An offline time calibration accuracy optimization method and system for a smart watch

Abstract

This invention discloses a method and system for optimizing offline time calibration accuracy in smartwatches, relating to the field of offline clock calibration technology. The method includes: deploying an offline calibration mode in the smartwatch, employing a combined setup of first-order dual-crystal differential common-mode compensation and second-order temperature gradient feedforward transient compensation. The first-order calibration mode eliminates common-mode temperature drift through differential mixing of the main and auxiliary crystal oscillators, generating a reference signal; the second-order calibration mode utilizes an analog differential architecture to obtain a temperature-voltage gradient signal, which is superimposed on the varactor diode bias terminal to achieve transient compensation. The two modes are combined to complete offline time calibration management. This invention solves the technical problem of timing deviations caused by changes in ambient temperature and actual operating conditions in smartwatches during offline operation, achieving the technical effect of improving the time calibration accuracy of smartwatches in offline mode while ensuring their long-term timing stability.

Description

Technical Field

[0001] This invention relates to the field of offline clock calibration technology, and in particular to a method and system for optimizing the offline time calibration accuracy of a smartwatch. Background Technology

[0002] Electronic timers are widely used in various electronic devices such as smart wearables. Their offline time calibration accuracy is crucial for timing accuracy, and clock drift suppression is a core performance indicator. Current technologies mostly use ordinary crystal oscillators as the timing core of electronic timers. While this meets basic timing requirements in online or stable environments with network or Bluetooth connections, the increasing demands for timing accuracy reveal many limitations of traditional technologies in offline applications. Electronic timers are susceptible to temperature drift and operating condition disturbances when operating offline. Traditional calibration methods cannot dynamically correct the clock frequency, leading to continuous accumulation of timing errors and insufficient calibration accuracy, making it difficult to meet the requirements for high-precision timing and long-term timing stability in offline conditions. Summary of the Invention

[0003] This application provides a method and system for optimizing the offline time calibration accuracy of smartwatches, which solves the technical problem that smartwatches are susceptible to timing deviations caused by changes in ambient temperature and disturbances in the actual operating conditions of the device when working offline.

[0004] The first aspect of this application provides a method for optimizing the offline time calibration accuracy of a smartwatch. The method includes: deploying an offline calibration mode in the smartwatch, wherein the offline calibration mode is jointly deployed using first-order dual-crystal differential common-mode compensation and second-order temperature gradient feedforward transient compensation; wherein the calibration steps based on the offline calibration mode include: constructing a first-order calibration mode using a dual-crystal differential structure, wherein the first-order calibration mode eliminates common-mode temperature drift through mixing and differential processing of the main crystal oscillator and auxiliary crystal oscillator to generate a pure reference signal; constructing a second-order calibration mode using an analog differential architecture, wherein the second-order calibration mode generates a gradient voltage signal by performing time-domain differentiation on the temperature voltage signal, and superimposes it onto the bias voltage terminal of the varactor diode of the main crystal oscillator to achieve temperature gradient feedforward transient compensation; and fusing the first-order calibration mode and the second-order calibration mode to perform offline time calibration management on the smartwatch.

[0005] A second aspect of this application provides an offline time calibration accuracy optimization system for a smartwatch. The system includes: an offline calibration mode deployment module for deploying an offline calibration mode in the smartwatch, wherein the offline calibration mode is jointly deployed using first-order dual-crystal differential common-mode compensation and second-order temperature gradient feedforward transient compensation; wherein the calibration steps based on the offline calibration mode include: a first-order calibration mode construction module for constructing a first-order calibration mode using a dual-crystal differential structure, wherein the first-order calibration mode eliminates common-mode temperature drift through mixing and differential processing of the main crystal oscillator and auxiliary crystal oscillator to generate a pure reference signal; a second-order calibration mode construction module for constructing a second-order calibration mode using an analog differential architecture, wherein the second-order calibration mode generates a gradient voltage signal by performing time-domain differentiation on the temperature voltage signal, which is then superimposed on the bias voltage terminal of the varactor diode of the main crystal oscillator to achieve temperature gradient feedforward transient compensation; and an offline time calibration execution module for fusing the first-order calibration mode and the second-order calibration mode to perform offline time calibration management on the smartwatch.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] This application deploys a combined first-order and second-order offline calibration mode for smartwatches. Through dual-crystal differential common-mode processing and temperature gradient feedforward transient compensation, an offline absolute clock reference is generated, and dynamic reverse compensation of the crystal oscillator frequency is achieved. Combined with the individual error curve of the crystal oscillator, residual phase error compensation is performed, and calibration parameters are stored and quickly restored upon wake-up. This effectively suppresses offline timing errors caused by temperature drift and operating condition disturbances, significantly improving the offline time calibration accuracy and long-term timing stability of smartwatches. It achieves the technical effect of improving the time calibration accuracy of smartwatches in offline mode while ensuring their long-term timing stability. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart illustrating a method for optimizing the offline time calibration accuracy of a smartwatch, as provided in an embodiment of this application.

[0010] Figure 2 This is a schematic diagram of the structure of an offline time calibration accuracy optimization system for a smartwatch provided in an embodiment of this application.

[0011] Figure labeling: Offline calibration mode deployment module 1, first-order calibration mode construction module 2, second-order calibration mode construction module 3, offline time calibration execution module 4. Detailed Implementation

[0012] This application provides a method and system for optimizing the offline time calibration accuracy of smartwatches, which solves the technical problem that smartwatches are susceptible to timing deviations caused by changes in ambient temperature and disturbances in the actual operating conditions of the device when working offline.

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0014] It should be noted that the terms "first," "second," etc., in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices.

[0015] Example 1, as Figure 1 As shown, a method for optimizing offline time calibration accuracy in a smartwatch includes:

[0016] An offline calibration mode is deployed in a smartwatch, wherein the offline calibration mode is jointly deployed with first-order dual-crystal differential common-mode compensation and second-order temperature gradient feedforward transient compensation.

[0017] In this embodiment, the offline calibration mode is a local automatic time calibration mode built into the smartwatch when it is offline and there is no network or external synchronization signal such as Bluetooth. It dynamically corrects the clock frequency and suppresses various timing errors through a combination of first-order and second-order calibration methods.

[0018] Specifically, the deployment of the offline calibration mode for smartwatches involves deploying a first-order calibration mode based on a dual-crystal differential structure and using first-order dual-crystal differential common-mode compensation; and deploying a second-order calibration mode based on an analog differential architecture and using main crystal varactor diode modulation under temperature gradient feedforward transient compensation. The two calibration modes are then integrated to complete the overall deployment. This step will be explained in detail later.

[0019] The calibration steps based on the offline calibration mode include:

[0020] A first-order calibration mode is constructed using a dual-crystal differential structure. The first-order calibration mode eliminates common-mode temperature drift by mixing and differential processing of the main crystal oscillator and the auxiliary crystal oscillator, thereby generating a pure reference signal.

[0021] Optionally, the first-order calibration mode uses a temperature-compensated crystal oscillator as the main crystal oscillator and a common crystal oscillator with the opposite temperature coefficient as the auxiliary crystal oscillator to form a dual-crystal differential structure. In the first offline time interval, the temperature-compensated reference frequency signal and the uncompensated local oscillator frequency signal are output. After mixing and differential processing of the two signals, a pure reference signal is obtained. The construction process of the above first-order calibration mode will be explained in detail in the following content.

[0022] The calibration steps based on the offline calibration mode include:

[0023] A second-order calibration mode is constructed using an analog differential architecture. The second-order calibration mode generates a gradient voltage signal by performing time-domain differentiation on the temperature-voltage signal, and then superimposes it onto the bias voltage terminal of the main crystal oscillator varactor diode to achieve transient compensation of the temperature gradient feedforward.

[0024] Optionally, in the second-order calibration mode, the temperature voltage signal output by the temperature sensor is input into an analog differential architecture composed of an operational amplifier and an RC network for time-domain differentiation to obtain a gradient voltage signal proportional to the rate of temperature change. This signal is amplified by a variable gain amplifier and superimposed on the bias voltage terminal of the varactor diode of the main crystal oscillator, so that the output frequency of the crystal oscillator can achieve instantaneous reverse compensation with the rate of temperature change. The construction process of the above-mentioned second-order calibration mode will also be explained in detail in the following content.

[0025] By combining the first-order calibration mode and the second-order calibration mode, offline time calibration management can be performed on smartwatches.

[0026] In one embodiment of this application, after the smartwatch enters offline mode, it simultaneously initiates a first-order calibration mode and a second-order calibration mode. The two modes run independently and in parallel, with no specific execution order. The first-order calibration mode calculates the steady-state frequency deviation of the main crystal oscillator in real time by comparing the pulses of the main and auxiliary crystal oscillators, and outputs the corresponding compensation amount for the steady-state temperature drift of the main crystal oscillator. The second-order calibration mode calculates the transient frequency hysteresis error of the main crystal oscillator in real time by extracting the rate of temperature change, and outputs the corresponding compensation amount for the thermal inertia hysteresis of the crystal oscillator.

[0027] Next, the steady-state compensation output from the first-order calibration mode and the transient compensation output from the second-order calibration mode are fused together using a passive adder circuit at the bias voltage terminal of the varactor diode of the main crystal oscillator, jointly adjusting the oscillation frequency of the main crystal oscillator. The fused clock signal is input to a digital phase-locked loop, and after phase locking, a stable system clock reference is output to provide an offline timing signal for the smartwatch.

[0028] The entire offline time calibration management process runs continuously. As long as the smartwatch is in working condition, the first-order calibration mode and the second-order calibration mode will collect data and perform compensation operations simultaneously to correct the frequency error of the main crystal oscillator in real time and ensure the accuracy of offline timekeeping.

[0029] Furthermore, the method provided in this application embodiment includes:

[0030] Based on the dual-crystal differential structure, a first-order calibration mode is deployed using first-order dual-crystal differential common-mode compensation. Based on the analog differential architecture, a second-order calibration mode is deployed using main crystal varactor diode modulation under temperature gradient feedforward transient compensation. The first-order and second-order calibration modes are then combined to deploy an offline calibration mode for the smartwatch.

[0031] Specifically, when deploying the offline calibration mode in a smartwatch, the process begins by using a dual-crystal differential structure as a baseline. A common crystal oscillator with an opposite temperature coefficient is selected as the auxiliary oscillator, and this is combined with a temperature-compensated crystal oscillator to construct the dual-crystal differential structure. First-order dual-crystal differential common-mode compensation is then used as the compensation method to deploy the first-order calibration mode. Next, using an analog differential architecture as a baseline, the main crystal oscillator varactor diode modulation under temperature gradient feedforward transient compensation is used as the compensation method to deploy the second-order calibration mode. The deployment steps for the first and second-order calibration modes will be explained accordingly in subsequent sections.

[0032] Next, the first-order calibration mode and the second-order calibration mode are integrated. The first-order calibration mode completes the compensation for common-mode temperature drift and generates a stable initial clock reference. The second-order calibration mode captures the temperature change rate in real time and dynamically corrects the transient errors caused by rapid temperature changes. The initial clock reference output by the first-order calibration mode and the correction results of the second-order calibration mode are integrated to form a jointly calibrated clock reference. This completes the overall deployment of the offline calibration mode, enabling the offline calibration mode to operate in a combined form of first-order dual-crystal differential common-mode compensation and second-order temperature gradient feedforward transient compensation.

[0033] Furthermore, the method provided in this application embodiment includes:

[0034] A temperature-compensated crystal oscillator (TCC) is used as the main crystal oscillator, and a conventional crystal oscillator with an opposite temperature coefficient to the main crystal oscillator is used as the auxiliary crystal oscillator, forming a dual-crystal differential structure. For the first offline time interval, the main crystal oscillator in the dual-crystal differential structure outputs a reference frequency signal, and the auxiliary crystal oscillator outputs a local oscillator frequency signal. The reference frequency signal is a temperature-compensated reference frequency, and the local oscillator frequency signal is an uncompensated auxiliary frequency. The reference frequency signal and the local oscillator frequency signal are mixed and differentially processed to obtain a pure reference signal.

[0035] Optionally, the main crystal oscillator can be a surface-mount temperature-compensated crystal oscillator with the nominal frequency, for example, a nominal frequency of 32.768kHz, a negative temperature coefficient, and a load capacitance of 12.5pF. An industrial-grade surface-mount crystal oscillator, model TCXO-32768-0.5ppm, could be used. The auxiliary crystal oscillator is a common quartz crystal oscillator with the same nominal frequency, a positive temperature coefficient that is opposite to that of the main crystal oscillator, and the same load capacitance as the main crystal oscillator. For example, a surface-mount quartz crystal oscillator, model XTAL-32768-30ppm, could be used. The main and auxiliary crystal oscillators together form a dual-crystal differential structure, utilizing the opposite temperature coefficients of the two oscillators to offset the frequency error caused by common-mode temperature drift.

[0036] Both the main crystal oscillator and the auxiliary crystal oscillator are connected to the smartwatch's system power supply, for example, 3.3V. Their grounding terminals are connected to the system ground plane. The output of the main crystal oscillator is connected to the first input of the four-quadrant analog multiplier after a 100Ω current-limiting resistor in series. The output of the auxiliary crystal oscillator is connected to the second input of the same analog multiplier after a 100Ω current-limiting resistor in series. Both signal input impedances are matched to 50Ω to avoid signal reflection and interference. A low-power four-quadrant multiplier with a single power supply is selected, and its input / output voltage range is consistent with the system power supply, covering 0V to 3.3V. For example, an AD835 analog multiplier can be used.

[0037] When the smartwatch detects no cellular network or Bluetooth connection for more than 30 seconds, it automatically enters offline calibration mode. During this time, both the main and auxiliary crystal oscillators are powered on and enter a stable operating state. The first offline time interval is defined as the entire offline duration after the smartwatch enters offline calibration mode. The first-order calibration mode runs continuously within this interval without interruption or switching. For the first offline time interval, the main crystal oscillator uses its internally integrated temperature compensation circuit to correct its output frequency in real time, outputting a temperature-compensated 32.768kHz reference frequency signal. The auxiliary crystal oscillator directly outputs an uncompensated 32.768kHz local oscillator frequency signal. Both signals are simultaneously output to the analog multiplier.

[0038] Furthermore, the first-order calibration mode and the subsequent second-order calibration mode operate in parallel and independently. They start and stop simultaneously, each completing its corresponding error compensation task, without any sequential dependency. The first-order calibration mode continuously processes the frequency signals output by the main and auxiliary crystal oscillators to compensate for steady-state common-mode temperature drift, while the second-order calibration mode synchronously acquires and processes temperature signals to compensate for transient temperature gradient errors. The outputs of both are ultimately merged into a unified calibrated clock reference.

[0039] The dual crystal oscillators are arranged adjacent to each other, with a center-to-center distance of no more than 5mm, and are kept away from heat-generating components such as the CPU and power management chip inside the smartwatch to ensure that the ambient temperature of the two is consistent, with a deviation of no more than 0.2℃. The crystal oscillator output signal lines adopt the shortest wiring design, with a length of no more than 10mm. Grounded copper foil is laid on both sides of the signal lines for shielding. The analog circuit area and the digital circuit area are separated by a grounding isolation strip to avoid electromagnetic interference affecting signal quality.

[0040] In addition, the initial frequency difference between the main and auxiliary crystal oscillators needs to be calibrated before the smartwatch leaves the factory. Specifically, the device is placed in a standard constant temperature environment of 25°C for 30 minutes, the output frequencies of the main and auxiliary crystal oscillators are collected and the initial difference is calculated. The difference is stored in a non-volatile register as the initial calibration value for subsequent differential processing.

[0041] Finally, the reference frequency signal and the local oscillator frequency signal are mixed and differentially processed. This involves inputting both signals into a multiplier to obtain a product signal containing the sum frequency and the difference frequency. The sum frequency component is then filtered out using a low-pass filter to obtain a differential signal reflecting the frequency difference between the main and auxiliary crystal oscillators. After obtaining the differential signal, peak detection and amplitude normalization are performed to obtain a clean reference signal. This signal is then converted from analog to digital and its frequency and phase are locked using a digital phase-locked loop (PLL), serving as the absolute clock reference in offline mode. These steps will be explained in detail later.

[0042] By constructing a differential structure using dual crystal oscillators with opposite temperature coefficients and performing common-mode compensation, the steady-state common-mode temperature drift of the smartwatch in offline mode is effectively eliminated, providing a stable and reliable initial reference signal for high-precision offline time calibration.

[0043] Furthermore, the method provided in this application embodiment includes:

[0044] The reference frequency signal and the local oscillator frequency signal are input into a multiplier to obtain a product signal containing the sum frequency and the difference frequency; the sum frequency component of the product signal is filtered out according to a low-pass filter to obtain a difference signal, wherein the difference signal reflects the frequency difference between the main crystal oscillator and the auxiliary crystal oscillator.

[0045] Specifically, the reference frequency signal and the local oscillator frequency signal are first input into a four-quadrant analog multiplier. The four-quadrant analog multiplier performs a multiplication operation on the two input signals, generating a product signal containing a sum frequency component and a difference frequency component. The temperature coefficients of the main crystal oscillator and the auxiliary crystal oscillator exhibit inverse characteristics. When the ambient temperature changes slowly, the output frequency of the main crystal oscillator decreases with increasing temperature, while the output frequency of the auxiliary crystal oscillator increases with increasing temperature. The frequency drifts of the two are similar in magnitude but opposite in direction, belonging to the common-mode temperature drift component. This common-mode temperature drift component cancels out each other in the difference frequency component output by the four-quadrant analog multiplier, retaining only the inherent frequency difference between the two signals.

[0046] The product signal is then filtered using a low-pass filter. A low-pass filter with a cutoff frequency between the maximum value of the difference frequency component and the minimum value of the sum frequency component is selected to filter out the high-frequency sum frequency component, retaining only the low-frequency difference frequency component, thus obtaining the differential signal. This differential signal accurately reflects the frequency difference between the main crystal oscillator and the auxiliary crystal oscillator, and has eliminated the steady-state common-mode temperature drift caused by slow changes in ambient temperature.

[0047] The above processing can only effectively eliminate steady-state temperature drift, i.e., the common-mode frequency deviation caused by slow changes in ambient temperature. When the ambient temperature changes rapidly, due to the thermal inertia of the crystal itself, there is a slight difference in the temperature response speed between the main crystal oscillator and the auxiliary crystal oscillator, which causes the frequency drift of the two to be unable to be completely synchronized and canceled out. A certain amount of transient temperature error will still remain in the differential signal, and this error will be specifically compensated in the subsequent second-order calibration mode.

[0048] Through the above mixing and differential processing steps, the steady-state common-mode temperature drift of the smartwatch in offline mode is effectively eliminated, providing a basic signal for subsequent high-precision offline time calibration.

[0049] Furthermore, the method provided in this application embodiment includes:

[0050] The differential signal is subjected to peak detection and amplitude normalization to determine the normalized pure reference signal; the pure reference signal is then converted from analog to digital and entered into a digital phase-locked loop for frequency and phase locking, serving as the absolute clock reference in offline mode.

[0051] In this embodiment, the digital phase-locked loop is a closed-loop feedback module based on digital circuits, which can keep its output signal synchronized with the frequency and phase of the input signal through phase comparison and frequency adjustment.

[0052] Specifically, the differential signal is first processed using peak detection. A diode peak detection method is employed, where the differential signal is input into a peak detection circuit composed of diodes, resistors, and capacitors. The peak envelope of the differential signal is extracted, yielding a DC voltage signal proportional to the differential signal amplitude. Next, the DC voltage signal obtained from peak detection undergoes linear amplitude normalization. This is achieved by dividing the DC voltage signal by the standard amplitude calibrated at the equipment's factory, mapping the signal amplitude uniformly to the standard range of 0 to 1, resulting in a normalized, pure reference signal. This processing eliminates amplitude fluctuations in the differential signal caused by individual differences between different batches of main and auxiliary crystal oscillators, ensuring consistency and stability in subsequent signal processing.

[0053] The clean reference signal is then subjected to analog-to-digital conversion (ADC) using a successive approximation ADC. This converts the analog reference signal into a corresponding digital signal, with the sampling frequency set to eight times the frequency of the clean reference signal to ensure the accuracy and integrity of the signal sampling. The converted digital signal is then input into a digital phase-locked loop (PLL). The PLL compares the phase difference between the input digital signal and the local oscillator signal using a phase detector. After filtering out high-frequency noise, the phase difference signal is passed through a loop filter, and a control signal is output to a numerically controlled oscillator (CNC). The output frequency and phase of the CNC oscillator are adjusted until the frequency and phase of the input signal and the local oscillator signal are completely locked. Once locked, the stable output of the CNC oscillator serves as the absolute clock reference for the smartwatch in offline mode.

[0054] Through continuous processing including peak detection, linear amplitude normalization, analog-to-digital conversion, and digital phase-locked loop locking, a stable and reliable offline absolute clock reference is obtained, providing core support for high-precision time calibration of smartwatches in offline mode.

[0055] Furthermore, the method provided in this application embodiment includes:

[0056] The temperature voltage signal output from the temperature sensor is input to an analog differential architecture, which consists of an operational amplifier and an RC network. The temperature voltage signal is differentiated in the time domain using the analog differential architecture to output a gradient voltage signal, which is proportional to the rate of temperature change. This gradient voltage signal is amplified by a variable gain amplifier and superimposed on the bias voltage terminal of the varactor diode of the main crystal oscillator. Through signal superposition, the crystal oscillator output frequency undergoes instantaneous reverse compensation with respect to the rate of temperature change.

[0057] In this embodiment, the main crystal oscillator varactor diode is a semiconductor device whose junction capacitance changes with the reverse bias voltage. It is connected in series in the crystal oscillation circuit, and the output frequency of the crystal oscillator can be finely adjusted by adjusting the bias voltage.

[0058] Specifically, the second-order calibration mode and the first-order calibration mode deployed in the aforementioned steps operate in parallel and independently. They start and stop simultaneously, each completing its corresponding error compensation task, without any sequential dependency. The analog temperature voltage signal output from the smartwatch's built-in temperature sensor, with a voltage range consistent with the smartwatch's system power supply (e.g., 0V to 3.3V), first passes through a 100nF decoupling capacitor to filter out high-frequency electromagnetic interference before being input to the analog differential architecture. This analog differential architecture uses an inverting operational amplifier to form the basic differential circuit. An RC network is formed by a capacitor connected in series between the inverting input and the signal input of the operational amplifier, and a resistor connected in parallel between the inverting input and the output of the operational amplifier. The operational amplifier is a low-power general-purpose operational amplifier powered by a single 3.3V supply, such as the LM321, with a 10nF capacitor, a 100kΩ resistor, and a time constant of 1ms, effectively responding to temperature changes from 0.1℃ / s to 10℃ / s.

[0059] Then, the input temperature and voltage signals are differentiated in the time domain according to the analog differential architecture. The calculation process is based on the virtual short and virtual open characteristics of the operational amplifier: when the input temperature and voltage signals flow through the series capacitor, the magnitude of the current generated by the capacitor is proportional to the instantaneous rate of change of the input voltage; the faster the voltage changes, the larger the capacitor current. According to the virtual open characteristic, all the current flowing through the capacitor will flow through the parallel feedback resistor. According to the virtual short characteristic, the voltage at the inverting input terminal of the operational amplifier is equal to the ground voltage at the non-inverting input terminal; therefore, the voltage across the feedback resistor is the output voltage of the operational amplifier. Finally, a gradient voltage signal is output. The amplitude of this gradient voltage signal is proportional to the rate of change of the ambient temperature, and the signal polarity is opposite to the rate of change of temperature. It can reflect the instantaneous trend of ambient temperature change in real time, rather than just reflecting the current absolute temperature value, providing a basis for subsequent transient temperature compensation.

[0060] Next, a variable gain amplifier is used to amplify the gradient voltage signal. The specific gain adjustment and calibration methods will be explained in detail later. Then, the amplified gradient voltage signal is superimposed on the original DC bias voltage of the main crystal oscillator varactor diode through a passive adder circuit consisting of two resistors of equal resistance. One end of each resistor is connected to the output of the amplified gradient voltage signal and the output of the original DC bias voltage, respectively, while the other end is connected to the bias voltage terminal of the main crystal oscillator varactor diode. The original DC bias voltage is provided by the smartwatch's 3.3V system power supply through a voltage divider circuit and is set to 1.75V, which is the midpoint of the linear operating range of the main crystal oscillator varactor diode. The superimposed bias voltage ranges from 1.0V to 2.5V, which is the general linear operating range of the varactor diode paired with the 32.768kHz crystal oscillator. Within this range, the junction capacitance of the varactor diode and the bias voltage have an approximately linear relationship, ensuring the accuracy of frequency compensation. The oscillation frequency of the main crystal oscillator is adjusted in real time by changing the junction capacitance of the varactor diode, thus providing instantaneous reverse compensation for the rate of temperature change of the crystal oscillator output frequency. This compensation method is a feedforward compensation, which can pre-adjust the output frequency of the crystal oscillator before the temperature of the crystal itself catches up with the change in ambient temperature, effectively offsetting the frequency lag error caused by the thermal inertia of the crystal oscillator.

[0061] Finally, the first-order calibration mode continuously processes the frequency signals output by the main and auxiliary crystal oscillators to compensate for steady-state common-mode temperature drift, while the second-order calibration mode synchronously acquires and processes the temperature signal to compensate for transient temperature gradient error. The outputs of both are eventually merged into a unified calibrated clock reference.

[0062] By simulating the differential architecture to extract the temperature change rate and realize feedforward instantaneous compensation of the main crystal oscillator frequency, the problem that the existing technology can only compensate for steady-state temperature drift and cannot cope with rapid temperature changes is effectively solved, and the offline time calibration accuracy of smartwatches in complex temperature environments is significantly improved.

[0063] Furthermore, the method provided in this application embodiment includes:

[0064] The gain coefficient of the variable gain amplifier is dynamically adjusted according to the amplitude of the temperature change rate; wherein the gain coefficient has a non-linear positive correlation with the temperature change rate; if the gain coefficient is a low bias voltage, the gain amplifier maintains a low gain and performs low-order amplification of small gradient signals; if the gain coefficient is a high bias voltage, the gain amplifier maintains a high gain and performs high-intensity compensation for large gradient signals.

[0065] In one embodiment, the current ambient temperature change rate is indirectly obtained by first acquiring the real-time amplitude of the gradient voltage signal output by the analog differential architecture. Two gradient voltage thresholds, 0.1V and 1.0V, are preset, corresponding to temperature change rates of 0.1℃ / s and 1.0℃ / s, respectively, dividing the temperature change rate into three intervals: small gradient, medium gradient, and large gradient. The gain coefficient exhibits a non-linear positive correlation with the temperature change rate. In the small gradient interval, the gain coefficient remains constant at a low value; in the medium gradient interval, the gain coefficient increases linearly with the temperature change rate; and in the large gradient interval, the gain coefficient remains constant at a high value. The aforementioned thresholds and gain coefficient can be calibrated and adjusted at the factory according to the temperature characteristics of different main crystal oscillators.

[0066] When the gradient voltage amplitude is less than 0.1V, it is determined to be a small temperature gradient change. At this time, a low bias voltage is output to the gain control terminal of the variable gain amplifier, so that the gain amplifier maintains a low gain coefficient of 1, and performs low-order amplification of the small gradient signal. This process can avoid excessive amplification of small temperature fluctuations and circuit noise in the environment, prevent the introduction of unnecessary frequency disturbances, and ensure the clock accuracy of the system when the temperature is stable.

[0067] When the gradient voltage amplitude is between 0.1V and 1.0V, it is determined to be a moderate temperature gradient change. At this time, a bias voltage that increases linearly with the gradient voltage amplitude is output to the gain control terminal of the variable gain amplifier, causing the gain coefficient to increase linearly from 1 to 10. This process can achieve a smooth transition of compensation intensity, avoid clock jumps caused by sudden gain changes, and ensure the continuity and stability of the compensation process.

[0068] When the gradient voltage amplitude exceeds 1.0V, it is determined to be a large temperature gradient change. At this time, a high bias voltage is output to the gain control terminal of the variable gain amplifier, keeping the gain amplifier at a high gain factor of 10, and performing high-intensity amplification of the large gradient signal. This process ensures that the gradient voltage signal can obtain sufficient amplification when the ambient temperature changes rapidly, providing sufficient driving capability for subsequent frequency compensation and avoiding clock errors caused by insufficient compensation intensity.

[0069] By dynamically adjusting the gain coefficient of the variable gain amplifier nonlinearly based on the amplitude of the temperature change rate, noise interference under small temperature gradients is effectively suppressed while ensuring the compensation strength under large temperature gradients, significantly improving the adaptability and accuracy of second-order transient temperature compensation under different operating conditions.

[0070] Furthermore, the method provided in this application embodiment includes:

[0071] Based on the service life of the smartwatch, individual error curves are extracted by calling historical data. The individual error curves are determined by the individual differences of the crystal oscillator and the aging and decay trend. After performing calibration based on the first-order calibration mode and the second-order calibration mode, residual phase error compensation is performed by performing positioning based on the individual error curves.

[0072] Optionally, each time the smartwatch connects to the internet and synchronizes with the standard network time, it calculates the difference between the local clock and the standard time during the current offline period, using the standard network time as a reference, as the cumulative time error. Simultaneously, it automatically collects the average ambient temperature, cumulative runtime data, and the aforementioned cumulative time error data during the current offline period. These three types of data are stored in the device's internal non-volatile memory in timestamp order, forming a historical error dataset. This dataset accumulates continuously over the device's service life, providing raw data support for subsequent analysis of individual error curves.

[0073] Every 30 days, the smartwatch automatically retrieves the historical error dataset from its non-volatile memory. It then uses the least squares method to perform piecewise fitting on the data, extracting the static error component related to individual crystal oscillator differences and the aging attenuation error component related to service time. The specific steps are as follows:

[0074] First, all historical data are divided into temperature intervals of 5°C based on the average ambient temperature. Within each temperature interval, a sample group with similar cumulative runtime is selected, and the average error per unit runtime within that temperature interval is calculated to obtain the influence coefficient of temperature on time error. The cumulative time error of each sample is then subtracted from the temperature influence error of the corresponding temperature interval to obtain the pure error data after eliminating temperature interference. This pure error data only includes the errors caused by the inherent individual differences of the crystal oscillator and long-term aging attenuation.

[0075] Using cumulative runtime as the x-axis and pure error data as the y-axis, a linear fit is performed using the least squares method to obtain a fitted straight line. The intercept of this fitted line on the y-axis represents the static error component of the crystal oscillator, reflecting the inherent frequency deviation of the crystal oscillator at the time of manufacture; the slope of the fitted line represents the aging decay error component, reflecting the frequency decay rate of the crystal oscillator as its service time increases. The changes of the static error component and the aging decay error component over time are superimposed to generate a unique individual error curve specific to this device. After each network synchronization with standard network time, the smartwatch automatically adds the latest error data to the historical error dataset, repeating the temperature error elimination and linear fitting steps to incrementally update the individual error curve, ensuring that the curve reflects the current characteristics of the crystal oscillator in real time.

[0076] Next, after performing the basic calibration processing in both first-order and second-order calibration modes, the cumulative runtime and average ambient temperature parameters of the current offline cycle are obtained. First, the temperature range corresponding to the individual error curve is located based on the average ambient temperature. Then, the cumulative runtime is input into the linear error curve corresponding to this range, and linear interpolation is used for positioning calculation. The specific process of linear interpolation is as follows: First, two calibrated time points adjacent to the current cumulative runtime are determined on the linear error curve of this temperature range, and the reference error values ​​corresponding to these two time points are read respectively. The time difference and error difference between the two calibrated time points are calculated to obtain the unit time error change of this curve segment. Then, the time difference between the current cumulative runtime and the previous calibrated time point is calculated, and this time difference is multiplied by the unit time error change to obtain the error increment at the current moment. Finally, the reference error value of the previous calibrated time point is added to the error increment to obtain the residual phase error value corresponding to the current operating condition. This residual phase error value is a unique error caused by the individual characteristics of the crystal oscillator and long-term aging, which cannot be covered by first-order and second-order general calibration.

[0077] Finally, the calculated residual phase error value is directly added to the phase accumulator of the digital phase-locked loop (PLL) to correct the phase of the absolute clock reference output by the PLL, thus completing the compensation process for the residual phase error. After compensation, the smartwatch automatically records the remaining time error after compensation and updates it to the historical error dataset during the next network synchronization, forming a closed-loop process of data acquisition, curve construction, error compensation, and data update.

[0078] By mining historical operating data to construct a device-specific individual error curve, and performing residual phase error compensation on the basis of general two-level calibration, a combination of general calibration and personalized compensation is achieved. This effectively eliminates long-term time errors caused by individual crystal oscillator differences and aging attenuation during the service life, and significantly improves the long-term stability and accuracy of offline time calibration for smartwatches.

[0079] Furthermore, the method provided in this application embodiment includes:

[0080] After offline time calibration is completed, the feedback management circuit is triggered by the locking indication signal of the digital phase-locked loop to package and store the calibrated feature parameters in a non-volatile register. When the smartwatch wakes up from offline mode, it first reads the calibrated feature parameters stored in the non-volatile register for rapid recovery.

[0081] In one embodiment, after offline time calibration is completed, when the digital phase-locked loop (PLL) achieves complete frequency and phase lock with the local oscillation signal, the PLL's lock indicator output pin will output a high-level active signal. This lock indicator signal is connected to the clock terminal of a feedback management circuit composed of D flip-flops. When the lock indicator signal transitions from low to high, the D flip-flops output a trigger signal, initiating the calibration parameter storage process. This triggering method is hardware-triggered, with a response delay of less than 1 microsecond, ensuring parameter storage is completed as soon as the locked state stabilizes.

[0082] The calibration characteristic parameters are all configurable parameters generated during the operation of first-order calibration mode, second-order calibration mode, and residual phase compensation mode. The currently valid calibration characteristic parameters are packaged according to a preset engineering format. These parameters include the reference bias voltage of the main crystal oscillator varactor diode, the current gain coefficient of the variable gain amplifier, the initial value of the phase accumulator of the digital phase-locked loop, and the latest update version number of the individual error curve. These parameters are then concatenated into a 32-bit fixed-length data block, with each parameter occupying a fixed number of bytes. A 1-byte checksum is added to the end of the data block to verify data integrity. The packaged data block is then written to a designated storage sector in the smartwatch's internal non-volatile register, overwriting the previously stored parameter data.

[0083] When the smartwatch wakes up from sleep or power failure, the system startup process prioritizes reading the non-volatile registers. First, it reads the calibration parameter data block from the specified sector of the non-volatile register, verifying the data's integrity through checksums. If the data verification passes, the parameters are directly loaded into the corresponding hardware registers, allowing the system to immediately enter the stable operating state after the last calibration, without needing to re-execute the complete first-order and second-order calibration processes. If the data verification fails, the system automatically loads the factory default parameters and restarts the complete calibration process.

[0084] The hardware automatically stores calibration parameters by locking the indication signal with a digital phase-locked loop, and prioritizes reading the stored calibration parameters for rapid recovery when the device wakes up. This effectively solves the problem that a complete recalibration is required after the device wakes up in the prior art, realizes the continuity of offline time calibration status, and significantly shortens the clock stabilization time after the device wakes up.

[0085] In summary, the offline time calibration accuracy optimization method for smartwatches provided in this application has the following technical effects:

[0086] This application achieves compensation by simultaneously initiating first-order steady-state calibration and second-order transient calibration modes, superimposing and fusing compensation voltages, using linear interpolation to locate residual errors, and combining the individual error curve of the crystal oscillator. It also relies on a digital phase-locked loop to realize parameter storage and wake-up recovery, thereby eliminating crystal oscillator temperature drift, thermal inertia, and aging decay errors. This makes the offline time calibration results of the smartwatch more accurate and reliable, achieving the technical effect of improving the time calibration accuracy of the smartwatch in offline mode while ensuring its long-term timekeeping stability.

[0087] Example 2, as Figure 2 As shown, based on the same inventive concept as in Embodiment 1 above, this application provides an offline time calibration accuracy optimization system for smartwatches, the system comprising:

[0088] Offline calibration mode deployment module 1, which is used to deploy an offline calibration mode in a smartwatch, wherein the offline calibration mode is jointly deployed with first-order dual-crystal differential common-mode compensation and second-order temperature gradient feedforward transient compensation.

[0089] The calibration steps based on the offline calibration mode include:

[0090] The first-order calibration mode construction module 2 is used to construct a first-order calibration mode with a dual-crystal differential structure. The first-order calibration mode eliminates common-mode temperature drift and generates a pure reference signal by mixing and differential processing of the main crystal oscillator and the auxiliary crystal oscillator.

[0091] Second-order calibration mode construction module 3 is used to construct a second-order calibration mode with an analog differential architecture. The second-order calibration mode generates a gradient voltage signal by performing time-domain differentiation on the temperature voltage signal, and superimposes it onto the bias voltage terminal of the main crystal oscillator varactor diode to achieve temperature gradient feedforward transient compensation.

[0092] The offline time calibration execution module 4 is used to integrate the first-order calibration mode and the second-order calibration mode to perform offline time calibration management on the smartwatch.

[0093] Furthermore, the offline calibration mode deployment module 1 is used to perform the following steps:

[0094] Based on the dual-crystal differential structure, a first-order calibration mode is deployed using first-order dual-crystal differential common-mode compensation. Based on the analog differential architecture, a second-order calibration mode is deployed using main crystal varactor diode modulation under temperature gradient feedforward transient compensation. The first-order and second-order calibration modes are then combined to deploy an offline calibration mode for the smartwatch.

[0095] Furthermore, the first-order calibration mode construction module 2 is used to perform the following steps:

[0096] A temperature-compensated crystal oscillator (TCC) is used as the main crystal oscillator, and a conventional crystal oscillator with an opposite temperature coefficient to the main crystal oscillator is used as the auxiliary crystal oscillator, forming a dual-crystal differential structure. For the first offline time interval, the main crystal oscillator in the dual-crystal differential structure outputs a reference frequency signal, and the auxiliary crystal oscillator outputs a local oscillator frequency signal. The reference frequency signal is a temperature-compensated reference frequency, and the local oscillator frequency signal is an uncompensated auxiliary frequency. The reference frequency signal and the local oscillator frequency signal are mixed and differentially processed to obtain a pure reference signal.

[0097] Furthermore, the first-order calibration mode construction module 2 is used to perform the following steps:

[0098] The reference frequency signal and the local oscillator frequency signal are input into a multiplier to obtain a product signal containing the sum frequency and the difference frequency; the sum frequency component of the product signal is filtered out according to a low-pass filter to obtain a difference signal, wherein the difference signal reflects the frequency difference between the main crystal oscillator and the auxiliary crystal oscillator.

[0099] Furthermore, the first-order calibration mode construction module 2 is used to perform the following steps:

[0100] The differential signal is subjected to peak detection and amplitude normalization to determine the normalized pure reference signal; the pure reference signal is then converted from analog to digital and entered into a digital phase-locked loop for frequency and phase locking, serving as the absolute clock reference in offline mode.

[0101] Furthermore, the second-order calibration mode construction module 3 is used to perform the following steps:

[0102] The temperature voltage signal output from the temperature sensor is input to an analog differential architecture, which consists of an operational amplifier and an RC network. The temperature voltage signal is differentiated in the time domain using the analog differential architecture to output a gradient voltage signal, which is proportional to the rate of temperature change. This gradient voltage signal is amplified by a variable gain amplifier and superimposed on the bias voltage terminal of the varactor diode of the main crystal oscillator. Through signal superposition, the crystal oscillator output frequency undergoes instantaneous reverse compensation with respect to the rate of temperature change.

[0103] Furthermore, the second-order calibration mode construction module 3 is used to perform the following steps:

[0104] The gain coefficient of the variable gain amplifier is dynamically adjusted according to the amplitude of the temperature change rate; wherein the gain coefficient has a non-linear positive correlation with the temperature change rate; if the gain coefficient is a low bias voltage, the gain amplifier maintains a low gain and performs low-order amplification of small gradient signals; if the gain coefficient is a high bias voltage, the gain amplifier maintains a high gain and performs high-intensity compensation for large gradient signals.

[0105] Furthermore, the offline time calibration execution module 4 is used to perform the following steps:

[0106] Based on the service life of the smartwatch, individual error curves are extracted by calling historical data. The individual error curves are determined by the individual differences of the crystal oscillator and the aging and decay trend. After performing calibration based on the first-order calibration mode and the second-order calibration mode, residual phase error compensation is performed by performing positioning based on the individual error curves.

[0107] Furthermore, the offline time calibration execution module 4 is used to perform the following steps:

[0108] After offline time calibration is completed, the feedback management circuit is triggered by the locking indication signal of the digital phase-locked loop to package and store the calibrated feature parameters in a non-volatile register. When the smartwatch wakes up from offline mode, it first reads the calibrated feature parameters stored in the non-volatile register for rapid recovery.

[0109] The offline time calibration accuracy optimization system for a smartwatch provided in this embodiment of the invention can execute the offline time calibration accuracy optimization method for a smartwatch provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0110] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0111] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for optimizing the offline time calibration accuracy of a smartwatch, characterized in that, The method includes: An offline calibration mode is deployed in a smartwatch, wherein the offline calibration mode is jointly deployed with first-order dual-crystal differential common-mode compensation and second-order temperature gradient feedforward transient compensation. The calibration steps based on the offline calibration mode include: A first-order calibration mode is constructed using a dual-crystal differential structure. The first-order calibration mode eliminates common-mode temperature drift by mixing and differential processing of the main crystal oscillator and the auxiliary crystal oscillator, thereby generating a pure reference signal. A second-order calibration mode is constructed using an analog differential architecture. The second-order calibration mode generates a gradient voltage signal by performing time-domain differentiation on the temperature voltage signal, and superimposes it onto the bias voltage terminal of the main crystal oscillator varactor diode to achieve temperature gradient feedforward transient compensation. By combining the first-order calibration mode and the second-order calibration mode, offline time calibration management can be performed on smartwatches.

2. The method for optimizing offline time calibration accuracy of a smartwatch as described in claim 1, characterized in that, Deploying an offline calibration mode in smartwatches includes: Based on the dual-crystal differential structure, and using first-order dual-crystal differential common-mode compensation as the compensation method, a first-order calibration mode is deployed. Based on the analog differential architecture, a second-order calibration mode is deployed using the main crystal oscillator varactor diode modulation under temperature gradient feedforward transient compensation as the compensation method. By integrating the first-order calibration mode and the second-order calibration mode, an offline calibration mode is deployed for the smartwatch.

3. The method for optimizing offline time calibration accuracy of a smartwatch as described in claim 2, characterized in that, The first-order calibration mode includes: A temperature-compensated crystal oscillator is used as the main crystal oscillator, and a common crystal oscillator with the opposite temperature coefficient to the main crystal oscillator is used as the auxiliary crystal oscillator, forming a dual-crystal differential structure. For the first offline time interval, the main crystal oscillator in the dual-crystal differential structure outputs a reference frequency signal, and the auxiliary crystal oscillator outputs a local oscillator frequency signal. The reference frequency signal is a temperature-compensated base frequency, and the local oscillator frequency signal is an uncompensated auxiliary frequency. The reference frequency signal and the local oscillator frequency signal are mixed and differentially processed to obtain a pure reference signal.

4. The method for optimizing offline time calibration accuracy of a smartwatch as described in claim 3, characterized in that, The mixing and differential processing of the reference frequency signal and the local oscillator frequency signal includes: The reference frequency signal and the local oscillator frequency signal are input into a multiplier to obtain a product signal containing the sum frequency and the difference frequency; The sum-frequency component of the product signal is filtered out using a low-pass filter to obtain a differential signal, wherein the differential signal reflects the frequency difference between the main crystal oscillator and the auxiliary crystal oscillator.

5. The method for optimizing offline time calibration accuracy of a smartwatch as described in claim 4, characterized in that, After obtaining the differential signal, the following is included: The differential signal is subjected to peak detection and amplitude normalization to determine the normalized pure reference signal; The pure reference signal is converted from analog to digital and then entered into a digital phase-locked loop for frequency and phase locking, serving as the absolute clock reference in offline mode.

6. The method for optimizing offline time calibration accuracy of a smartwatch as described in claim 2, characterized in that, The second-order calibration mode includes: The temperature voltage signal output by the temperature sensor is input to an analog differential architecture, which consists of an operational amplifier and an RC network. Based on the analog differential architecture, the temperature voltage signal is differentiated in the time domain to output a gradient voltage signal, wherein the gradient voltage signal is proportional to the rate of temperature change. The gradient voltage signal is amplified by a variable gain amplifier and superimposed on the bias voltage terminal of the main crystal varactor diode. By superimposing the signals, the crystal oscillator output frequency generates instantaneous reverse compensation with the rate of change of temperature.

7. The method for optimizing offline time calibration accuracy of a smartwatch as described in claim 6, characterized in that, The gain coefficient of the variable gain amplifier is dynamically adjusted based on the magnitude of the temperature change rate. Among them, the gain coefficient has a non-linear positive correlation with the rate of temperature change; If the gain coefficient is a low bias voltage, the gain amplifier maintains a low gain and performs low-order amplification of small gradient signals; if the gain coefficient is a high bias voltage, the gain amplifier maintains a high gain and performs high-intensity compensation for large gradient signals.

8. The method for optimizing offline time calibration accuracy of a smartwatch as described in claim 1, characterized in that, The method further includes: Based on the service life of smartwatches, individual error curves are extracted by calling historical data. The individual error curves are determined by the individual differences of crystal oscillators and the aging and decay trend. After performing calibration based on first-order and second-order calibration modes, residual phase error compensation is performed by performing positioning based on individual error curves.

9. The method for optimizing offline time calibration accuracy of a smartwatch as described in claim 1, characterized in that, After performing offline time calibration management on the smartwatch, the following is included: After offline time calibration is completed, the feedback management circuit is triggered by the locking indication signal of the digital phase-locked loop to package and store the calibrated characteristic parameters in a non-volatile register. When the smartwatch wakes up from offline mode, it prioritizes reading the calibrated feature parameters stored in the non-volatile register for rapid recovery.

10. An offline time calibration accuracy optimization system for a smartwatch, characterized in that, The system is used to implement the offline time calibration accuracy optimization method for a smartwatch according to any one of claims 1-9, the system comprising: An offline calibration mode deployment module is used to deploy an offline calibration mode in a smartwatch, wherein the offline calibration mode is jointly deployed with first-order dual-crystal differential common-mode compensation and second-order temperature gradient feedforward transient compensation. The calibration steps based on the offline calibration mode include: A first-order calibration mode construction module is used to construct a first-order calibration mode with a dual-crystal differential structure. The first-order calibration mode eliminates common-mode temperature drift by mixing and differential processing of the main crystal oscillator and the auxiliary crystal oscillator, thereby generating a pure reference signal. The second-order calibration mode construction module is used to construct a second-order calibration mode using an analog differential architecture. The second-order calibration mode generates a gradient voltage signal by performing time-domain differentiation on the temperature voltage signal, and superimposes it onto the bias voltage terminal of the main crystal oscillator varactor diode to achieve temperature gradient feedforward transient compensation. The offline time calibration execution module is used to integrate the first-order calibration mode and the second-order calibration mode to perform offline time calibration management for the smartwatch.

Images