A laser precision angle rotation system based on multilayer PCB

Abstract

The application relates to the technical field of laser precision positioning equipment, and discloses a laser precision angle rotating system constructed based on a multilayer PCB. The laser precision angle rotating system comprises a multilayer PCB, integrated driving circuit, sensing circuit and control circuit, and is provided with an electromagnetic coil array on the multilayer PCB; a rotor is arranged in cooperation with the electromagnetic coil array, and the rotor is fixedly connected with a laser reflector and a grating; an angle feedback system comprises a fixedly arranged photoelectric sensor array, is used for reading signals generated when the grating rotates, and forms digital square wave signals after amplification, filtering and shaping of the signals; original angle data are obtained after four times frequency processing and position counting of the digital square wave signals; high-precision angle values are output after Kalman filtering processing of the original angle data; and the control module comprises an upper computer, a main controller and a driving chip. The system is suitable for scenes requiring high-precision angle control, such as laser scanning, optical image stabilization and precision measurement.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of laser precision positioning equipment, and particularly relates to a laser precision angle rotation system based on a multilayer PCB. BACKGROUND

[0002] In the fields of laser precision machining, optical measurement and precision instruments, a high-precision angle rotation system is a core component for realizing laser beam pointing control. A traditional angle rotation system usually adopts a separated design, and a driving circuit, a sensing circuit and a control circuit are independently laid out. A mechanical transmission structure (such as a gear or a worm gear) is used to connect a driving unit and a rotor, which leads to a large system size, angle errors caused by mechanical clearances, low integration and insufficient reliability. In addition, the angle feedback of the traditional system is usually achieved by a simple counting method, noise and jitter are not effectively filtered out, thermal deformation compensation relies on an empirical formula, and the compensation precision is limited, which is difficult to meet the requirements of precision angle control in microns or even nanometers.

[0003] In the prior art, an electromagnetic driving structure usually adopts a layout in which independent coils are separated from a PCB, and the installation precision of the coils depends on mechanical positioning, which increases assembly complexity. Angle signal processing is only achieved by simple filtering, and intelligent algorithms such as Kalman filtering are not combined to improve precision. Thermal management compensation lacks a real-time prediction mechanism and cannot dynamically adapt to changes in working conditions. Therefore, a system is needed to solve at least one of the above problems. SUMMARY

[0004] The present application provides a laser precision angle rotation system based on a multilayer PCB, which aims to solve the problems in the prior art that an electromagnetic driving structure usually adopts a layout in which independent coils are separated from a PCB, the installation precision of the coils depends on mechanical positioning, which increases assembly complexity; angle signal processing is only achieved by simple filtering, and intelligent algorithms such as Kalman filtering are not combined to improve precision; and thermal management compensation lacks a real-time prediction mechanism and cannot dynamically adapt to changes in working conditions.

[0005] In a first aspect, the embodiments of the present application provide a laser precision angle rotation system based on a multilayer PCB, which comprises:

[0006] A multilayer PCB, which integrates a driving circuit, a sensing circuit and a control circuit, and is provided with an electromagnetic coil array;

[0007] A rotor, which is arranged in cooperation with the electromagnetic coil array, and is fixedly connected with a laser reflector and a grating;

[0008] An angle feedback system comprises a fixedly arranged photoelectric sensor array for reading signals generated when the grating rotates, the signals are sequentially amplified, filtered and shaped to form a digital square wave signal, the digital square wave signal is subjected to four times frequency processing and position counting to obtain original angle data, and the original angle data is subjected to Kalman filtering processing to output a high-precision angle value.

[0009] A control module comprises an upper computer, a main controller and a driving chip, the upper computer is used for inputting instructions, the instructions include two modes of continuous trajectory and point-to-point and generating an angle instruction sequence, the angle instruction sequence is transmitted to the main controller through a protocol, the main controller generates a three-phase PWM signal and inputs the three-phase PWM signal to the driving chip according to the angle instruction sequence and the high-precision angle value, the driving chip outputs a three-phase current to drive the electromagnetic coil array.

[0010] In some embodiments, the electromagnetic coil array is integrated in the conductive layers of the multilayer PCB board in a layered manner, the electromagnetic coils of each layer are arranged in a ring shape at a preset angle interval, the electromagnetic coils of adjacent layers are distributed in a staggered manner in the circumferential position, and the electromagnetic coil array is directly electrically connected with the driving circuit of the multilayer PCB board, the electrical connection of the coils of each layer is realized through the conductive through holes of the multilayer PCB board, and an embedded electromagnetic driving structure without mechanical transmission components is formed.

[0011] In some embodiments, the analog signal output by the photoelectric sensor array is first input to a preamplifier for gain-adjustable amplification processing, the amplified signal is filtered to remove high-frequency and low-frequency noise through a band-pass filter with a fixed center frequency and a limited bandwidth, and the filtered signal is compared with a preset threshold voltage through a voltage comparator to convert the analog signal into a digital square wave signal with fixed high and low levels.

[0012] In some embodiments, the rotor is a centrally symmetric disc structure, the central axis of the rotor coincides with the central axis of the electromagnetic coil array, the laser reflector is fixedly installed on the upper surface of the rotor and rotates synchronously with the rotor, the grating is fixedly installed in the edge region of the rotor in a concentric manner, and the scale line direction of the grating is consistent with the radial direction of the rotor, so that the grating can be read by the photoelectric sensor array when the grating rotates with the rotor.

[0013] In some embodiments, the digital square wave signal is divided into A-phase and B-phase signals with a phase difference of 90 degrees, which are input to a programmable logic device for quadrupling processing, that is, a counting pulse is triggered at the rising and falling edges of each signal period, and the quadrupled pulse signal is counted by a 32-bit counter. The original rotation angle data of the rotor is calculated based on the preset number of grating lines and the pulse counting result.

[0014] In some embodiments, the original angle data is output as a high-precision angle value after Kalman filtering processing, including: establishing a state space model containing angle state and angular velocity state, inputting the original angle data as a measurement value into a Kalman filter, and iteratively estimating the angle state through a prediction step and an update step; wherein the prediction step predicts the current state based on the state at the previous time and the system dynamics model, and the update step corrects the predicted state combined with the current measurement value, and finally outputs a high-precision angle value filtered from random noise and jitter.

[0015] In some embodiments, the main controller inputs the difference between the angle instruction sequence and the high-precision angle value into a position PID controller to calculate a speed instruction; the speed instruction is superimposed with an acceleration feedforward compensation amount and input into a field-oriented control module, the current in a three-phase stationary coordinate system is converted into direct-axis and quadrature-axis current instructions in a rotating coordinate system through coordinate transformation, and the direct-axis and quadrature-axis currents are controlled in a closed loop based on a PI controller; the control output voltage vector generates three PWM signals through a space vector pulse width modulation algorithm, and the duty cycle of the PWM signal is determined according to the amplitude and phase of the voltage vector and input into the drive chip.

[0016] In some embodiments, it further includes a thermal management compensation system, including temperature sensors distributed at key positions of the multi-layer PCB board, which collect temperature data in real time and input into a thermal deformation model, the thermal deformation model predicts the thermal expansion trend of the structure combined with historical data and outputs a thermal deformation angle prediction value, which is superimposed on the angle instruction sequence as a feedforward compensation value to achieve real-time compensation.

[0017] In some embodiments, the main controller inputs the temperature data and the current coil current value into a pre-trained long short-term memory neural network model, which outputs a predicted thermal deformation angle value; the main controller superimposes the thermal deformation angle value as a feedforward compensation value into the angle instruction sequence to achieve real-time compensation for the angle deviation caused by rotor thermal expansion, and the system uses new data to incrementally update the parameters of the long short-term memory neural network model during idle periods.

[0018] In some embodiments, further comprising: a safety monitoring module for monitoring the three-phase coil current, the temperature of each critical position, and the difference between the command angle and the feedback angle in real time, and triggering a protection mechanism to shut down the drive and report fault information when the monitored values exceed the set threshold.

[0019] The application relates to the technical field of laser precision positioning equipment, in particular to a high-precision electromagnetic driving angle rotation system based on multi-layer PCB integrated design, which is suitable for scenes requiring high-precision angle control such as laser scanning, optical image stabilization and precision measurement.

[0020] The driving circuit, the sensing circuit, the control circuit and the electromagnetic coil array are integrated by the multi-layer PCB, mechanical connecting components in traditional separate design are eliminated, the system volume is significantly reduced, the structural compactness and reliability are improved, and angle errors caused by mechanical transmission gaps are avoided. The angle feedback system improves the angle measurement accuracy to sub-arcsecond level through four times frequency processing and Kalman filtering algorithm, and effectively filters out grating signal noise and rotor jitter. The control module adopts field-oriented control (FOC) and space vector pulse width modulation (SVPWM), and combines speed feedforward compensation to realize dynamic response speed improvement and torque ripple suppression.

[0021] The safety monitoring module monitors the current, temperature and angle deviation in real time, quickly shuts down the drive when the threshold is triggered, avoids overcurrent, overheating and out-of-control risks, and improves the system fault protection response speed. The electromagnetic coil array and the circuit are integrated based on mature PCB manufacturing process, the independent component assembly steps are reduced, the processing cost is reduced, and batch production is suitable.

[0022] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the application. BRIEF DESCRIPTION OF DRAWINGS

[0023] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor.

[0024] Figure 1 is a structure schematic diagram of a laser precision angle rotation system based on multi-layer PCB construction based on self-focusing optical fiber coupling scheme provided by an embodiment of the application.

[0025] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the application. DETAILED DESCRIPTION

[0026] With reference to the drawings, the technical solutions in the embodiments of the present application will be clearly and completely described below. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of the present application.

[0027] The flowcharts shown in the drawings are only illustrative, and do not necessarily include all contents and operations / steps, nor are they necessarily executed in the order described. For example, some operations / steps can be further decomposed, combined or partially merged, so that the actual execution order can be changed according to the actual situation.

[0028] It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present application, in the embodiments of the present application, the terms "first", "second", etc. are used to distinguish the same items or similar items with basically the same functions and effects. Those skilled in the art can understand that the terms "first", "second", etc. do not limit the quantity and execution order, and the terms "first", "second", etc. also do not necessarily mean that they are different.

[0029] It should be understood that the terms used in the present application specification are only for the purpose of describing specific embodiments and do not intend to limit the present application. As used in the present application specification and the appended claims, unless otherwise clear from the context, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0030] It should also be understood that the term "and / or" used in the present application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations.

[0031] Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The following embodiments and features in the embodiments can be combined with each other without conflict.

[0032] In the field of laser precision machining, optical measurement and precision instruments, a high-precision angle rotation system is a core component for realizing laser beam pointing control. Traditional angle rotation systems usually adopt a separate design, with the driving circuit, sensing circuit and control circuit independently laid out, relying on mechanical transmission structures (such as gears, worm gears and worm shafts) to connect the driving unit and the rotor, resulting in a large system size, angle errors caused by mechanical clearances, low integration and insufficient reliability. In addition, the angle feedback of the traditional system is mostly in the form of simple counting, without effectively filtering out noise and jitter, and the thermal deformation compensation relies on empirical formulas, with limited compensation accuracy, making it difficult to meet the requirements of micron-level or even nanometer-level precision angle control.

[0033] In the prior art, the electromagnetic driving structure is mostly arranged separately from the PCB, the installation precision of the coil depends on mechanical positioning, and the assembly complexity is increased; the angle signal processing is only through simple filtering, without combining intelligent algorithms such as Kalman filtering to improve the precision; the thermal management compensation lacks a real-time prediction mechanism and cannot dynamically adapt to changes in working conditions. Therefore, there is an urgent need for a system to solve at least one of the above problems.

[0034] To solve the above problems, please refer to Figure 1 The embodiment of the application provides a laser precision angle rotation system based on a multi-layer PCB, which comprises a multi-layer PCB, an integrated driving circuit, a sensing circuit and a control circuit, and an electromagnetic coil array arranged on the multi-layer PCB; a rotor, which is arranged in cooperation with the electromagnetic coil array, and a laser reflector and a grating fixedly connected to the rotor; an angle feedback system comprising a fixedly arranged photoelectric sensor array, the photoelectric sensor array being used to read signals generated when the grating rotates, the signals being sequentially subjected to amplification, filtering and shaping to form a digital square wave signal, the digital square wave signal being subjected to four times frequency processing and position counting to obtain original angle data, and the original angle data being subjected to Kalman filtering processing to output a high-precision angle value; a control module comprising an upper computer, a main controller and a driving chip, the upper computer being used to input instructions, the instructions comprising two modes of continuous trajectory and point-to-point and generating an angle instruction sequence, the angle instruction sequence being transmitted to the main controller through a protocol, the main controller sequentially performing position control loop calculation, speed feedforward compensation, magnetic field oriented control and space vector pulse width modulation processing according to the angle instruction sequence and the high-precision angle value, generating a three-phase PWM signal and inputting the three-phase PWM signal to the driving chip, and the driving chip outputting a three-phase current to drive the electromagnetic coil array.

[0035] Specifically, the system is designed by integrating the multi-layer PCB, the electromagnetic driving, the angle sensing, the control algorithm and the like, forms a full-electronic precision angle control scheme without mechanical transmission, and uses the multi-layer PCB to embed the electromagnetic coil array, integrates the driving / sensing / control circuit, eliminates the mechanical connection gap of the traditional separated driving, acquires the angle signal through the grating-photoelectric sensor combination, improves the measurement precision through multi-stage signal processing and Kalman filtering, adopts the magnetic field oriented control (FOC) combined with the speed feedforward compensation to realize the dynamic response optimization of the electromagnetic driving, and compensates the temperature drift through the thermal deformation prediction model.

[0036] Multi-layer PCB board: 6-10 layers of FR-4 or high thermal conductivity substrate, 2-4 layers as conductive layers, integrating the following functional modules: drive circuit layer: containing power MOSFET, freewheeling diode, etc. Driver devices are directly connected to the electromagnetic coil array through copper foil lines; Sensing circuit layer: integrating preamplifier, bandpass filter, voltage comparator, etc. Signal conditioning circuit, processing the analog signals output by the photoelectric sensor; Control circuit layer: arranging main controller (such as DSP or ARM chip), programmable logic device (FPGA), communication interface (such as SPI / I2C / UART), realizing control algorithm and data interaction.

[0037] Through etching a ring-shaped arrangement of planar spiral coils on each conductive layer, the number of single coils is 6-12 groups, evenly distributed at an interval of 30°-60°; The coils of adjacent layers are staggered by 15°-30° in the circumferential position, and are connected in series or parallel through conductive vias, forming a three-dimensional magnetic field distribution; The coil wire width is 0.2-0.5mm, and the wire spacing is 0.3-0.6mm, which improves the conductivity through PCB electroplating process and reduces copper loss. After the electromagnetic coil array is powered on, it generates an alternating magnetic field, which interacts with the permanent magnet (or induced eddy current) on the rotor to drive the rotor to rotate around the central axis. Since the coil is directly integrated on the PCB, no additional mechanical fixation is needed, eliminating the problem of magnetic field non-uniformity caused by traditional coil assembly deviation.

[0038] Adopting a central symmetric disc structure, the material is aluminum alloy or titanium alloy, and the surface is plated with nickel-chromium alloy to improve hardness; The central shaft is coaxially installed with the center hole of the PCB board, and cooperates with high-precision bearings (radial runout ≤1μm) to ensure rotation accuracy; The laser mirror is fixed on the upper surface of the rotor by vacuum adsorption or epoxy resin adhesion, with a mirror surface flatness of ≤λ / 10 (λ=632.8nm); The grating uses a glass transmission type incremental grating with 1000-2000 lines per circle, and the scale lines are distributed along the radial direction of the rotor, forming a reading system with the photoelectric sensor array (fixed on the edge of the PCB board). The distance between the rotor and the PCB board is controlled at 0.5-1mm, and the axial movement is prevented by limiting structure to ensure that the electromagnetic air gap uniformity error is less than 5%.

[0039] Photoelectric sensor array: contains 4 groups of opposite photoelectric sensors, output two-way orthogonal (phase difference 90 °) analog signal (A, B phase), amplitude 0-5V; signal conditioning process includes: amplification: through the instrument amplifier (such as AD620) gain adjustable amplification (gain range 10-100 times), compensation grating signal attenuation; filter: using four-order Butterworth band-pass filter (center frequency 10kHz, bandwidth 5kHz), filter out the low frequency noise (<1kHz) and high frequency electromagnetic interference (>20kHz) introduced by mechanical vibration; shaping: through the hysteresis comparator (such as LM311) to convert the filtered sine wave into TTL level square wave signal, hysteresis voltage 50-100mV to prevent noise jitter.

[0040] Four times frequency and counting through FPGA receive A, B phase square wave signal, detect each edge (rising edge / falling edge), realize four times frequency (resolution is improved 4 times), through 32-bit counter accumulates pulse number. Angle calculation formula: original angle = (pulse count × 360°) / (grating line number × 4).

[0041] By establishing a two-dimensional state space model, the state quantity is angle θ and angular velocity ω, and the observation value is the original angle data after four times frequency. Through the prediction equation (θk = θk-1 + ωk-1*Δt, ωk= ωk-1) and the update equation, the measurement noise is corrected in real time, and the output noise standard deviation is less than 0.1″ high precision angle value.

[0042] The host computer inputs the target angle (point-to-point mode) or trajectory curve (continuous mode) through the human-machine interface, generates an instruction sequence containing time-angle parameters, and transmits it to the main controller through the CAN / LAN bus. The main controller calculates the deviation between the instruction angle and the feedback angle, and outputs the speed instruction through the PID controller, the proportional coefficient Kp=200-500, the integral time Ti=0.01-0.05s, and the differential time Td=0.001-0.005s; speed feedforward compensation: according to the instruction angular velocity pre-computes the required electromagnetic torque, superimposed to the PID output, reduces the dynamic response lag (response time <1ms);

[0043] FOC (Field Oriented Control) converts the measured three-phase current into direct-axis current i_d (field current) and quadrature-axis current i_q (torque current) through Clark transformation (three-phase static→two-phase static) and Park transformation (two-phase static→two-phase rotating); PI controller is used to control i_d and i_q (i_d*=0 to realize unit power factor control) respectively, and output voltage commands u_d and u_q in rotating coordinate system; SVPWM (Space Vector Pulse Width Modulation) converts u_d and u_q into PWM signals of three-phase bridge arm, carrier frequency is 20 kHz, and seven-segment modulation strategy is used to reduce switching loss, and output voltage harmonic distortion is less than 5%. The driving chip receives the PWM signal, controls the three-phase bridge circuit to output 0-48V and 0-5A current to drive the electromagnetic coil, and the current sampling resistor (precision 1%) feeds back to the main controller in real time to realize overcurrent protection.

[0044] Exemplarily, the corresponding parameters are shown in the following table:

[0045]

[0046] In some embodiments, the electromagnetic coil array is integrated in the conductive layers of the multi-layer PCB board in a layered manner, the electromagnetic coils of each layer are arranged in a ring shape at a preset angle interval, the electromagnetic coils of adjacent layers are distributed in a staggered manner in the circumferential position, and the electromagnetic coil array is directly electrically connected with the driving circuit of the multi-layer PCB board, the electrical connection of the coils of each layer is realized through the conductive vias of the multi-layer PCB board, and an embedded electromagnetic driving structure without mechanical transmission components is formed.

[0047] The core of the embodiment is to realize layered embedded integration of the electromagnetic coil array through the conductive layers of the multi-layer PCB, and eliminate the mechanical connection components of the traditional driving structure. Specifically, the electromagnetic coils are layered and deployed in the conductive layers of the multi-layer PCB according to functions, each layer of coils is independently etched and interconnected through conductive vias; the single-layer coils are arranged in a ring shape at a preset angle interval (such as 30°, 45°), the coils of adjacent layers are misaligned by 15°-30° in the circumferential position, and a three-dimensional magnetic field superposition effect is formed; the coils are directly connected with the driving circuit through the copper foil of the PCB, without external cables or mechanical fixing structure, and the integration of the driving unit and the circuit is realized.

[0048] The PCB layered structure adopts an 8-layer PCB board, in which the 2nd, 4th and 6th layers are coil layers, each layer has a thickness of 50μm copper foil; each coil layer is etched with 8 groups of planar spiral coils, which are distributed in a ring shape at an interval of 45° with the center of the PCB as the center, the outer diameter of a single coil is 5mm, the inner diameter is 3mm, the line width is 0.3mm, and the line spacing is 0.4mm.

[0049] The adjacent layer staggered design is formed by the 0° starting position of the 2nd layer coil, the 22.5° starting position of the 4th layer coil (i.e. staggered by 1 / 8 period), and the 45° starting position of the 6th layer, so that each layer of the coil is staggered and superimposed in the circumferential direction; the layers are connected through the conductive through holes with a diameter of 0.3 mm, the adjacent layer coils are connected in series (for example, the output end of the 2nd layer coil 1 is connected to the input end of the 4th layer coil 1 through the through hole), and the total inductance value is adjustable through the interlayer connection mode (series connection increases the inductance, and parallel connection reduces the impedance). The electrical connection and the driving circuit are integrated, the copper foil of the coil layer is directly connected to the power MOSFET of the driving circuit layer, the drain electrode is connected to the input end of the coil, and the source electrode is connected in series with a current diode (for example, a fast recovery diode); the driving circuit is powered by the power supply layer (the 3rd and 5th layers) in the PCB, the voltage range is 12-48V, and the current carrying capacity is 5A / mm 2 Copper foil (satisfying the peak current requirement of the coil).

[0050] In some embodiments, the analog signal output by the photoelectric sensor array is first input to a preamplifier for gain-adjustable amplification processing, the amplified signal is filtered by a band-pass filter with a fixed center frequency and a limited bandwidth to filter out high-frequency and low-frequency noise, and the filtered signal is compared with a preset threshold voltage by a voltage comparator to convert the analog signal into a digital square wave signal with fixed high and low levels.

[0051] The present embodiment clearly defines the three-stage preprocessing process of the analog signal in the angle feedback system, including adjustable gain amplification, band-pass filtering, and threshold comparison, to ensure the reliability of the digital square wave signal. Specifically, the weak analog signal output by the photoelectric sensor is amplified with adjustable gain to compensate for signal attenuation; a filter with a fixed center frequency and a limited bandwidth is used to filter out mechanical vibration (low frequency) and electromagnetic interference (high frequency) noise; and the filtered sinusoidal wave is converted into a square wave signal with standard TTL level by a voltage comparator, eliminating noise jitter.

[0052] The preamplification circuit includes: device selection: instrument amplifier AD620, gain set by external resistor Rg (gain formula G=1+49.4kΩ / Rg, adjustable range 10-200 times); input signal: orthogonal analog signals (A phase, B phase) output by the photoelectric sensor (peak-to-peak value 0.5-2V, DC bias 2.5V); circuit design: differential input to suppress common-mode noise, input end in parallel with 100nF decoupling capacitor to suppress high-frequency interference.

[0053] The band-pass filter circuit includes: filter type: fourth-order Butterworth band-pass filter, center frequency f0=10 kHz, passband width BW=8 kHz (cut-off frequency 2 kHz-18 kHz); element parameters: designed with operational amplifier OPA227, resistance R=10 kΩ, capacitance C=1.59 nF (calculated according to f0=1 / (2πRC)); filter effect: attenuate mechanical vibration noise (<1 kHz, such as bearing vibration) and PWM switching noise (>20 kHz, suppression amplitude >40 dB).

[0054] The voltage comparator circuit includes: device selection: hysteresis comparator, reference voltage Vref=2.5 V (consistent with the sensor bias voltage); hysteresis voltage setting: ±50 mV (achieved by feedback resistance voltage division), prevent output signal oscillation caused by noise; output signal: square wave signal converted to 3.3 V TTL level, rise / fall time <100 ns, duty cycle 50%±5%. The pre-processing circuit is integrated in the PCB sensor layer, close to the photoelectric sensor (distance <5 mm), reducing the noise introduced by long wires;

[0055] Analog ground and digital ground are single-point common ground through 0Ω resistance, suppressing ground loop interference.

[0056] In some embodiments, the rotor is a central-symmetric disc structure, the central axis of the rotor coincides with the central axis of the electromagnetic coil array, the laser mirror is fixedly installed on the upper surface of the rotor and rotates synchronously with the rotor, the grating is fixed concentrically on the edge region of the rotor, and the scale line direction of the grating is consistent with the radial direction of the rotor, so that the grating can be read by the photoelectric sensor array when the grating rotates with the rotor.

[0057] The embodiment defines the mechanical structure of the rotor and the installation method of the optical element, and ensures that the laser mirror and the grating rotate synchronously with high precision. The core design includes: central-symmetric disc structure: the geometric center of the rotor coincides with the central axis of the electromagnetic coil, reducing the rotation error caused by eccentricity; mirror and grating coaxial installation: the mirror is fixed on the central region of the upper surface of the rotor, and the grating is arranged concentrically on the edge, and the scale lines are distributed radially, facilitating reading by the photoelectric sensor.

[0058] The rotor body design includes: material: aviation aluminum alloy (such as 6061-T6), density 2.7 g / cm 3 , thermal expansion coefficient 23.6×10 -6 / ℃, surface anodizing treatment; structure: diameter 60 mm, thickness 5 mm, central axis hole diameter 8 mm (cooperate with high-precision bearing), edge region thickness 3 mm (reduce weight, moment of inertia ≤5×10 -6 kg*m 2Machining accuracy: perpendicularity of the central shaft hole to the rotor end face ≤5μm, circumferential runout ≤2μm.

[0059] Laser reflector installation: Reflector specifications: diameter 20mm, thickness 2mm, surface coated with a high-reflectivity film (reflectivity > 99.5% @ 632.8nm), flatness ≤ λ / 20; fixing method: low expansion coefficient epoxy resin (such as EPO-TEK 353ND) is used to bond to the center of the upper surface of the rotor, the adhesive layer thickness is ≤ 50μm, and the angle is calibrated after curing (the perpendicularity between the reflector normal and the rotor axis is ≤ 10″).

[0060] The grating installation and scale design include: Grating type: transmissive glass incremental grating, 2000 lines / revolving loop, outer diameter 50mm, inner diameter 40mm; Installation position: The grating is fixed to the rotor edge (25mm from the central axis) by three evenly distributed elastic pressure plates, and the scale lines extend radially along the rotor (angle ≤1° with the radial direction); Reading fit: The photoelectric sensor array (4 sets of through-beam type) is fixed to the PCB edge, maintaining a 0.5mm gap with the grating to ensure light transmittance >80%. Mechanical fit: Bearing selection: Deep groove ball bearing EZO R8804 is used, with radial runout ≤1μm and axial runout ≤2μm; Air gap control: The distance between the rotor and the PCB coil layer is 1mm, axially positioned by a limiting ring (1mm thick), and the air gap uniformity error is <3%.

[0061] In some embodiments, the digital square wave signal is divided into phase A and phase B signals with a 90-degree phase difference. The phase A and phase B signals are input to a programmable logic device for frequency quadrupling processing, that is, a counting pulse is triggered at the rising and falling edges of each signal cycle. The frequency quadrupled pulse signal is accumulated and counted by a 32-bit counter. The original rotation angle data of the rotor is calculated based on the preset number of grating lines and the pulse counting result.

[0062] This embodiment improves the angle measurement resolution by quadrating the frequency of the orthogonal A / B phase signals. Specifically, it includes: orthogonal signals with a 90° phase difference: the A and B phase signals are output by a photoelectric sensor to determine the rotation direction and achieve edge triggering; the quadrating principle: counting is triggered on both the rising and falling edges of each signal cycle, increasing the resolution to four times the original number of grating lines; and a 32-bit counter: the pulse count is accumulated and converted into an angle value to meet the requirements of high-precision angle calculation.

[0063] The quadrature signal characteristics include: a 90° phase difference between phase A and phase B; phase A leading phase B by 90° when the rotor rotates forward; and phase B leading phase A by 90° when rotating in reverse; signal frequency: f = ω × N / (2π), where ω is the rotor angular velocity (rad / s), N is the number of grating lines (e.g., 2000 lines / revolution), and the maximum frequency is ≤50kHz (corresponding to a rotation speed of 3000rpm).

[0064] Quadruple frequency multiplication: Hardware carrier: Programmable logic device (FPGA, such as Xilinx Spartan-6), with an integrated edge detection module; Edge detection: The rising edge (↑) and falling edge (↓) of phase A and phase B are detected separately to generate 4 counting trigger signals (A↑, A↓, B↑, B↓); Direction determination: By comparing the current state of phase A / B with the previous state (e.g., A(n)=1, B(n)=0, A(n-1)=0, B(n-1)=0, it is determined to be forward rotation), the counter is controlled to increment or decrement.

[0065] The counting and angle calculation includes: Counter: 32-bit unsigned register, maximum count value 4,294,967,296, corresponding angle measurement range > 360°×4,294,967,296 / (2000×4)=193,712,424° (satisfies full circumference measurement);

[0066] Angle calculation formula: θ = (pulse count × 360°) / (number of grating lines × 4) = (pulse count × 360°) / (2000 × 4) = pulse count × 0.045°. The resolution is 0.045° / pulse = 162″ / pulse. After frequency quadrupling, the resolution is improved by 4 times to 40.5″ / pulse compared to the original grating resolution (360° / 2000 = 108″ / line) (in practice, it is further improved to sub-arcsecond level through Kalman filtering). Metastability noise is filtered out by setting two-level register synchronization at the FPGA input port; overflow detection of the counter is performed every 1ms (when the count value > the upper limit, it is reset to zero and an overflow flag is marked) to ensure the continuity of angle calculation.

[0067] In some embodiments, the high-precision angle value is output after the raw angle data is processed by Kalman filtering, including: establishing a state-space model containing angle state and angular velocity state; inputting the raw angle data as a measurement value into the Kalman filter; iteratively estimating the angle state through a prediction step and an update step; wherein, the prediction step predicts the current state based on the state at the previous moment and the system dynamics model, and the update step corrects the predicted state in combination with the current measurement value, and finally outputs a high-precision angle value after filtering out random noise and jitter.

[0068] This embodiment establishes a two-dimensional state-space model and performs Kalman filtering on the original angle data to filter out random noise and high-frequency jitter, outputting high-precision angle values. The core steps include: State-space modeling: defining angle θ and angular velocity ω as state variables, and constructing prediction equations incorporating system dynamics; Kalman filtering iteration: using a "prediction-update" loop, predicting the current state using the previous state, and combining this with measured values ​​to correct prediction errors, thus achieving noise suppression.

[0069] State vector: xk=[θk, ωk]T Where θ is the angle (rad) and ω is the angular velocity (rad / s). Prediction equation (based on uniform motion model): xk∣k⁻¹=Fxk⁻¹∣k⁻¹+Buk; State transition matrix F= Control input matrix B = (Considering the influence of acceleration u, u=0 in this embodiment, simplifying to a uniform velocity model). Observation equation: zk=Hxk+vk, H=[1,0], the observed value z is the original angle data after fourth harmonic, and vk is the measurement noise (Gaussian distribution, standard deviation σz=10″).

[0070] Kalman filter parameter settings include: process noise covariance matrix Q = Angular noise σθ = 5″, angular velocity noise σω = 2″ / ms. Measurement noise covariance matrix: R = σz 2 =(10″) 2 Initial state estimation: x0 = [0,0] T The covariance matrix P0 = diag([(30″)) 2 (5″ / ms) 2 The iterative steps include calculating the state prediction, calculating the prediction covariance, calculating the Kalman gain, correcting the state estimate, and updating the covariance.

[0071] In some embodiments, the main controller inputs the difference between the angle command sequence and the high-precision angle value into the position PID controller to calculate the speed command; the speed command is superimposed with the speed feedforward compensation and then input into the field-oriented control module, which converts the current in the three-phase stationary coordinate system into direct-axis and quadrature-axis current commands in the rotating coordinate system through coordinate transformation, and performs closed-loop control of the direct-axis and quadrature-axis currents based on the PI controller; the voltage vector of the control output is used to generate three PWM signals through a space vector pulse width modulation algorithm, and the duty cycle of the PWM signal is determined according to the amplitude and phase of the voltage vector and input to the driver chip.

[0072] This embodiment constructs a closed-loop control chain of "position loop - PID - speed feedforward - FOC - SVPWM" to achieve high-precision angle tracking. Position error calculation uses the deviation between the target angle and the filtered angle as input to the PID controller to generate speed commands; speed feedforward compensation reduces control lag by pre-calculating the torque required for dynamic response; field-oriented control separates the excitation current and torque current through coordinate transformation, enabling independent control and improving response speed; SVPWM modulation is used to generate a high-efficiency PWM signal to drive the electromagnetic coil.

[0073] Position PID controller: Input / output: Error e(t) = θ∗(t) - θ(t), Output speed command vcmd(t); where:

[0074] ;

[0075] Parameter range: Kp=300-500, Ki=10-20, Kd=0.5-1.0 (adaptively adjusted according to rotor inertia); Anti-saturation design: Set output limit ±1000° / s to prevent integral saturation.

[0076] Speed ​​feedforward compensation includes: Torque pre-calculation: Tff = Jω'cmd + Bωcmd, where J is the rotor moment of inertia (5 × 10⁻⁶). -6 kg*m 2 B is the viscous damping coefficient (0.001 N*m*s / rad); current command mapping: iqff =Tff / Kt (Kt is the torque constant, 0.1 N*m / A), which is superimposed on iq* of the PID output.

[0077] Field-oriented control (FOC) includes: coordinate transformation: Clark transformation (three-phase → two-phase stationary), Park transformation (two-phase stationary → two-phase rotating), current loop PI control, carrier parameters, voltage vector synthesis, and drive output.

[0078] In some embodiments, the system further includes a thermal management compensation system, comprising temperature sensors distributed at key locations on the multilayer PCB board. The temperature sensors collect temperature data in real time and input it into a thermal deformation model. The thermal deformation model combines historical data to predict the thermal expansion trend of the structure and outputs a predicted value of the thermal deformation angle. The predicted value of the thermal deformation angle is superimposed on the angle command sequence as a feedforward compensation value to achieve real-time compensation.

[0079] This embodiment uses temperature sensors to collect real-time temperature rise data at key locations on the PCB, inputs it into a thermal deformation model to predict angular deviations caused by structural expansion, and eliminates temperature drift through feedforward compensation. The temperature acquisition network deploys NTC sensors at multiple points, covering heat sources (coils, driver chips) and sensitive areas (encoders); thermal deformation prediction is based on historical data or physical models, outputting predicted thermal deformation angle values; real-time compensation dynamically corrects the temperature effect by superimposing the prediction deviation onto the instruction sequence.

[0080] Temperature sensor deployment includes: location distribution: center of coil layer (1), edge of coil layer (2) - monitoring electromagnetic heating; surface of driver chip (1), encoder circuit board (1) - monitoring circuit heating; sensor type: NTC thermistor (B value 3950K, accuracy ±0.3℃, model MF52E), with 10kΩ voltage divider resistor to collect voltage signal, which is converted by 12-bit ADC (resolution 0.1℃).

[0081] The thermal deformation model includes: thermal expansion formula: Δθ=αΔTL / r×180° / π, where α is the coefficient of thermal expansion of the rotor material (23.6×10⁻⁻⁶). 6 / ℃), ΔT is the temperature rise (℃), L is the distance from the rotor edge to the central shaft (25mm), r is the grating radius (25mm), simplified to Δθ=αΔT×360° / 2π (angular deviation is related to circumferential expansion); dynamic correction is achieved by collecting temperature data every 5ms, calculating the average temperature rise ΔT, and updating the compensation amount in real time (response delay <10ms).

[0082] The compensation control logic includes: instruction superposition: target angle θcomp∗ =θcmd∗+Δθpred, where Δθpred is the prediction deviation (when the rotor outer diameter expands during positive temperature rise, the angle feedback lags, and the compensation direction is opposite to the expansion direction); when the temperature change rate at any measuring point is >0.5℃ / min, dynamic compensation is enabled; the compensation amount is set to 0 in the stable temperature zone (ΔT<1℃).

[0083] In some embodiments, the main controller inputs the temperature data and the current coil current value into a pre-trained long short-term memory neural network model, and the model outputs a predicted thermal deformation angle value; the main controller superimposes the thermal deformation angle value as a feedforward compensation value into the angle command sequence to realize real-time compensation for the angle deviation caused by rotor thermal expansion, and the system incrementally updates the parameters of the long short-term memory neural network model using new data during idle periods.

[0084] This embodiment employs a Long Short-Term Memory (LSTM) neural network to construct a thermal deformation model. Inputting temperature data and coil current (reflecting heating power), it outputs a predicted thermal deformation angle, achieving nonlinear, dynamic thermal drift compensation. Data-driven modeling utilizes historical temperature rise-angle deviation data to train the model, adapting to complex thermal coupling effects. An incremental update mechanism optimizes model parameters with new data during system idle periods, maintaining prediction accuracy. Feedforward compensation overlays real-time predicted values ​​onto the instruction sequence to compensate for rapid temperature rise scenarios (such as high-power operation).

[0085] The LSTM model architecture includes: Input layer: 8 nodes – data from 4 temperature sensors (current time t-2, t-1, t) and RMS values ​​of 3-phase coil current (current time t); Hidden layer: 64 LSTM units with a time step of 3 (inputting data from the last 3 time steps to capture temperature change trends); Output layer: 1 node – predicting the thermal deformation angle Δθ_pred (unit: μrad) for the next 1ms; Activation function: linear activation for the input / output layer and tanh activation for the hidden layer.

[0086] Training data preparation includes: Sample collection: Temperature rise test in a temperature chamber, control coil current 0-5A (corresponding power consumption 0-24W), collect temperature (50Hz) and angle deviation (1kHz), and generate 5000 sets of samples (4000 training set and 1000 validation set); Data preprocessing: Temperature normalization (0-1), angle deviation standardization (mean 0, standard deviation 1), and use a sliding window (window size 3) to generate time series samples.

[0087] Model training and updates include: optimizer: Adam (learning rate 0.001), loss function MSE, training until validation set MSE < (0.5 μrad). 2 Incremental update: When the system is idle (e.g., when there is no angle command), the latest 100 sets of data are added to the training set, and the last 10 hidden layer units are retrained (cold start update, time <100ms).

[0088] Real-time prediction and compensation include: prediction frequency: 1kHz (synchronized with the control cycle), input of temperature / current data at the current and the previous two moments, output Δθpred; compensation logic: θcmd∗=θref∗-Δθpred (thermal expansion causes the actual rotor angle to lag behind the feedback value, and the compensation direction is opposite).

[0089] In some embodiments, it further includes: a safety monitoring module, used to monitor the three-phase coil current, the temperature at each critical position, and the difference between the command angle and the feedback angle in real time; when the monitored value exceeds a set threshold, it triggers a protection mechanism to shut down the drive and report fault information.

[0090] This embodiment monitors key parameters such as current, temperature, and angle deviation in real time, triggering hardware-level protection and fault reporting in case of anomalies to prevent system damage. Multi-parameter monitoring covers three categories of risk indicators: electrical, thermal, and kinematic. A tiered protection mechanism combines rapid hardware response (μs level) with software logic judgment. Fault diagnosis records fault codes for easy future maintenance.

[0091] The monitoring parameters and thresholds are shown in the table below:

[0092]

[0093] The overcurrent protection compares the current sampling signal with the 5A threshold voltage via a comparator (LM339), outputs a low level to the enable terminal (EN) of the driver chip, and directly turns off the MOSFET drive signal;

[0094] The over-temperature protection triggers a relay to disconnect the power input via a temperature sensor signal output through a hysteresis comparator (a backup battery maintains the fault record). After the fault is cleared, a "reset" command must be sent from the host computer, and the system will restart according to the process of "sensor self-test → hardware initialization → zero-point calibration" to prevent malfunctions.

[0095] In some embodiments, to address the differences in thermal characteristics under different operating modes (such as high-precision positioning mode and fast scanning mode), transfer learning technology is employed. Based on a pre-trained LSTM model, fine-tuning is performed using a small amount of target operating condition data to achieve rapid adaptation for cross-mode thermal deformation prediction. This solves the prediction delay problem of traditional models when switching operating conditions and improves the real-time performance of dynamic compensation.

[0096] The transfer learning architecture design includes: Source domain model: A general LSTM model is trained using historical full-condition data (including 10 typical operating modes) to extract common features of the heat-deformation relationship (such as temperature conduction delay and coil heating coupling law); Target domain fine-tuning: When the system switches to a new operating condition (such as entering "high-speed rotation mode" for the first time), real-time data (approximately 60,000 samples) for the first 10 minutes is collected, the parameters of the first two LSTM layers are frozen, and only the last fully connected layer is trained, enabling the model to quickly adapt to the thermal dynamic characteristics under specific operating conditions. Operating condition feature vector: Input the current control mode (classes 1-10), coil current waveform (frequency domain features), and temperature gradient (temperature difference between sensors), and identify the current operating condition category through a lightweight CNN (3 convolutional layers); Model storage and retrieval: The main control chip has 8 built-in lightweight sub-models (approximately 200KB each), corresponding to different priority operating conditions. The corresponding sub-model is loaded or the fine-tuning process is triggered based on the recognition result. After each change of operating conditions, the new operating condition data is added to the source domain dataset, and the global model is updated (incremental training) every 24 hours during the system's sleep period to avoid "catastrophic forgetting". The prediction latency is reduced from 15ms of traditional LSTM to 5ms after transfer learning (when the operating conditions are known), and the first adaptation time for unknown operating conditions is less than 3 minutes.

[0097] This invention relates to the field of laser precision positioning equipment technology, specifically a high-precision electromagnetic drive angle rotation system based on multi-layer PCB integrated design, which is suitable for scenarios requiring high-precision angle control such as laser scanning, optical image stabilization, and precision measurement.

[0098] By integrating the drive circuit, sensing circuit, control circuit, and electromagnetic coil array into a single multi-layer PCB, the mechanical connecting parts of the traditional separate design are eliminated, significantly reducing the system size, improving structural compactness and reliability, and avoiding angle errors caused by mechanical transmission backlash. The angle feedback system improves angle measurement accuracy to sub-arcsecond levels through quadruple frequency harmonic processing and Kalman filtering algorithms, effectively filtering out grating signal noise and rotor jitter. The control module employs field-oriented control (FOC) and space vector pulse width modulation (SVPWM), combined with speed feedforward compensation, to achieve improved dynamic response speed and torque ripple suppression.

[0099] The safety monitoring module monitors current, temperature, and angle deviation in real time, and quickly shuts down the drive when a threshold is triggered to avoid overcurrent, overheating, and runaway risks, thus improving the system's fault protection response speed. Based on mature PCB manufacturing processes, the electromagnetic coil array is integrated with the circuit, reducing assembly steps for independent components, lowering processing costs, and making it suitable for mass production.

[0100] In some embodiments, for cluster systems consisting of multiple devices, federated learning technology is employed to train security monitoring models locally on each device. Only the model parameters are uploaded to a central server for aggregation, achieving "model movement without data movement." This protects user data privacy while improving the generalization ability of fault prediction. It addresses the problem of insufficient fault data for single devices and is particularly suitable for early warning of rare faults (such as coil insulation aging).

[0101] The device-side (edge ​​node) includes: a local safety monitoring module that collects data such as current waveforms, temperature sequences, and angle deviations (100 sets per second) and labels abnormal data (such as waveforms 10 seconds before overcurrent); a local fault prediction model trained based on 1D-CNN (detecting 6 types of faults such as overtemperature and inter-turn short circuits) is sent to the server every 10 minutes (not raw data); the central server aggregates gradients from multiple devices to update the global model (FedAvg algorithm), and distributes the updated model (parameter size < 50KB) after every 50 aggregations; and maintains a global fault feature library, labels rare fault samples (such as bearing wear faults occurring in only 3 devices), and protects the data source through differential privacy technology (ε=0.5).

[0102] The specific model includes: CNN structure: 2 convolutional layers (3×3 kernels, stride 1) + global average pooling + fully connected layer (outputting 6 types of fault probabilities); Training strategy: the device only starts training during idle periods (such as at night), with CPU utilization <20%, and battery-powered devices can extend battery life through low power mode; Anomaly detection: when the local model predicts a fault probability >0.95, the local protection mechanism is immediately triggered, and an alert (including the anonymous device ID) is sent to the server.

[0103] In some embodiments, for extreme operating conditions that are difficult to reproduce in real systems (such as sudden power outages or complete sensor failure), Generative Adversarial Networks (GANs) are used to generate virtual extreme scenario data to enhance the robustness of the safety monitoring module. By training the discriminator to distinguish between real data and generated data, the generator is forced to learn extreme fault characteristics, thereby improving the system's fault tolerance under unknown faults.

[0104] The GAN model construction includes: Generator (G): Input random noise vector (100-dimensional), which generates sensor data under extreme conditions (such as a sudden current increase of 10A, a sudden temperature change of +50℃, and an angle feedback jump of ±10°) through a deconvolutional network, and the output dimension is consistent with the real sensor data (6-dimensional); Discriminator (D): 3-layer fully connected network, which determines whether the input data is a real sample or a generated sample. The loss function adopts WGAN-GP (Wasserstein distance + gradient penalty) to avoid mode collapse; Training data: The initial real data includes 2000 normal samples + 500 known fault samples. The generator's goal is to deceive the discriminator, and the discriminator's goal is to classify correctly.

[0105] The security monitoring model is enhanced by adding the generated extreme data (accounting for 30% of the training set) to the training set of the security monitoring module and retraining the fault detection classifier (such as SVM or lightweight CNN).

[0106] Real-time fault tolerance strategies include: triggering an "extreme operating condition mode" when sensor data is detected to exceed the historical normal range of ±3σ; switching to a backup control strategy trained on GAN-generated data (such as open-loop compensation relying only on current and temperature) to maintain a safe system shutdown rather than direct power failure.

[0107] The generator and discriminator are trained offline on the PC only. After training, key fault features (such as high-frequency harmonic components of the current waveform) are extracted and converted into rules that can be recognized by the embedded system (such as "the coil is short-circuited when the current rise rate is >5A / ms and the temperature is >90℃"). The response time under extreme conditions is <20μs, which is implemented by hardware comparators and logic circuits to avoid software delay.

[0108] It should be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of the application. It should be understood that when an element or layer is referred to as “on,” “adjacent to,” “connected to,” or “coupled to” other elements or layers, it may be directly on, adjacent to, connected to, or coupled to other elements or layers, or there may be intervening elements or layers. Conversely, when an element is referred to as “directly on,” “directly adjacent to,” “directly connected to,” or “directly coupled to” other elements or layers, there are no intervening elements or layers. It should be understood that although the terms first, second, third, etc., may be used to describe various elements, components, areas, layers, and / or portions, these elements, components, areas, layers, and / or portions should not be limited by these terms. These terms are merely used to distinguish one element, component, area, layer, or portion from another element, component, area, layer, or portion. Therefore, without departing from the teachings of this application, the first element, component, area, layer, or portion discussed below may be referred to as a second element, component, area, layer, or portion.

[0109] Spatial relation terms such as “below,” “under,” “below,” “under,” “above,” “above,” etc., are used herein for convenience of description to describe the relationship between one element or feature shown in the figure and other elements or features. It should be understood that, in addition to the orientation shown in the figure, spatial relation terms are intended to also include different orientations of the device in use and operation. For example, if the device in the figure is flipped, then the element or feature described as “below,” “under,” or “below” other elements or features will be oriented “above” other elements or features. Therefore, the exemplary terms “below” and “under” can include both above and below orientations. The device may be otherwise oriented (rotated 90 degrees or otherwise) and the spatial descriptive terms used herein will be interpreted accordingly.

[0110] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. When used herein, the singular forms “a,” “an,” and “the” are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising” and / or “including,” when used in this specification, identify the presence of the stated features, integers, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups. When used herein, the term “and / or” includes any and all combinations of the associated listed items.

[0111] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0112] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A laser precision angle rotation system constructed based on a multi-layer PCB, characterized in that, The application relates to a multi-layer PCB board integrated with a driving circuit, a sensing circuit and a control circuit, wherein an electromagnetic coil array is arranged on the multi-layer PCB board. A rotor is arranged in cooperation with the electromagnetic coil array, and a laser reflector and a grating are fixedly connected to the rotor. An angle feedback system comprises a fixedly arranged photoelectric sensor array, which is used for reading signals generated when the grating rotates, and the signals are sequentially subjected to amplification, filtering and shaping treatment to form a digital square wave signal, the digital square wave signal is subjected to four times frequency processing and position counting to obtain original angle data, and the original angle data is subjected to Kalman filtering treatment to output a high-precision angle value. A control module comprises an upper computer, a main controller and a driving chip, the upper computer is used for inputting instructions, the instructions comprise two modes of continuous trajectory and point-to-point and generate angle instruction sequences, the angle instruction sequences are transmitted to the main controller through a protocol, the main controller generates three-phase PWM signals according to the angle instruction sequences and the high-precision angle value, and sequentially subjects to position control loop calculation, speed feedforward compensation, magnetic field orientation control and space vector pulse width modulation treatment, and inputs the three-phase PWM signals to the driving chip, and the driving chip outputs three-phase current to drive the electromagnetic coil array; the electromagnetic coil array is integrated in a conductive layer of the multi-layer PCB board in a layered mode, electromagnetic coils of each layer are arranged in a ring shape at preset angle intervals, electromagnetic coils of adjacent layers are distributed in a staggered mode in a circumferential position, and the electromagnetic coil array is directly electrically connected with a driving circuit of the multi-layer PCB board, the electrical connection of coils of each layer is realized through conductive through holes of the multi-layer PCB board, and an embedded electromagnetic driving structure without mechanical transmission components is formed. An analog signal output by the photoelectric sensor array is first input to a preamplifier for gain-adjustable amplification treatment, the amplified signal is filtered through a band-pass filter with fixed center frequency and limited bandwidth to remove high-frequency and low-frequency noises, and the filtered signal is compared with a preset threshold voltage through a voltage comparator to convert the analog signal into a digital square wave signal with fixed high and low levels.

2. The laser precision angle rotation system constructed based on multi-layer PCB according to claim 1, characterized in that, The rotor is a center-symmetrical disc structure, a central shaft of the rotor coincides with a central axis of the electromagnetic coil array, the laser reflector is fixedly installed on an upper surface of the rotor and synchronously rotates with the rotor, the grating is concentrically fixed to an edge region of the rotor, and a scale line direction of the grating is consistent with a radial direction of the rotor, so that the grating can be read by the photoelectric sensor array when the grating rotates with the rotor.

3. The laser precision angle rotation system constructed based on multi-layer PCB according to claim 1, characterized in that, The digital square wave signal is divided into A-phase and B-phase signals with a phase difference of 90 degrees, the A-phase and B-phase signals are input to a programmable logic device for four times frequency processing, that is, a counting pulse is triggered at a rising edge and a falling edge of each signal period, the pulse signal after four times frequency processing is counted by a 32-bit counter, and original rotation angle data of the rotor is calculated according to preset grating line numbers and pulse counting results.

4. The laser precision angle rotation system constructed based on multi-layer PCB according to claim 1, characterized in that, The original angle data is subjected to Kalman filtering treatment to output a high-precision angle value, which comprises:

5. The laser precision angle rotation system constructed based on multi-layer PCB according to claim 1, characterized in that, ​ A state space model containing angle state and angular velocity state is established, the original angle data is input into the Kalman filter as measurement value, and the angle state is iteratively estimated through the prediction step and the update step; wherein, the prediction step predicts the current state based on the state of the previous time and the system dynamics model, and the update step corrects the predicted state combined with the current measurement value, and finally outputs the high-precision angle value after filtering out random noise and jitter.

6. The laser precision angle rotation system constructed based on multi-layer PCB according to claim 1, characterized in that, The main controller inputs the difference value between the angle instruction sequence and the high-precision angle value into the position PID controller, and calculates the speed instruction; the speed instruction is superimposed with the acceleration feedforward compensation amount and input into the field-oriented control module, the current in the three-phase static coordinate system is converted into the direct-axis and quadrature-axis current instructions in the rotating coordinate system through coordinate transformation, and the PI controller is used for closed-loop control of the direct-axis and quadrature-axis currents; The control output voltage vector generates three PWM signals through the space vector pulse width modulation algorithm, and the duty cycle of the PWM signal is determined according to the amplitude and phase of the voltage vector and input into the drive chip.

7. The laser precision angle rotation system constructed based on multi-layer PCB according to claim 1, characterized in that, Further comprising: A thermal management compensation system including temperature sensors distributed at key positions of the multi-layer PCB, the temperature sensors collect temperature data in real time and input into a thermal deformation model, the thermal deformation model predicts the thermal expansion trend of the structure combined with historical data and outputs a thermal deformation angle prediction value, and the thermal deformation angle prediction value is superimposed on the angle instruction sequence as feedforward compensation to realize real-time compensation.

8. The laser precision angle rotation system constructed based on multi-layer PCB according to claim 7, characterized in that, The main controller inputs the temperature data and the current coil current value into a pre-trained long short-term memory neural network model, and the model outputs a predicted thermal deformation angle value; the main controller superimposes the thermal deformation angle value as feedforward compensation value into the angle instruction sequence to realize real-time compensation of the angle deviation caused by rotor thermal expansion, and the system uses the newly added data to incrementally update the parameters of the long short-term memory neural network model during the idle period.

9. The laser precision angle rotation system constructed based on multi-layer PCB according to claim 1, characterized in that, Further comprising: A safety monitoring module for real-time monitoring of three-phase coil current, temperature at each key position, and difference between instruction angle and feedback angle, and triggering a protection mechanism to shut down the drive and report fault information when the monitored value exceeds the set threshold.

Images