Grating scale measurement coupling error compensation method and device

Abstract

The application discloses a grating line scale measurement coupling error compensation method and device, which comprises a measuring device, a length measuring laser module and an elastic device, a microscope objective is arranged on the measuring device, the measuring device is connected with the length measuring laser module through the elastic device, the measuring device comprises a sliding plate, a line scale and a measuring scale cylinder, the line scale is arranged on the measuring scale cylinder, one end of the measuring scale cylinder is connected with the length measuring laser module through the elastic device, and an L-shaped connecting rod is arranged on the measuring scale cylinder and connected with the length measuring laser module. The application realizes accurate compensation of coupling errors in the grating line scale measurement process through comprehensive analysis and modeling of various coupling error factors, and improves the measurement precision.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of linear encoder measurement, and in particular to a grating linear encoder measurement coupling error compensation method and device. BACKGROUND

[0002] In the field of modern precision measurement, grating linear encoders are widely used in high-precision measurement systems such as numerical control machine tools, coordinate measuring machines, and semiconductor manufacturing equipment due to their high precision and high resolution. They are key components for accurate measurement of length, displacement, and other parameters. However, during actual measurement, factors such as the complexity of the measurement environment (e.g., temperature changes, mechanical vibrations, installation errors, etc.), and the connection method between the measurement device and the measured object can cause measurement coupling errors. These coupling errors can cause the measurement results to deviate from the true value, severely affecting the measurement accuracy and reliability, and limiting the application of grating linear encoders in higher precision measurement scenarios. For example, in the process of high-precision semiconductor chip manufacturing, nanometer-level measurement errors can cause chip performance to decline or even be scrapped; in precision machine tool processing, coupling errors can make the dimensional accuracy of processed parts unable to meet design requirements. Currently, existing error compensation methods mostly compensate for single error factors, making it difficult to effectively solve the problem of complex errors caused by multi-factor coupling. Therefore, there is an urgent need for a more effective grating linear encoder measurement coupling error compensation method and system. SUMMARY

[0003] The purpose of the present application is to provide a grating linear encoder measurement coupling error compensation method.

[0004] To achieve the above purpose, the present application is implemented according to the following technical solutions:

[0005] The present application comprises the following steps:

[0006] The linear encoder is placed in the center position; two groups of laser position sensors and line image sensors are arranged on both sides of the linear encoder; the laser position sensors are used to detect the position information of the linear encoder in space, including the horizontal attitude α and the vertical attitude β, as well as the height h; the line image sensors are used to obtain the image information of the linear encoder lines, assisting in accurate installation and positioning;

[0007] After obtaining the accurate installation and positioning, the measurement data and related data of the preset grating are obtained, and the measurement data and the related data are preprocessed; the related data includes grating encoder attitude parameters, installation offset, grating encoder body temperature, base temperature, environmental temperature, temperature distribution along the length direction of the grating encoder, displacement, speed, vibration, jitter, moire fringe signal characteristics, interpolation subdivision nonlinear error, mechanical stress, motion direction load, timestamp; the measurement data includes grating linear encoder measurement data and actual measurement data;

[0008] The difference between the measurement data from the grating line scale and the actual measured data is used as the measurement error. Based on the relevant data, the measurement error is compensated by a neural network to obtain first compensation data. Feedback correction is performed based on the first compensation data to obtain second compensation data. The second compensation data is then output as the compensation result.

[0009] Further, the method for obtaining first compensation data by performing neural network compensation on the measurement error based on the relevant data includes:

[0010] The relevant data is input into the neural network compensation model, and the relevant data is input into the shared feature extraction layer. Common high-order features are extracted from multi-source sensor data through a fully connected network.

[0011] By automatically filtering out noise-related redundant residuals through adaptive soft thresholding, effective error characteristics are preserved. The expression is as follows:

[0012]

[0013] τ(x)=Sigmoid(Wx+b o )

[0014] in τ(x) is the feature scaling factor, τ(x) is the adaptive threshold, x is the input feature, y is the output feature, ReLU(·) is the rectified linear unit, W is the learnable weight matrix, and b o Let Wx+b be a learnable bias matrix, and let Sigmoid(·) be a sigmoid function. o When Wx+b approaches positive infinity, the adaptive threshold is set to 1. o When the value approaches negative infinity, values ​​less than the threshold are set to zero.

[0015] The spatiotemporal attention branch combines convolutional-recurrent neural networks with a self-attention mechanism to dynamically focus on key time-frequency regions of vibration signals and capture their time-frequency characteristics; the sensor attention branch is based on a multi-head cross-attention mechanism to weightedly fuse the feature contributions from multiple sensors.

[0016] Feature fusion integrates the outputs of the spatiotemporal attention branch and the sensor attention branch through channel splicing and weighted summation to form a global coupled error representation.

[0017] Embed thermal expansion and Abbe error into neural network predictions, and calculate the loss function:

[0018]

[0019] Where α is the tradeoff coefficient, and α is the coefficient of thermal expansion of the grating material. Let ΔL be the total loss function. pred The predicted displacement error value, is a physical constraint loss, is a smooth L1 loss, φ is a neural network weight, ||φ||2 is a ridge regression value of the neural network weight, β is a vibration sensitivity coefficient, ΔT is a temperature difference between the grating ruler and the base body, vib_rms is a root mean square value of a vibration signal, ||·||1 is a Manhattan norm;

[0020] When the loss function value is lower than 0.127, outputting a global coupling error of the neural network compensation model as the first compensation data.

[0021] Further, the method for obtaining the second compensation data through feedback correction according to the first compensation data comprises:

[0022] Taking the loss function value of the neural network compensation as a prediction error value, obtaining a measured error value, and calculating a residual error of the first compensation data:

[0023] e(k) = ΔL1(k) - ΔL 直实 (k)

[0024] Wherein e(k) is the residual error of the first compensation data, k is a current time step, ΔL 直实 (k) is a measured error value at the kth time step, and ΔL1(k) is a prediction error value of the neural network;

[0025] A dynamic adjustment of the second compensation data is performed by using an incremental proportional-integral-derivative controller, and the expression is:

[0026]

[0027] Wherein ΔL2(k) is the second compensation data, K p is a proportional term, K i is an integral term, and K d is a differential term, and e(k-1) is a residual error of the first compensation data at the k-1th time step;

[0028] A residual error reset threshold is set, and the proportional term, the integral term and the differential term are adjusted according to the dynamic error, and the expression is:

[0029]

[0030] Wherein λ is a decay coefficient, is a dynamically adjusted proportional term, is a dynamically adjusted integral term, is a dynamically adjusted differential term, T is a temperature, vib is a vibration intensity, f(T,vib) is a weighted function of the temperature and the vibration intensity, g(ΔT) is a correction function of a temperature change rate, ρ and μ are adjustment coefficients, vib_rms is a root mean square value of a vibration signal, and e th is the residual error reset threshold.

[0031] The second compensation data is calculated and outputted by using the adjusted proportional term, integral term and differential term.

[0032] In a second aspect, a grating scale measurement coupling error compensation device comprises a measurement device, a length measurement laser module and an elastic device, a microscope objective is arranged on the measurement device, the measurement device is connected with the length measurement laser module through the elastic device, the measurement device comprises a slide plate, a scale and a measurement scale cylinder, the scale is arranged on the measurement scale cylinder, one end of the measurement scale cylinder is connected with the length measurement laser module through the elastic device, and an L-shaped connecting rod is arranged on the measurement scale cylinder and connected with the length measurement laser module.

[0033] Further, the elastic device comprises a tension spring and a coupling rod, one end of the coupling rod is connected with the measurement scale cylinder, the other end of the coupling rod is connected with the length measurement laser module, and the tension spring is sleeved on the coupling rod.

[0034] Further, the length measurement laser module comprises a measurement mirror, a reference mirror, an interference measurement mirror and a light path support, the interference measurement mirror is arranged in the light path support, the measurement mirror and the reference mirror are sequentially arranged on a coupling interference light path of the interference measurement mirror, the tension spring of the elastic device is arranged on the light path support, the measurement mirror is fixed on the L-shaped connecting rod and located on the same horizontal axis as the measurement scale cylinder, and an adjustable support is arranged at the bottom of the measurement scale cylinder.

[0035] The present application has the following beneficial effects:

[0036] Compared with the prior art, the present application has the following technical effects:

[0037] The present application comprehensively analyzes and models various coupling error factors, accurately compensates the coupling error in the grating scale measurement process, greatly improves the measurement accuracy, and reduces the measurement error by more than one order of magnitude compared with the traditional measurement method. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 The present application is a grating scale measurement coupling error compensation method, and the step flow chart is shown in the figure;

[0039] Figure 2 The present application is a grating scale measurement coupling error compensation device in a specific embodiment, and the diagram is shown in the figure;

[0040] Figure 3 The present application is a precise installation and positioning diagram of the scale in a specific embodiment. Detailed Implementation

[0041] The present invention will be further described below through specific embodiments. The illustrative embodiments and descriptions herein are used to explain the present invention, but are not intended to limit the present invention.

[0042] The present invention discloses a method and apparatus for compensating coupling errors in the measurement of grating line scales, comprising the following steps:

[0043] like Figure 1 As shown, this embodiment includes the following steps:

[0044] The etched ruler is placed in the center position; two sets of laser position sensors and etched image sensors are arranged on both sides of the etched ruler; the laser position sensors are used to detect the position information of the etched ruler in space, including the lateral attitude α, the vertical attitude β and the height h, and the etched image sensors are used to acquire the image information of the etched lines of the etched ruler to assist in accurate installation and positioning.

[0045] After precise installation and positioning, measurement data and related data of the preset grating are obtained, and the measurement data and related data are preprocessed. The related data includes grating ruler attitude parameters, installation offset, grating ruler body temperature, substrate temperature, ambient temperature, temperature distribution along the length of the grating ruler, displacement, velocity, vibration, jitter, moiré fringe signal characteristics, interpolation subdivision nonlinearity error, mechanical stress, load in the motion direction, and timestamp. The measurement data includes grating line scale measurement data and actual measurement data.

[0046] The difference between the measurement data of the grating line scale and the measured data is used as the measurement error. The measurement error is compensated by a neural network based on the relevant data to obtain the first compensation data. The second compensation data is obtained by feedback correction based on the first compensation data. The second compensation data is output as the compensation result.

[0047] In actual assessment, such as Figure 2 As shown, in this embodiment, the microscope objective lens is the Leica MPlanApo50x / 0.95 objective lens from Germany. This objective lens has high resolution and clear imaging effect, and can accurately capture the scale information on the grating line scale to meet the high-precision measurement requirements. The slide plate adopts a linear rolling guide slide plate, model HIWIN HGH30CA, which has the characteristics of high precision, low friction and high rigidity, and can ensure the smooth movement of the measuring device during the measurement process. The line scale is a high-precision line scale selected from Renishaw's XL-80 laser interferometer from the UK, with an accuracy of ±0.5μm / m, which can provide an accurate measurement reference.

[0048] The measurement ruler cylinder is customized with aluminum alloy, with a size of 300 mm in length, 100 mm in width, and 80 mm in height. The aluminum alloy material has the advantages of light weight, high strength, and corrosion resistance, facilitating installation and use. The measurement mirror is a P / NFB19-A from Thorlabs Company, which has high reflectivity and low loss, ensuring stable transmission of laser signals. The reference mirror is a P / NFB19-R from Thorlabs Company, which is in the same series as the measurement mirror, ensuring consistent optical performance and improving measurement accuracy. The interference measurement mirror is a Model 1931-C from Newport Company, which has high sensitivity and stability, enabling accurate measurement of laser interference signals. The optical path support is customized with steel, with a size of 400 mm in length, 150 mm in width, and 200 mm in height. The steel material has good rigidity and stability, supporting various optical elements and ensuring stable optical path;

[0049] The tension spring is a stainless steel tension spring with a spring constant of 5 N / mm and a length of 50 mm, which has good elasticity and corrosion resistance, providing stable tension during measurement.

[0050] The coupling rod is a stainless steel coupling rod with a diameter of 8 mm and a length of 100 mm, ensuring the strength and rigidity of the coupling rod and the reliable connection between the measurement device and the length measurement laser module.

[0051] The L-shaped connecting rod is an aluminum alloy connecting rod with two right-angle edges of 150 mm and 100 mm in length, respectively, used to connect the measurement ruler cylinder and the length measurement laser module, ensuring that the measurement mirror and the measurement ruler cylinder are on the same horizontal axis.

[0052] The adjustable supports are four SBR16 adjustable supports installed at the bottom of the measurement ruler cylinder, each with an adjustable height range of 0-50 mm, facilitating horizontal adjustment of the measurement ruler cylinder.

[0053] The linear scale is fixed on the measurement ruler cylinder to ensure firm installation and clear graduation. The slide plate is installed on the measurement ruler cylinder to ensure smooth sliding. The microscope objective is installed on the slide plate, and the position and angle of the objective are adjusted to clearly observe the graduation on the linear scale.

[0054] The interference measurement mirror is installed inside the optical path support according to the optical path design requirements. The measurement mirror and the reference mirror are installed in sequence on the coupling interference light path of the interference measurement mirror, ensuring accurate installation position and consistency with the optical axis of the interference measurement mirror.

[0055] The measurement mirror is fixed on the L-shaped connecting rod, and the position and angle of the L-shaped connecting rod are adjusted to ensure that the measurement mirror and the measurement ruler cylinder are on the same horizontal axis.

[0056] One end of the coupling rod is connected with the measuring scale cylinder, and the other end is connected with the length measuring laser module on the light path support, ensuring firm connection; the tension spring is sleeved on the coupling rod, and the extension length of the tension spring is adjusted to provide appropriate tension to ensure the elastic connection between the measuring device and the length measuring laser module; four adjustable supports are installed at the bottom of the measuring scale cylinder, and the height of the adjustable supports is adjusted to keep the measuring scale cylinder in a horizontal state, ensuring the accuracy of the measurement;

[0057] Turn on the power supply of the length measuring laser module, adjust the laser light path, and ensure that the laser can accurately irradiate the measuring mirror and the reference mirror and produce a stable interference signal;

[0058] Move the slide plate of the measuring device, observe the line scale scale under the microscope objective and the measurement data of the length measuring laser module, and check the cooperation between the measuring device and the length measuring laser module;

[0059] According to the measurement data and actual measurement requirements, the tension of the tension spring of the elastic device and the height of the adjustable support are fine-tuned to calibrate and optimize the device, ensuring the accuracy and stability of the measurement results;

[0060] The line scale is in the center position; laser position sensors (1-4) and line image sensors (1 and 2) are arranged on both sides of the line scale; the laser position sensor is used to detect the position information of the line scale in space, including the horizontal attitude α and the vertical attitude β, and the height h; the horizontal attitude α reflects the inclination of the line scale in the horizontal direction, the vertical attitude β reflects the inclination in the vertical direction, and the height h is the distance of the line scale relative to the reference surface; the line image sensor is used to obtain the image information of the line scale, which assists in accurate installation and positioning;

[0061] The data transmitted by the sensors of the mechanical and electrical parts is received and processed by the computer; Figure 3 The example display software obtains the horizontal attitude α = 0.001°, the vertical attitude β = 0.001°, and the height h = 150.000mm, which can be used to judge whether the installation attitude and position of the line scale meet the requirements, and then guide the adjustment to realize the accurate automatic installation and positioning of the line scale;

[0062] The first compensation data is 0.107, and the second compensation data is 0.038;

[0063] In the size measurement of parts in a high-precision machining enterprise, the part to be measured is placed on the workbench of the measuring device, the microscope objective is aligned with the measurement position of the part by moving the slide, and the initial scale on the wire gauge is read; then, the length measuring laser module is started to measure, the laser signal is reflected by the measuring mirror and interferes with the reference mirror, the interference measuring mirror converts the interference signal into an electrical signal and transmits it to the data processing system, and the data processing system calculates the actual size of the part; during the measurement process, the elastic device can effectively compensate the coupling error between the measuring device and the length measuring laser module, ensuring the accuracy of the measurement results; through multiple measurement verification, the measurement error of the device is controlled within ±1μm, meeting the requirements of the enterprise for high-precision part size measurement.

[0064] In the embodiment, the method for neural network compensation of the measurement error based on the correlation data to obtain first compensation data comprises:

[0065] The correlation data is input into the neural network compensation model, the correlation data is input into the shared feature extraction layer, and common high-order features are extracted from the multi-source sensor data through the full connection network;

[0066] Redundant residuals related to noise are automatically filtered out through adaptive soft thresholding, and effective error features are retained, and the expression is:

[0067]

[0068] τ(x)=Sigmoid(Wx+b o )

[0069] wherein is a feature scaling coefficient, τ(x) is an adaptive threshold, x is an input feature, y is an output feature, ReLU(·) is a rectified linear unit, W is a learnable weight matrix, b o is a learnable bias matrix, Sigmoid(·) is a sigmoid function, when Wx+b o tends to positive infinity, the adaptive threshold is set to 1, and when Wx+b o tends to negative infinity, the threshold less than the threshold is set to zero;

[0070] The space-time attention branch combines convolution-recurrent neural network and self-attention mechanism to dynamically focus on the key time-frequency area of the vibration signal and capture the time-frequency features of the vibration signal; the sensor attention branch is based on a multi-head cross-attention mechanism to weight and fuse the feature contributions of multiple sensors;

[0071] Feature fusion integrates the outputs of the space-time attention branch and the sensor attention branch through channel splicing and weighted summation to form a global coupling error representation;

[0072] Embedding thermal expansion, Abbe error in neural network prediction, calculating loss function:

[0073]

[0074] Wherein α is the trade-off coefficient, α is the thermal expansion coefficient of the grating material, is the total loss function, ΔL pred is the predicted displacement error value, is the physical constraint loss, is the smooth L1 loss, φ is the neural network weight, ||φ||2 is the ridge regression value of the neural network weight, β is the vibration sensitivity coefficient, ΔT is the temperature difference between the grating ruler and the base, vib_rms is the root mean square value of the vibration signal, ‖·‖1 is the Manhattan norm;

[0075] When the loss function value is lower than 0.127, the global coupling error output of the neural network compensation model is the first compensation data.

[0076] In this embodiment, the method for obtaining the second compensation data according to the feedback correction of the first compensation data comprises:

[0077] The loss function value of the neural network compensation is taken as the predicted error value, the measured error value is obtained, and the residual error of the first compensation data is calculated:

[0078] e(k) = ΔL1(k) - ΔL 直实 (k)

[0079] Wherein e(k) is the residual error of the first compensation data, k is the current time step, ΔL 直实 (k) is the measured error value at the kth time step, and ΔL1(k) is the predicted error value of the neural network;

[0080] The incremental proportional-integral-derivative controller is used to dynamically adjust the second compensation data, and the expression is:

[0081]

[0082] Wherein ΔL2(k) is the second compensation data, K p is the proportional term, K i is the integral term, and K d is the differential term, and e(k-1) is the residual error of the first compensation data at the k-1th time step.

[0083] The residual error reset threshold is set, the proportional term, the integral term and the differential term are adjusted according to the dynamic error, and the expression is:

[0084]

[0085] Wherein λ is the attenuation coefficient, a dynamic adjusted proportional term, a dynamic adjusted integral term, a dynamic adjusted differential term, T is temperature, vib is vibration intensity, f(T, vib) is a weighting function of temperature and vibration intensity, g(ΔT) is a correction function of temperature change rate, ρ, μ are adjustment coefficients, vib_rms is the root mean square value of the vibration signal, e th is a residual reset threshold value;

[0086] The adjusted proportional term, integral term and differential term are used to calculate and output the second compensation data.

[0087] In the embodiment, the grating line scale measurement coupling error compensation device comprises a measuring device, a length measuring laser module and an elastic device, a microscope objective is arranged on the measuring device, the measuring device is connected with the length measuring laser module through the elastic device, the measuring device comprises a slide plate, a line scale and a measuring scale cylinder, the line scale is arranged on the measuring scale cylinder, one end of the measuring scale cylinder is connected with the length measuring laser module through the elastic device, and an L-shaped connecting rod is arranged on the measuring scale cylinder and connected with the length measuring laser module.

[0088] In the embodiment, the grating line scale measurement coupling error compensation device, the elastic device comprises a tension spring and a coupling rod, one end of the coupling rod is connected with the measuring scale cylinder, the other end of the coupling rod is connected with the length measuring laser module, and the tension spring is sleeved on the coupling rod.

[0089] In the embodiment, the grating line scale measurement coupling error compensation device, the length measuring laser module comprises a measuring mirror, a reference mirror, an interference measuring mirror and a light path support, the interference measuring mirror is arranged in the light path support, the measuring mirror and the reference mirror are sequentially arranged on the coupling interference light path of the interference measuring mirror, the tension spring of the elastic device is arranged on the light path support, the measuring mirror is fixed on the L-shaped connecting rod and located on the same horizontal axis as the measuring scale cylinder, and an adjustable support is arranged at the bottom of the measuring scale cylinder.

[0090] The above only describes the preferred embodiments of the present application and is not intended to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for compensating coupling error in grating line scale measurement, characterized in that, Includes the following steps: The etched ruler is positioned at the center. Two sets of laser position sensors and etched line image sensors are arranged on both sides of the etched ruler. The laser position sensors are used to detect the position information of the etched ruler in space, including the lateral attitude α, the vertical attitude β, and the height h. The etched line image sensors are used to acquire image information of the etched lines of the etched ruler to assist in precise installation and positioning. The lateral attitude α reflects the tilt of the etched ruler in the horizontal direction, the vertical attitude β reflects the tilt in the vertical direction, and the height h is the distance of the etched ruler relative to the reference plane. After precise installation and positioning, measurement data and related data of the preset grating are obtained, and the measurement data and related data are preprocessed. The related data includes grating ruler attitude parameters, installation offset, grating ruler body temperature, substrate temperature, ambient temperature, temperature distribution along the length of the grating ruler, displacement, velocity, vibration, jitter, moiré fringe signal characteristics, interpolation subdivision nonlinearity error, mechanical stress, load in the motion direction, and timestamp. The measurement data includes grating line scale measurement data and actual measurement data. The difference between the measurement data of the grating line scale and the measured data is used as the measurement error. The measurement error is compensated by a neural network based on the relevant data to obtain the first compensation data. The second compensation data is obtained by feedback correction based on the first compensation data. The second compensation data is output as the compensation result. include: The relevant data is input into the neural network compensation model, and the relevant data is input into the shared feature extraction layer. Common high-order features are extracted from multi-source sensor data through a fully connected network. By automatically filtering out noise-related redundant residuals through adaptive soft thresholding, effective error characteristics are preserved. The expression is as follows: in For feature scaling factor, For adaptive threshold, For input features, For output features, It is a rectifier linear unit. The weight matrix is ​​a learnable matrix. The bias matrix is ​​a learnable matrix. For a sigmoid function, when When it approaches positive infinity, the adaptive threshold is set to 1. When the value approaches negative infinity, values ​​less than the threshold are set to zero. The spatiotemporal attention branch combines convolutional-recurrent neural networks with a self-attention mechanism to dynamically focus on key time-frequency regions of vibration signals and capture their time-frequency characteristics; the sensor attention branch is based on a multi-head cross-attention mechanism to weightedly fuse the feature contributions from multiple sensors. Feature fusion integrates the outputs of the spatiotemporal attention branch and the sensor attention branch through channel splicing and weighted summation to form a global coupled error representation. Embed thermal expansion and Abbe error into neural network predictions, and calculate the loss function: in As a weighting factor, The coefficient of thermal expansion of the grating material is... For the total loss function, The predicted displacement error value, For physical constraint loss, To smooth out L1 loss, For neural network weights, The ridge regression values ​​are the weights of the neural network. The vibration sensitivity coefficient, The temperature difference between the grating ruler and the substrate. The root mean square value of the vibration signal. It is the Manhattan norm; When the loss function value is below 0.127, the global coupling error of the neural network compensation model is output as the first compensation data. Using the loss function value of the neural network compensation as the prediction error value, the measured error value is obtained, and the residual of the first compensation data is calculated: in The residual of the first compensated data, k is the current time step. Let be the measured error value at the k-th time step. This represents the prediction error value of the neural network. The second compensation data is dynamically adjusted using an incremental proportional-integral-derivative controller, expressed as: in This is the second compensation data. For the proportion term, For integration, For differential terms, The residual of the first compensated data at the (k-1)th time step; Set a residual reset threshold, and adjust the proportional, integral, and derivative terms according to the dynamic error. The expression is: in The attenuation coefficient is... This is the dynamically adjusted proportion. This is the dynamically adjusted integral term. The differential term is dynamically adjusted, where T is temperature. For vibration intensity, It is a weighted function of temperature and vibration intensity. This is a correction function for the rate of temperature change. , For adjustment coefficients, The root mean square value of the vibration signal. This is the residual reset threshold; The second compensation data is calculated and output using the adjusted proportional, integral, and differential terms.

Images