio of the high-gain channel, the signal-to-noise ratio of the low-gain channel, and the equivalent signal-to-noise ratio after synthesis are calculated. The final detected displacement is obtained based on an adaptive weighted fusion complete mathematical algorithm.

[0011] Furthermore, the calculation method for the standard Hall induced voltage estimate is as follows: the Hall output voltage enters two low-temperature drift follower circuits for impedance transformation and buffer isolation, outputting two stable Hall output voltages. The two stable Hall output voltages are then subjected to gain conditioning by high-gain conditioning circuits and low-gain conditioning circuits, and output to the signal-to-noise ratio adaptive data fusion unit for quantization and correction to obtain the standard Hall induced voltage estimate.

[0012] Furthermore, quantization and correction are performed using an ADC, and the estimated standard Hall effect voltage is obtained as follows:

[0013] ;

[0014] in, This is the sampled data from the high-gain channel after analog-to-digital conversion. This is the sampled data from the low-gain channel after analog-to-digital conversion. These are the gain coefficients for the high-gain and low-gain channels. This refers to the gain zeros of the high-gain and low-gain channels. Estimation of the standard Hall effect voltage for the high-gain channel. Estimated standard Hall effect voltage for low-gain channel.

[0015] Furthermore, the signal-to-noise ratios of the high-gain channel and the low-gain channel are as follows:

[0016] ;

[0017] in, For high-gain channel signal-to-noise ratio. For low-gain channel signal-to-noise ratio. This represents the total noise power of the high-gain channel. This represents the total noise power of the low-gain channel. This is the Hall output voltage.

[0018] Furthermore, the combined equivalent signal-to-noise ratio is obtained based on the set fusion weights and the Hall output voltage; the fusion weights are set as follows: , ,satisfy The combined equivalent signal-to-noise ratio is:

[0019] ;

[0020] in, This represents the total noise power of the high-gain channel. This represents the total noise power of the low-gain channel. This is the Hall output voltage.

[0021] Furthermore, the adaptive weighted fusion complete mathematical algorithm includes displacement interval determination threshold calibration, calculation of standard Hall sensing voltage based on interval weight mathematical expression, and final detection displacement based on standard Hall sensing voltage.

[0022] Furthermore, the displacement interval determination threshold calibration includes calculating the basic displacement estimate through a low-gain channel. Through basic displacement estimation Set two fixed remote thresholds , Two remote intervals and one transition interval are determined based on two fixed remote thresholds.

[0023] Furthermore, the interval weight expression includes:

[0024] distant interval High-gain channels are used.

[0025] distant interval Use a low-gain channel;

[0026] transition interval Based on basic displacement estimation The fusion weight is determined by a fixed remote threshold, and the standard Hall effect voltage is obtained based on the fusion weight and the standard induced voltage estimate.

[0027] Furthermore, the final expression for the detected displacement is:

[0028] ;

[0029] in, This is the standard Hall effect sensing voltage. , To fix the remote threshold.

[0030] A Hall displacement detection circuit achieving high signal-to-noise ratio across the entire displacement range using dual ADC collaborative sampling includes a Hall sensor, a low-temperature drift follower circuit, a gain conditioning circuit, and a signal-to-noise ratio adaptive data fusion unit. The Hall sensor outputs a Hall output voltage, and the gain conditioning circuit implements impedance matching buffering and eliminates temperature drift offset. The gain conditioning circuit includes a high-gain conditioning circuit and a low-gain conditioning circuit. The signal-to-noise ratio adaptive data fusion unit uses dual ADCs synchronized by the same source clock to acquire the preceding signal in parallel, and combines it with an adaptive weighted fusion complete mathematical algorithm to achieve uniform high signal-to-noise ratio across the entire Hall displacement range, thus obtaining the final detected displacement.

[0031] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0032] This invention provides a Hall displacement detection circuit and method that achieves high signal-to-noise ratio (SNR) across the entire displacement range through dual ADC collaborative sampling. It leverages a high-gain channel to amplify the amplitude of weak remote sensing signals and employs an adaptive weighting transition mechanism for SNR to significantly reduce the noise floor at the remote end without requiring changes to the hardware gain level. This results in a uniform and consistent high SNR output across the entire displacement range, completely resolving the inherent defects of traditional single fixed-gain acquisition schemes, such as insufficient resolution of remote micro-displacements, unbalanced SNR distribution across the entire range, and large output jitter. Micrometer-level minute displacement changes can be stably and accurately identified, and the output signal exhibits no abrupt jumps. This meets the stringent requirements of industrial precision valve cores and servo positioning closed-loop control for low-jitter, uniform, and high-precision detection signals. Attached Figure Description

[0033] Figure 1 This is a block diagram of the technical solution;

[0034] Figure 2 A schematic diagram of mathematical modeling for the hardware architecture;

[0035] Figure 3 A schematic diagram illustrating the complete mathematical derivation of signal-to-noise ratio (SNR);

[0036] Figure 4 A schematic diagram of the complete mathematical algorithm for adaptive weighted fusion;

[0037] Figure 5 This is a schematic diagram of a high-gain conditioning circuit.

[0038] Figure 6 This is a schematic diagram of a low-gain conditioning circuit.

[0039] Figure 7 This is a schematic diagram of the signal-to-noise ratio adaptive data fusion unit circuit.

[0040] Figure 8 The waveform diagram of the 1μm periodic transformation of the far-end interval;

[0041] Figure 9 The waveform diagram is shown as a 1μm waveform of the near-end interval periodic transformation.

[0042] Figure 10 The waveform of the far-end interval periodic transformation after adaptive signal-to-noise ratio is shown in 1μm.

[0043] Figure 11 The waveform of the near-end interval periodic transformation after adaptive signal-to-noise ratio is shown in 1μm. Detailed Implementation

[0044] The present invention will be further illustrated below with reference to specific embodiments. These embodiments are implemented based on the technical solutions of the present invention, and it should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention.

[0045] Example 1

[0046] like Figure 1 As shown, this embodiment provides a Hall displacement detection circuit that achieves high signal-to-noise ratio across the entire travel distance using dual ADC collaborative sampling. The circuit includes a Hall sensor, a low-temperature drift follower circuit, a gain conditioning circuit, and a signal-to-noise ratio adaptive data fusion unit. The Hall sensor outputs a Hall output voltage. The gain conditioning circuit implements impedance matching buffering and eliminates temperature drift offset. The gain conditioning circuit includes a high-gain conditioning circuit and a low-gain conditioning circuit. The signal-to-noise ratio adaptive data fusion unit uses dual ADCs synchronized by the same source clock to acquire the preceding signal in parallel. Combined with an adaptive weighted fusion complete mathematical algorithm, it achieves uniform high signal-to-noise ratio across the entire Hall displacement travel distance, obtaining the final detected displacement.

[0047] Hall effect sensor: output and displacement Corresponding induced voltage The larger the air gap (the farther the displacement), the better. The smaller the amplitude, the smaller the air gap (near the displacement). The larger the amplitude.

[0048] Upper branch: Low temperature drift follower circuit → High gain conditioning circuit (gain) → .

[0049] Lower branch: Low temperature drift follower circuit → Low gain conditioning circuit (gain) , → .

[0050] Two completely symmetrical preamps: the low temperature drift follower circuit of the upper branch and the low temperature drift follower circuit of the lower branch have perfectly matched op-amp and RC parameters, with only the back-end gain being different.

[0051] Synchronous Dual ADC Sampling + Signal-to-Noise Ratio Adaptive Data Fusion Unit: Simultaneously acquires two ADC code values ​​using the same clock source, performs normalization correction, signal-to-noise ratio calculation, and segmented weighted fusion, outputting the final high-precision displacement. .

[0052] like Figure 5 As shown, the high-gain conditioning circuit in this embodiment consists of two cascaded zero-drift, low-temperature-drift operational amplifiers, OPA186DBVR. The first-stage operational amplifier U19 forms a low-temperature-drift voltage follower buffer circuit. The Hall input signal is fed to pin 3 of the non-inverting input of U19 through a 1kΩ resistor R87 with 1% accuracy. A 5.6μF capacitor C143 with a withstand voltage of 50V is connected in parallel with R87 to analog ground AGND, forming an input RC low-pass filter network to filter out high-frequency electromagnetic interference. A 1kΩ, 1% accuracy feedback resistor R85 is connected between pin 4 of the inverting input and pin 1 of the output of U19 to form a unity-gain voltage follower. As per the structure, a 10μF / 16V capacitor C141 and a 0.1μF / 50V capacitor C142 are connected in parallel to AGND at pin 5 of the positive power supply of U19, and a 0.1μF / 50V capacitor C145 and a 10μF / 16V capacitor C146 are connected in parallel to AGND at pin 2 of the negative power supply. The combination of dual-specification capacitors achieves power supply ripple suppression and low temperature drift voltage regulation buffer. This stage circuit relies on the zero drift characteristic of OPA186 to complete the electrical isolation between the high output impedance of the Hall chip and the low input impedance of the subsequent amplifier circuit, eliminating the temperature drift offset and channel mismatch caused by impedance coupling between the front and rear stages. It only performs impedance matching buffering for the Hall sensing signal and does not change the signal amplitude.

[0053] The Hall signal from the buffered output of U19 is fed into pin 3 of the non-inverting input of the second-stage operational amplifier U27 through a 1kΩ, 1% precision resistor R97. Pin 4 of the inverting input of U27 is connected to AGND through a 1kΩ, 1% precision resistor R99. A 9kΩ, 1% precision feedback resistor R86 is connected between pin 4 of the inverting input and pin 1 of the output, forming a non-inverting proportional amplifier circuit with a closed-loop voltage amplification factor of 10. A 10μF / 16V capacitor C170 and a 0.1μF / 50V capacitor C171 are connected in parallel to AGND at pin 5 of the positive power supply of U27. A 0.1μF / 50V capacitor C179 and a 10μF / 16V capacitor C171 are connected in parallel to AGND at pin 2 of the negative power supply. Capacitor C180 to AGND also adopts a dual-capacitor power supply decoupling design to ensure the low temperature drift and low noise operation performance of the amplifier circuit. The amplified analog signal HALL_GAIN10 is output from pin 1 of U27 and connected to the ADC_IN0 port of the signal-to-noise ratio adaptive data fusion unit. This high-gain conditioning branch is specifically designed to boost the amplitude of the small displacement sensing signal corresponding to the weak gradient magnetic field of the permanent magnet at a long distance, reduce the background noise ratio of the circuit, and significantly improve the signal-to-noise ratio and micron-level displacement resolution of the far-end micro-displacement detection. All resistors and capacitors are selected with 1% precision matching devices to ensure the consistency of hardware parameters of the two symmetrical conditioning channels and reduce the inherent measurement deviation of dual-channel acquisition.

[0054] like Figure 6 As shown, the low-gain conditioning circuit in this embodiment has a completely symmetrical hardware architecture to the high-gain conditioning circuit described above. It is also composed of two cascaded zero-drift, low-temperature-drift operational amplifiers, OPA186DBVR. The first-stage operational amplifier U25 is a low-drift voltage follower buffer circuit. The Hall input signal is fed to pin 3 of the non-inverting input of U25 via a 1kΩ resistor R98 with 1% accuracy. A 5.6μF capacitor C144 with a withstand voltage of 50V is connected in parallel with R98 to analog ground AGND, forming an input RC low-pass filter network to filter out high-frequency electromagnetic interference in the industrial environment. A 1kΩ resistor with 1% accuracy is connected between pin 4 of the inverting input and pin 1 of the output of U25. A feed resistor R96 forms a unity-gain voltage follower structure. A 10μF / 16V capacitor C168 and a 0.1μF / 50V ceramic capacitor C169 are connected in parallel to AGND at pin 5 of the positive power supply of U25. A 0.1μF / 50V capacitor C172 and a 10μF / 16V capacitor C173 are connected in parallel to AGND at pin 2 of the negative power supply. A two-stage power supply decoupling scheme with large-capacity electrolytic capacitors and high-frequency ceramic capacitors is used to suppress power supply ripple and reduce baseline shift caused by temperature drift. This follower stage achieves impedance isolation between the high internal resistance output of the Hall effect and the subsequent amplifier circuit, eliminates the impedance mismatch error of the dual channels, and only buffers the Hall effect sensing signal without changing the signal amplitude.

[0055] The Hall signal, buffered by U25, is fed into pin 3 of the non-inverting input of the second-stage operational amplifier U28 via a 1kΩ, 1% precision resistor R102. Pin 4 of the inverting input of U28 is connected to a pre-reserved adjustable resistor R101 (NC, not soldered by default) to AGND. A 1kΩ, 1% precision feedback resistor R100 is connected between pin 4 of the inverting input and pin 1 of the output. When R101 is left unsoldered, the circuit forms a unit non-inverting amplifier with a closed-loop gain of 1. Different values ​​of R101 can be soldered to adjust the gain levels. Pin 5 of the positive power supply of U28 is connected in parallel with a 10μF / 16V capacitor C178 and a 0.1μF / 50V capacitor C181 to AGND. Pin 2 of the negative power supply is connected in parallel with a 0.1μF / 50V capacitor C182 and a 10μF / 16V capacitor C18. From 3 to AGND, a dual-capacitor combination is also used to complete the power supply filtering and voltage regulation, ensuring the low-gain amplifier circuit operates stably with low temperature drift and low noise. The processed analog signal HALL_GAIN1 is output from pin 1 of U28 and connected to the ADC_IN1 port of the signal-to-noise ratio adaptive data fusion unit. The amplification factor of this low-gain conditioning branch is much lower than that of the high-gain branch, and it has an ultra-large dynamic input range of analog signals. There will be no voltage saturation clipping distortion in the near-end displacement range where the permanent magnet is close to the Hall and the magnetic field strength increases significantly. The effective induction signal corresponding to the large displacement of the entire stroke is completely preserved. All resistors and capacitors are selected with 1% precision matching components, and the resistor and capacitor parameters and filtering characteristics are completely consistent with those of the high-gain branch circuit, ensuring the hardware benchmark of the two acquisition channels is unified and reducing the inherent measurement deviation of the two channels.

[0056] like Figure 7As shown, the signal-to-noise ratio adaptive data fusion unit in this embodiment uses the STM32G431CB6 microcontroller U29 as the core control chip. The chip's digital power supply pins VDD1, VDD2, VDD, and VDDA are all connected to the VCC+3V3 power supply. Four power supply pins are connected in parallel with 0.1uF / 50V ceramic decoupling capacitors C184, C185, C186, and C187 to analog ground AGND to filter out high-frequency switching noise from the digital power supply and avoid interference from digital circuits with analog ADC sampling. The chip reset pin... Pin PG10-NRST is connected to VCC+3V3 via a 10K pull-up resistor R103 with 1% accuracy. A 0.1uF / 50V capacitor C188 is connected in parallel to AGND to form a power-on reset delay circuit, ensuring stable system power-on timing. The analog reference voltage is provided by the REF3033AIDBZT precision reference source U30. The input pin IN of U30 is connected to AVCC+5V, and a 1uF / 50V capacitor C189 is connected in parallel to AGND to filter the input power supply. The output pin OUT outputs 3V.A precise 3V reference voltage is fed into pin 20 of the VREF+ pin of U29. A 1uF / 50V capacitor C190 is connected to AGND on the output side to suppress reference voltage ripple, providing a high-precision, low-temperature-drift sampling reference level for the on-chip dual-channel synchronous ADC. The Hall high-gain conditioning branch output signal HALL_GAIN10 is connected to pin 8 of the PA0 pin of U29, and the low-gain conditioning branch output signal HALL_GAIN1 is connected to pin 9 of the PA1 pin of U29. These two analog signals are directly connected to the two independent synchronous ADC channels on the microcontroller, allowing parallel analog-to-digital conversion of the two analog voltages within the same clock cycle, eliminating dynamic phase deviation caused by time-division sampling. Pin 46 of PB8 is connected to AGND through a 10KΩ, 1% accuracy pull-down resistor R104 to achieve hardware configuration for startup mode. The chip peripherals include serial communication pins USART1_TX (pin 31 of PA9), USART1_RX (pin 32 of PA10), and USART1_EN (pin 33 of PA11), as well as SWDIO (pin 33 of PA11). Pin 36) and SWCLK (PA14, pin 37) are the program debugging and download pins; the U29 internal hardware implements a synchronous dual ADC sampling function with the same clock source. After synchronously acquiring the original sampling codes of PA0 and PA1 channels, it relies on the on-chip computing unit to complete a complete set of signal-to-noise ratio adaptive fusion algorithms: first, it uses VREF+ precision reference to normalize and convert the two ADC code values ​​into standard Hall sensing voltages, and calculates the real-time signal-to-noise ratio index of the two channels respectively. Based on the displacement conversion of the low-gain channel, it divides the working interval into three segments: far-end weak magnetic field, middle transition, and near-end strong magnetic field. According to the interval matching linear adaptive weight, it completes smooth weighted fusion calculation, and finally converts and outputs a high-precision displacement value with uniform and continuous signal-to-noise ratio throughout the entire stroke. The displacement detection result can be output externally through the USART serial port. The entire unit relies on a precision voltage reference to ensure the ADC sampling accuracy, and the synchronous ADC hardware architecture ensures instantaneous acquisition and alignment of the two signals. The digital fusion algorithm realizes the signal-to-noise ratio balance of the far and near end signals and eliminates the output jump when switching the range, fully realizing the core data processing function of the invention of full-stroke high signal-to-noise ratio Hall displacement detection.

[0057] Example 2

[0058] This embodiment provides a Hall displacement detection method with high signal-to-noise ratio across the entire stroke using dual ADC collaborative sampling, including the following steps:

[0059] Step 1: Use a Hall sensor to acquire signals and output Hall output voltage;

[0060] Step 2: Use a low-temperature drift follower circuit for impedance transformation and buffer isolation to output two stable Hall output voltages;

[0061] Step 3: The two stable Hall output voltages are gain conditioned based on high-gain conditioning circuits and low-gain conditioning circuits;

[0062] Step 4: Perform ADC quantization and correction, signal-to-noise ratio calculation, and segmented weighted fusion through the signal-to-noise ratio adaptive data fusion unit to output the final detected displacement.

[0063] Step 1 Hall sensor induced voltage model:

[0064] Output and displacement Corresponding induced voltage The larger the air gap (the farther the displacement), the better. The smaller the amplitude, the smaller the air gap (near the displacement). The larger the amplitude.

[0065] The distance between the permanent magnet and the Hall air gap is (Measured displacement), the Hall output voltage satisfies the gradient magnetic field relationship:

[0066]

[0067] in, For Hall chip sensitivity, Let be the magnetic flux density function that varies with displacement x. The magnetic field gradient constant of the permanent magnet. To compensate for the air gap at the zero point of the magnetic field, The measured linear displacement is denoted as .

[0068] From Formula 1, we can obtain that... Enlarge (remote), Decreasing the signal-to-noise ratio worsens. Reduce (proximal), The gain increases rapidly, and high-gain branches are prone to saturation.

[0069] Step 2: Low-temperature drift follower circuit output model:

[0070] The two low-temperature drift follower circuits are completely symmetrical, with perfectly matched op-amp and RC parameters, differing only in the back-end gain, thus introducing a uniform follower offset. Unrelated circuit thermal noise The two outputs are completely identical:

[0071]

[0072] in, This is the voltage value output by the high-gain, low-temperature drift follower circuit. The voltage value is the output of the low-gain, low-temperature drift follower circuit, and the two voltages are equal. This is the original voltage value output by the Hall sensor. This is the equivalent input offset voltage of the low-temperature drift follower circuit. This is noise from the low-temperature drift follower circuit.

[0073] Step 3: Output models of high-gain conditioning circuit and low-gain conditioning circuit:

[0074] In a high-gain conditioning circuit, the conditioning gain Additional op-amp mismatch Channel noise ADC full-range input limit Saturation determination conditions .

[0075]

[0076] in, This is the output voltage of the high-gain conditioning circuit, which is the voltage before it is connected to the analog-to-digital converter.

[0077] In a low-gain conditioning circuit, the conditioning gain Additional op-amp mismatch Channel noise ADC full-range input limit The entire journey It is always true and has no saturation.

[0078]

[0079] in, Output voltage of low-gain conditioning circuit.

[0080] Step 4: ADC quantization output model

[0081] In ADC quantization, both ADCs have N bits, and the quantization step size is... Quantization noise , Uncorrelated, ADC raw sampling code:

[0082]

[0083] in, This is the sampled data from the high-gain channel after analog-to-digital conversion. The sampled data is from the low-gain channel after analog-to-digital conversion. Synchronous sampling constraint: both ADCs convert simultaneously. , Corresponding to the same instantaneous displacement There is no time phase difference.

[0084] Step 4: ADC correction model

[0085] During ADC calibration, the factory-set stored parameters are: high-gain zero. High gain coefficient Low gain zero point Low gain coefficient The two ADC codes are converted into a standard induced voltage estimate. , .

[0086]

[0087] in, Estimation of the standard Hall effect voltage for the high-gain channel. This is an estimate of the standard Hall effect voltage for a low-gain channel. Ideally, without error... The only difference is in the noise component.

[0088] The signal-to-noise ratio calculation model in step 4:

[0089] like Figure 3 As shown, the signal-to-noise ratio of the two channels is obtained based on the Hall output voltage, and the method is as follows:

[0090] Total noise power of high-gain channel:

[0091]

[0092] in, The noise variance of the low-temperature drift follower circuit. The noise variance of the high-gain conditioning circuit. This represents the sampling quantization noise variance of a high-gain analog-to-digital converter.

[0093] Total noise power of low-gain channel:

[0094]

[0095] in, For the noise variance of the low-gain conditioning circuit, This represents the sampling quantization noise variance of a low-gain analog-to-digital converter.

[0096] signal RMS value The signal-to-noise ratio is defined as signal power / noise power.

[0097]

[0098] in, This represents the effective value of the Hall sensor output signal. The square of the effective value is the power of the Hall sensor output signal. Represents the total noise power; for high gain, take... Take at low gain .

[0099] Substituting the two approaches, we get:

[0100]

[0101] in, For high-gain channel signal-to-noise ratio. For low-gain channels, the signal-to-noise ratio is [not specified].

[0102] Distal micro-displacement range: , Significantly increase the effective signal amplitude. Significant amplification, with noise amplification factor much lower than signal amplification factor, satisfying... .

[0103] Near-end strong magnetic field region: Saturation distortion Failure efficient.

[0104] Therefore, if only a single high-gain channel is retained: near-end saturation occurs; if only a single low-gain channel is retained: far-end saturation occurs. Extremely low, micro-displacements are indistinguishable. The dual-channel complementary architecture of this invention mathematically resolves this inherent contradiction.

[0105] The equivalent signal-to-noise ratio after fusion is obtained based on the set fusion weights, as follows:

[0106] Set fusion weights , satisfy The expression for the standard induced voltage after fusion is:

[0107]

[0108] Total noise power after synthesis (two noise sources are uncorrelated, with a covariance of 0):

[0109]

[0110] Equivalent signal-to-noise ratio after integration:

[0111]

[0112] Distal micro-displacement range: , , Preserving a high signal-to-noise ratio;

[0113] Near-end strong magnetic field region: , , To avoid saturation;

[0114] Intermediate transition interval: weights change continuously and gradually. Smooth transition.

[0115] Step 4: Segmented weighted fusion

[0116] like Figure 4 As shown, the standard Hall effect voltage estimation is based on an adaptive weighted fusion complete mathematical algorithm to obtain the final measured displacement, including displacement interval determination threshold calibration, calculation of the standard Hall effect voltage based on the interval weight mathematical expression, and obtaining the final detected displacement based on the standard Hall effect voltage.

[0117] Low-gain, non-saturated channel conversion to basic displacement estimate Set two fixed remote thresholds:

[0118] Remote threshold :

[0119] Remote threshold :

[0120] Two far-end intervals and one transition interval are defined based on two fixed far-end thresholds:

[0121] constraint :

[0122] distant interval : , Comprehensive output It fully utilizes high-gain, high-signal-to-noise ratio channels.

[0123] distant interval : , Comprehensive output To avoid high-gain saturation distortion.

[0124] transition interval :

[0125] The weights change linearly and continuously with displacement, eliminating step jumps:

[0126]

[0127] Substituting into the standard induced voltage formula, we get:

[0128]

[0129] The fused standard induced voltage Substituting the Hall magnetic field-displacement inverse function, the final detected displacement is output:

[0130]

[0131] In summary, the digital fusion algorithm achieves signal-to-noise ratio equalization between near and far ends, eliminates output jumps during range switching, and fully realizes the core data processing function of the full-stroke high signal-to-noise ratio Hall displacement detection of this invention.

[0132] Figure 8 The measured displacement is only 1μm in the range of weak magnetic field and micro-displacement. The original waveform was obtained by single-channel unity gain acquisition.

[0133] Figure 9 The original waveform was obtained by periodic reciprocating motion within the near-end strong magnetic field displacement range, with a fixed unity gain.

[0134] The horizontal axis represents the time axis, and the vertical axis represents the amplitude of the ADC sampling output (in μm). The vertical spikes in the waveform represent the noise floor resulting from the superposition of circuit thermal noise, quantization noise, and electromagnetic spurious noise.

[0135] The following is a comparison one by one:

[0136] (1) Difference between noise floor and signal-to-noise ratio:

[0137] Figure 8 The far-end waveform is covered with dense, large-amplitude longitudinal spikes. The signal waveform is severely enveloped by noise, with the peak-to-peak noise level approaching 50% of the effective signal amplitude. The edges of the effective signal are blurred, and the signal-to-noise ratio is extremely low. Figure 9 The near-end waveform curve is smooth with very few spikes, and the noise floor amplitude is only within 5% of the effective signal amplitude. The waveform outline is clear and regular, and the signal-to-noise ratio is far superior to that of the far end. For the same 1μm micro-displacement, the signal-to-noise ratio of a single fixed gain channel shows a cliff-like difference simply due to changes in displacement distance and magnetic field strength. The far-end micro-displacement signal is easily submerged by noise, making it impossible to accurately identify micron-level displacement fluctuations.

[0138] (2) Differences in effective signal recognition:

[0139] Figure 8 The distant waveform noise spikes cover the complete sine contour, the peaks and troughs jitter violently, making it difficult to accurately determine the extreme values ​​of instantaneous displacement, and small displacement changes cannot be stably distinguished. Figure 9 The near-end waveform is smooth and continuous, with clear peak and trough boundaries, complete and distortion-free displacement trajectory, and stable identification of micron-level fluctuations.

[0140] (3) The defects in the consistency of measurement throughout the entire stroke are visually apparent:

[0141] The two sets of tests only changed the air gap distance (displacement range) between the permanent magnet and the Hall, and the displacement change amplitude was 1μm. Theoretically, the sensor output fluctuation amplitude should be consistent. However, the signal-to-noise ratio of the far-end signal was greatly degraded under the single fixed gain scheme, while the signal-to-noise ratio of the near-end was excellent. The detection accuracy and data stability were seriously unbalanced throughout the process, which could not meet the hard requirements of industrial precision closed-loop control for uniform measurement accuracy and low jitter signal throughout the process.

[0142] Based on the above description, the following conclusions can be drawn:

[0143] The waveform in the far-end region is superimposed with a large amount of high-amplitude random noise spikes, with the noise floor amplitude approaching half of the effective signal amplitude. The sinusoidal contour is severely obscured by noise, making it difficult to identify micron-level displacement fluctuations, resulting in severe data jitter and an extremely poor signal-to-noise ratio. In contrast, the waveform in the near-end region is smooth and regular, with an extremely low noise floor, a clear and complete signal contour, and an excellent signal-to-noise ratio. Relying solely on a single fixed-gain acquisition channel without adaptive signal-to-noise ratio weighted fusion processing has an inescapable technical flaw:

[0144] First, the micro-displacement signal under the weak magnetic field at the far end is easily submerged by noise, and the micron-level resolution is greatly reduced. If a single high-gain channel is used for the strong magnetic field at the near end, signal saturation distortion will occur, and it is impossible to simultaneously achieve high precision in micro-displacement and complete measurement of large stroke.

[0145] Second, the signal-to-noise ratio distribution is severely uneven throughout the entire range, and the detection error and jitter of the same amplitude displacement change are huge in the near and far ranges, resulting in poor consistency of the measurement throughout the entire range.

[0146] Third, high-noise output signals at the remote end can cause continuous oscillations in the precision closed-loop control mechanism and substandard positioning accuracy. The conventional fixed threshold switching channel method can also cause output jumps at the interval boundary, further deteriorating control stability.

[0147] This invention utilizes a dual-ADC synchronous acquisition algorithm combined with an adaptive signal-to-noise ratio smoothing weighted fusion algorithm. At the far end, high-gain channel data is used to suppress noise, while at the near end, low-gain channels are switched to avoid saturation. The weights are adjusted linearly and smoothly in the transition interval, resulting in a low-noise and smooth output waveform throughout the entire range. This achieves uniform high signal-to-noise ratio detection throughout the entire range, completely overcoming the inherent defects of the single fixed-gain acquisition scheme.

[0148] Beneficial effects can be seen Figure 10 and Figure 11 :

[0149] Figure 10 This is the output waveform of the 1μm periodic back-and-forth motion in the far-end weak magnetic field region after processing by the dual ADC collaborative sampling and signal-to-noise ratio adaptive weighted fusion algorithm of this invention; combined with Figure 8A comparative analysis was conducted on the high-noise original waveform at the far end and the smooth waveform at the near end without adaptive fusion. Under the same test conditions of weak magnetic field at the far end and reciprocating motion with a 1μm micro-displacement, without the adaptive signal-to-noise ratio fusion scheme, the single low-gain acquisition waveform was filled with large random noise spikes, the noise floor amplitude was close to half of the effective signal amplitude, the sine contour was severely obscured by noise, micron-level displacement fluctuations were difficult to identify, and the output data fluctuated violently. However, after applying the adaptive smoothing weighted fusion processing of this invention, the waveform noise spikes were completely suppressed, the curve was smooth and continuous, the peak and trough boundaries were clear and regular, the noise floor amplitude was reduced to less than 5% of the effective signal amplitude, and the smoothness and signal-to-noise ratio of the waveform in the far end section reached a level basically consistent with that of the waveform in the near end section. Figure 11 The waveform in the near-end strong magnetic field displacement range was not distorted or saturated due to the algorithm.

[0150] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A Hall displacement detection method of full stroke high signal-to-noise ratio realized by dual-ADC cooperative sampling, characterized in that: The Hall output voltage of the Hall sensor is processed to obtain the standard Hall sensing voltage estimate. The signal-to-noise ratio of the high-gain channel, the signal-to-noise ratio of the low-gain channel, and the equivalent signal-to-noise ratio after synthesis are calculated. The final detected displacement is obtained based on the adaptive weighted fusion complete mathematical algorithm.

2. The Hall displacement detection method of claim 1, wherein the full stroke high SNR is achieved by the dual-ADC coordinated sampling. The standard Hall effect voltage estimation method is as follows: the Hall output voltage enters two low-temperature drift follower circuits for impedance transformation and buffer isolation, and outputs two stable Hall output voltages. The two stable Hall output voltages are then subjected to gain conditioning by high-gain conditioning circuits and low-gain conditioning circuits, and output to the signal-to-noise ratio adaptive data fusion unit for quantization and correction to obtain the standard Hall effect voltage estimation.

3. The Hall displacement detection method of claim 2, wherein: The standard Hall effect voltage estimate obtained by quantization and correction using an ADC is: ； wherein, is the sampling data of the high-gain channel after analog-to-digital conversion, is the sampling data of the low-gain channel after analog-to-digital conversion, is the gain coefficient of the high-gain channel and the low-gain channel, is the gain zero point of the high-gain channel and the low-gain channel, is the standard Hall induction voltage estimate of the high-gain channel, is the standard Hall induction voltage estimate of the low-gain channel.

4. The Hall displacement detection method with high signal-to-noise ratio across the entire stroke achieved by dual ADC collaborative sampling as described in claim 1, characterized in that: The signal-to-noise ratios of the high-gain channel and the low-gain channel are: ； in, For high-gain channel signal-to-noise ratio. For low-gain channel signal-to-noise ratio. This represents the total noise power of the high-gain channel. This represents the total noise power of the low-gain channel. This is the Hall output voltage.

5. The Hall displacement detection method with high signal-to-noise ratio across the entire stroke achieved by dual ADC collaborative sampling according to claim 1, characterized in that: The combined equivalent signal-to-noise ratio is obtained based on the set fusion weights and the Hall output voltage; the fusion weights are set as follows: , ,satisfy The combined equivalent signal-to-noise ratio is: ； in, This represents the total noise power of the high-gain channel. This represents the total noise power of the low-gain channel. This is the Hall output voltage.

6. The Hall displacement detection method with high signal-to-noise ratio across the entire stroke achieved by dual ADC collaborative sampling according to claim 1, characterized in that: The adaptive weighted fusion complete mathematical algorithm includes displacement interval determination threshold calibration, calculation of standard Hall sensing voltage based on interval weight mathematical expression, and final detection displacement based on standard Hall sensing voltage.

7. The Hall displacement detection method with high signal-to-noise ratio across the entire stroke achieved by dual ADC collaborative sampling as described in claim 6, characterized in that: The displacement interval determination threshold calibration includes calculating the basic displacement estimate through a low-gain channel. Through basic displacement estimation Set two fixed remote thresholds , Two remote intervals and one transition interval are determined based on two fixed remote thresholds.

8. The Hall displacement detection method with high signal-to-noise ratio across the entire stroke achieved by dual ADC collaborative sampling according to claim 6 or 7, characterized in that: The interval weight expression includes: distant interval High-gain channels are used. distant interval Use a low-gain channel; transition interval Based on basic displacement estimation The fusion weight is determined by a fixed remote threshold, and the standard Hall effect voltage is obtained based on the fusion weight and the standard induced voltage estimate.

9. The Hall displacement detection method with high signal-to-noise ratio across the entire stroke achieved by dual ADC collaborative sampling according to claim 6, characterized in that: The final detected displacement expression is: ； in, This is the standard Hall effect sensing voltage. , To fix the remote threshold.

10. A Hall displacement detection circuit that achieves high signal-to-noise ratio across the entire stroke through dual ADC collaborative sampling, characterized in that: It includes a Hall sensor, a low temperature drift follower circuit, a gain conditioning circuit, and a signal-to-noise ratio adaptive data fusion unit; the Hall sensor outputs a Hall output voltage, and the gain conditioning circuit realizes impedance matching buffering and eliminates temperature drift offset. The gain conditioning circuit includes a high-gain conditioning circuit and a low-gain conditioning circuit. The signal-to-noise ratio adaptive data fusion unit uses dual ADCs synchronized by the same source clock to acquire the front-end signal in parallel. Combined with an adaptive weighted fusion complete mathematical algorithm, it achieves a uniform high signal-to-noise ratio throughout the Hall displacement stroke to obtain the final detected displacement.

Images