Staged temperature control method and system for ovenized crystal oscillator

Abstract

This application discloses a staged temperature control method and system for a cryogenic crystal oscillator, comprising: acquiring the inflection point temperature and historical steady-state drive value of the cryogenic crystal oscillator; sampling the temperature of the cryogenic crystal oscillator, and performing a moving average of the sampled temperature at each sampling time to obtain the average temperature at the sampling time; subtracting the inflection point temperature from the average temperature to obtain the subsequent temperature difference at the corresponding sampling time; executing a startup strategy based on the validity of the historical steady-state drive value, the inflection point temperature, the start-up temperature, the reference room temperature, the average start-up temperature, and the start-up temperature difference; and dynamically adjusting the heating power according to the subsequent temperature difference at different sampling times. This application enables the cryogenic crystal oscillator to complete startup by executing a corresponding startup strategy according to its own inflection point parameters and historical steady-state drive value, so that the startup power no longer completely depends on a fixed empirical value. Furthermore, it dynamically adjusts the heating direction and amplitude according to the subsequent temperature difference to prevent problems such as temperature overshoot or excessively low temperature during startup.

Description

Technical Field

[0001] This application relates to the field of oscillator technology, and in particular to a staged temperature control method and system for a thermostatic crystal oscillator. Background Technology

[0002] The frequency of a quartz crystal changes with temperature, and the effect of temperature change on the frequency is minimal near the inflection point of the frequency-temperature curve.

[0003] In actual debugging, it can be seen that even for the same model of oven-controlled crystal oscillator (OCXO), different individuals will have different inflection point temperatures, inflection point frequency deviations, thermal inertia, and the amount of heating required to maintain the inflection point temperature. In addition, if the same oven-controlled crystal oscillator is started from room temperature, restarted in a high-temperature environment, or already at a temperature higher than the inflection point when powered on, using the same heating method will inevitably lead to problems such as insufficient heating or excessive heating. Summary of the Invention

[0004] The main objective of this application is to provide a staged temperature control method and system for a thermostatic crystal oscillator, which can solve the technical problem that the fixed and uniform heating method in the prior art cannot be adapted to different individual thermostatic crystal oscillators and thermostatic crystal oscillators under different conditions.

[0005] To achieve the above objectives, the first aspect of this application provides a staged temperature control method for a cryogenic crystal oscillator, the method comprising: Obtain the inflection point temperature of the cryogenic crystal oscillator and the historical steady-state drive value actually used when the oscillator was last stable within the target temperature window near the inflection point temperature; Temperature sampling is performed on the isothermal crystal oscillator, and the sampled temperature is averaged at each sampling time to obtain the average temperature at the corresponding sampling time. The average temperature includes the power-on average temperature. The inflection point temperature is used as the target temperature for temperature control. The difference between the target temperature and the average temperature is calculated to obtain the subsequent temperature difference at the corresponding sampling time. The subsequent temperature difference includes the start-up temperature difference. Based on the validity of historical steady-state driving values, inflection point temperature, startup temperature, reference room temperature, average startup temperature, and startup temperature difference, execute the corresponding startup strategy. The heating power is dynamically adjusted based on the subsequent temperature difference at different sampling times.

[0006] To achieve the above objectives, a second aspect of this application provides a staged temperature control system for a thermostatic crystal oscillator, the system comprising: a thermostatic crystal oscillator, a temperature sampling device, a heating actuator, and a controller; A temperature sampling device is used to sample the temperature of the thermostatic crystal oscillator and provide the sampled temperature to the controller; A heating actuator is used to heat a thermostatic crystal oscillator under the control of a controller. A controller for executing the staged temperature control method for the isothermal crystal oscillator described in any of the above-mentioned methods.

[0007] The embodiments of this application have the following beneficial effects: This application does not simply raise the temperature to a set value, but rather allows each isothermal crystal oscillator to execute a corresponding startup strategy based on its own inflection point parameters and historical steady-state drive values ​​to complete the startup. The startup power no longer relies entirely on fixed empirical values. Furthermore, the heating power is dynamically adjusted based on the subsequent temperature difference, reducing inflection point temperature overshoot while maintaining temperature rise, and preventing problems such as temperature overshoot or undershoot during startup. Attached Figure Description

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

[0009] in: Figure 1 This is a flowchart of the staged temperature control method for the isothermal crystal oscillator in the embodiments of this application; Figure 2 This is a structural block diagram of the staged temperature control system of the isothermal crystal oscillator in the embodiments of this application. Detailed Implementation

[0010] 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 some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0011] A temperature-controlled crystal oscillator (OCXO) is a frequency source device that achieves high frequency stability and low phase noise performance by heating a crystal resonator to keep it operating within a preset constant temperature range for an extended period.

[0012] Oven-controlled crystal oscillators (OCXOs) do not maintain the temperature at an arbitrary fixed value. The frequency of a quartz crystal changes with temperature, and the effect of temperature changes on the frequency is minimal near the inflection point of the frequency-temperature curve. Therefore, each OCXO needs to first determine its own inflection point temperature, and then maintain the crystal temperature as close to that point as possible during product operation.

[0013] PLL can perform frequency multiplication, frequency division, and jitter suppression, but it cannot eliminate the slow frequency deviation caused by temperature changes in the oven-controlled crystal oscillator (OCXO). Therefore, temperature control should be done well at the oven-controlled crystal oscillator (OCXO) end.

[0014] In actual debugging, it can be seen that even for the same model of oven-controlled crystal oscillator (OCXO), different individuals will have different inflection point temperatures, inflection point frequency deviations, thermal inertia, and the amount of heating required to maintain the inflection point. The heating methods for the same product cannot be exactly the same whether it is started at room temperature, restarted in a high-temperature environment, or already at a temperature above the inflection point. If the same starting power is always applied, the common result is slow heating when cold and easy overshoot when hot; if frequent adjustments are made after approaching the inflection point, the temperature will fluctuate around the target temperature.

[0015] A common approach is to pre-set a target temperature, heat the device with a fixed power initially after startup, and then switch to PID or PI control once the temperature approaches the target. This method is simple to implement, but it typically only considers the current temperature difference and does not take into account individual device differences, the initial temperature at startup, the current rate of heating, or the actual sustaining power required after stabilization.

[0016] Therefore, the existing technical solutions have the following drawbacks: Using the same starting power when the machine is cold, hot, or already above the inflection point may result in insufficient temperature rise for a cold machine, while a hot machine may easily exceed the target temperature.

[0017] Focusing only on the current temperature difference, without considering the rate at which the temperature is rising, means that even with reduced power near the inflection point, the heat previously accumulated inside the furnace may continue to push the temperature higher.

[0018] If the temperature is considered to have stabilized as soon as it passes the target value, without checking the trend and the magnitude of the fluctuation, it is easy to misjudge a short-term passage as stability.

[0019] Even after reaching steady state, if the heating drive is still modified every second, it is easy to create unnecessary small oscillations near the inflection point.

[0020] Using a fixed sustaining drive value cannot accommodate the differences in heat dissipation between individual OCXOs and the entire system; if the value is continuously updated from power-on, large errors during the temperature rise process may be saved as steady-state data.

[0021] Based on this, this application provides a staged temperature control scheme for a thermostatic crystal oscillator.

[0022] like Figure 1 As shown, in one embodiment, a staged temperature control method for a cryogenic crystal oscillator is provided. The staged temperature control method for the cryogenic crystal oscillator specifically includes the following steps: S100: Obtain the inflection point temperature of the isothermal crystal oscillator and the historical steady-state drive value actually used when the oscillator was last stable within the target temperature window near the inflection point temperature.

[0023] Specifically, through heating and temperature control, a cryogenic crystal oscillator (OCXO) can be maintained near the inflection point temperature of its frequency-temperature curve. Within this region, the first-order sensitivity of the frequency to temperature changes is low, thus enabling the cryogenic crystal oscillator (OCXO) to provide a stable reference clock for high-precision audio, measurement, communication, or synchronization systems.

[0024] Historical steady-state drive values ​​and inflection point temperatures are stored in association with the device identifier of the same oven-controlled crystal oscillator (OCXO). After power-on, the controller first reads the inflection point temperature T_k, historical steady-state drive value U_k, and valid flag of U_k of the oven-controlled crystal oscillator (OCXO), and then checks the parameter range and stored verification results.

[0025] Here, T_k determines the target temperature to which the isothermal crystal oscillator will eventually stabilize, or serves as the final temperature control target during operation; Δf_k is used to calculate the frequency correction for each output; and U_k represents the actual heating drive amount used when the isothermal crystal oscillator last stabilized within the target temperature window near T_k. The target temperature window includes the inflection point temperature, and the differences between its maximum and minimum temperatures and the inflection point temperature are both within a preset range. U_k can be a 16-bit digital value used for the starting position of the power-on drive and can be updated after stabilization.

[0026] In one specific embodiment, if T_k is invalid, the controller can use a safe default target value, which can be recalibrated later.

[0027] In one specific embodiment, during the factory calibration stage, the temperature-controlled crystal oscillator (TCXO) under test is heated at least once, and the time series data collected during each heating process is acquired. The acquired data includes: ambient temperature Ta, furnace temperature T, heating power P, crystal oscillator frequency OCXO frequency f, and optional voltage control voltage Vtune. Based on the frequency-temperature curve, the candidate temperature region where the first derivative of frequency with respect to temperature is zero or close to zero is determined. The target inflection point temperature T0 and target frequency f0 of the isothermal crystal oscillator under test are determined by combining the frequency variance, long-term drift and power consumption constraints.

[0028] S200: The temperature of the thermostatic crystal oscillator is sampled, and the sampled temperature is averaged at each sampling time to obtain the average temperature at the corresponding sampling time. The average temperature includes the power-on average temperature.

[0029] Specifically, the current average temperature is obtained by performing a sliding average on the current sampled temperature and the historical sampled temperature; that is, the current average temperature T_avg[n] is obtained by performing a sliding average on a first preset number of sampled temperatures, including the current sampled temperature and the historical sampled temperature.

[0030] For example, taking the time interval between adjacent sampling times as the first preset duration, the temperature value is read once every first preset duration (e.g., 1 second), and the effective temperature values ​​read in the first preset number (e.g., the most recent 10 points) are averaged to obtain the average temperature (i.e., the control filter temperature).

[0031] One of the first preset durations is one sampling period.

[0032] The average temperature at the current sampling moment is obtained by averaging the current sampled temperature at that moment with a subset of historical sampled temperatures. For example, the average temperature at the current sampling moment is obtained by averaging the current sampled temperature with the nine most recent historical sampled temperatures. The same principle applies to other sampling moments.

[0033] In addition, the temperature is read every second. If the reading exceeds -30°C to 100°C, it is reread, with a maximum of 3 retries; if it is still invalid, the previous valid temperature can be used in the current control cycle.

[0034] If the number of sampled temperatures is less than the first preset number, then the actual number of temperatures will be used for temperature averaging.

[0035] The temperature resolution of the sampled temperature is approximately 0.0078°C. The control thread performs a 10-point moving average on the effective temperature, and the temperature values ​​in the display cache can be saved in increments of 0.01°C.

[0036] Temperature can be collected by a digital temperature sensor or by a thermistor in conjunction with an analog-to-digital converter; if the furnace is large, multiple temperature measurement points can also be used.

[0037] For example, the furnace temperature is collected by a thermistor, and the heating actuator is driven by a controller loop that includes PID, fuzzy PID, or analog differential control.

[0038] S300: The inflection point temperature is used as the temperature control target temperature. The difference between the temperature control target temperature and the average temperature is calculated to obtain the subsequent temperature difference at the corresponding sampling time. The subsequent temperature difference includes the start-up temperature difference.

[0039] Specifically, the subsequent temperature difference e[n] is calculated according to the following formula 1: e[n] = T_k - T_avg[n] Formula 1 The start-up temperature difference is the difference between the target temperature and the average start-up temperature. The average start-up temperature is the temperature of the thermostatic crystal oscillator at startup.

[0040] A positive e[n] indicates that the average temperature is below the inflection point temperature T_k, and heating needs to be increased; a negative e[n] indicates that the average temperature is already above the inflection point temperature T_k, and heating needs to be reduced.

[0041] S400: Based on the validity of historical steady-state drive values, inflection point temperature, start-up temperature, reference room temperature, average start-up temperature, and start-up temperature difference, execute the corresponding startup strategy.

[0042] Specifically, the validity of the historical steady-state driving value U_k may be valid or invalid.

[0043] If U_k is invalid, T_k can still be used for temperature control, but the historical drive value will not be used at startup. Instead, cold start, warm start, or overheat start will be selected based on the current temperature. If U_k is valid, the temperature will start from near U_k to avoid restarting from a fixed high power level every time the machine is turned on.

[0044] S500: Dynamically adjusts heating power based on the subsequent temperature difference at different sampling times.

[0045] Specifically, as the temperature difference changes, the heating direction and heating amplitude can be adjusted, for example, dynamically increasing or decreasing the heating to maintain the temperature at or near the inflection point temperature as much as possible.

[0046] This embodiment achieves heating and heating control by driving a heating actuator.

[0047] The heating actuator can be driven by a DAC, current source, PWM duty cycle, or any other power control method. The controller maps the heating actuator input value to a drive value that monotonically corresponds to the heating power, and determines the direction of heating power change when the drive value increases or decreases based on hardware polarity. If a DAC control method is used, the DAC value is inversely related to the heating power; when using a forward driver, only the direction of the control increment needs to be changed.

[0048] It should be noted that this application can simultaneously perform independent temperature control on multiple cryogenic crystal oscillators. The PLL synthesizes the frequencies output by the cryogenic crystal oscillators at or near the inflection point temperature, outputting the independent frequencies of each cryogenic crystal oscillator. Then, the frequency offset Δf_k at each inflection point is used to statically correct the corresponding independent frequencies to obtain the final frequency of each cryogenic crystal oscillator.

[0049] This embodiment does not simply raise the temperature to a set value, but rather allows each isothermal crystal oscillator to execute a corresponding startup strategy based on its own inflection point parameters and historical steady-state drive values ​​to complete the startup. The startup power no longer relies entirely on fixed empirical values. Furthermore, the heating power is dynamically adjusted based on the subsequent temperature difference, reducing inflection point temperature overshoot while maintaining temperature rise. This prevents problems such as temperature overshoot or undershoot during startup, ensuring that the isothermal crystal oscillator remains stable near the true inflection point temperature, effectively preventing temperature and frequency overshoot.

[0050] In one embodiment, the method further includes: If the sampled temperature enters the target window, determine whether the sampled temperature has entered a steady state within the target window; If the sampling temperature enters a steady state, the steady-state driving value required to maintain the inflection point temperature is learned within the learning window. The candidate steady-state driving values ​​of the longest stable segment within the entire learning window are retained and updated to the latest steady-state driving values. The longest stable segment is the segment with the longest continuous stable time. If there is no continuous stable segment within the entire learning window, the driving value being used at the end of the learning window is retained as the latest steady-state driving value.

[0051] Specifically, if the sampling temperature has not yet reached the temperature control target temperature, and the difference between the temperature control target temperature and the sampling temperature is less than the difference threshold, then the sampling temperature has entered the target window.

[0052] The fact that the sampled temperature has entered the target window does not mean that it has stabilized. Therefore, it is necessary to determine whether the sampled temperature within the target window has entered a steady state.

[0053] If the sampled temperature within the target window has reached a steady state, then a learning window is opened (e.g., a 30-minute learning window). Within the learning window, the steady-state driving value U_k required to maintain the inflection point temperature T_k is learned. During the learning period, the temperature is still maintained by proximal control.

[0054] Within the learning window, compare the absolute value of the subsequent temperature difference at each sampling time with the fourth temperature difference threshold (e.g., 0.05°C), and compare the difference between the highest and lowest temperatures within the most recent seventh preset time period (e.g., within 60 seconds) with the fifth temperature difference threshold (e.g., 0.03°C).

[0055] Within the learning window, for any sampling moment, if the absolute value of the subsequent temperature difference does not exceed (is less than or equal to) the fourth temperature difference threshold, and the difference between the highest and lowest temperatures within the most recent seventh preset time period does not exceed the fifth temperature difference threshold, then the sampling moment is determined to meet the continuity condition and can be used as a continuous sampling moment; otherwise, the sampling moment is determined not to meet the continuity condition and is not continuous at that sampling moment, and a new round of continuity judgment will be performed subsequently.

[0056] Similarly, for each sampling time within the learning window, a judgment is made as to whether the continuity condition is met.

[0057] Candidate steady-state driving values ​​are driving values ​​provided based on the proximal control strategy within the learning window.

[0058] The candidate steady-state driving value or the mean of the candidate steady-state driving values ​​for the longest stable segment with the longest continuous stable time within the entire learning window is retained and updated as the latest steady-state driving value. For example, to determine which segment within a 30-minute learning window has the longest continuous stable time, the driving value currently in use at the end of that segment (the end of the stable segment) is saved as the latest steady-state driving value. After the window ends, the latest steady-state driving value is written to non-volatile memory and locked.

[0059] In addition, if there are two longest stable segments, the candidate steady-state driving value or the average of the candidate steady-state driving values ​​at the end of the later longest stable segment is taken as the latest steady-state driving value.

[0060] If there is no continuous stable segment within the entire learning window, the driving value being used at the end of the learning window is retained as the latest steady-state driving value.

[0061] Let W represent the current learning window, and L[n] represent the time until the nth second when the above two consecutive conditions are met. Then, the candidate steady-state driving value U_k* is taken from the position with the longest consecutive steady time. n* = arg max{L[n] | n∈W}, U_k* = u[n*] Formula 2 In addition, if the operating temperature range of the cryogenic crystal oscillator is wide, multiple sets of U_k can be saved according to the ambient temperature, and the values ​​of adjacent intervals can be selected or interpolated during operation.

[0062] This embodiment does not simply raise the temperature to a certain set value, but rather allows each thermostatic crystal oscillator to start up according to its own inflection point parameters: it can heat up as quickly as possible when the machine is cold, reduce the heating action in advance when it is close to the inflection point, and record the drive value required to maintain the temperature after confirming that the temperature is truly stable.

[0063] In one embodiment, the step of executing a corresponding startup strategy based on the validity of historical steady-state driving values, inflection point temperature, startup temperature, reference room temperature, average startup temperature, and startup temperature difference includes: If the historical steady-state driving values ​​are valid, the heat demand ratio is calculated based on the inflection point temperature, start-up temperature, and reference room temperature. If the heat demand ratio indicates that the isothermal crystal oscillator is in a cold state, the initial drive is to increase heating based on the historical steady-state drive value; If the heat demand ratio indicates that the isothermal crystal oscillator is in a thermodynamic state, then the initial drive is the historical steady-state drive value. If the historical steady-state driving value is invalid, the difference between the temperature control target temperature and the average start-up temperature is calculated to obtain the start-up temperature difference; The start-up state of the thermostatic crystal oscillator is determined based on the start-up temperature difference, wherein the start-up state is one of cold start, hot start, and overheat start; The corresponding startup strategy is executed based on the startup status.

[0064] Specifically, if the historical steady-state driving value U_k is valid, the ambient temperature needs to be considered, and it cannot be used directly. The controller first calculates the subsequent temperature difference from the current sampling temperature to the inflection point temperature, which is the proportion of the temperature difference required to rise from room temperature to the inflection point temperature. Let the start-up temperature be T[0] and the reference room temperature be Tref, then the heat demand ratio ρ is as shown in Formula 3: ρ = clip((T_k - T[0]) / (T_k - Tref), 0, 1) Formula 3 T_k is the inflection point temperature, and clip indicates that the calculation result is limited to between 0 and 1. If ρ is close to 1 (i.e., the difference between ρ and 1 is within the first preset range), it means that the cryogenic crystal oscillator is close to a cold state, and more heating can be added based on the historical steady-state drive value U_k; if ρ is small (i.e., the difference between ρ and 1 is outside the second preset range), it means that the cryogenic crystal oscillator itself is already quite hot, and the initial drive should be close to or equal to the historical steady-state drive value U_k. The smaller the DAC value of the current hardware, the greater the heating power. Therefore, when the temperature is below T_k, the correction is from U_k to a smaller value, and when the temperature is above T_k, the correction is from a larger value. The ratio here is only used to determine the start-up position; after entering the closed loop, the next action is still determined by the actual temperature and the heating rate.

[0065] If T_ref is 25°C, clip means clip(x,0,1)=min(max(x,0),1). When ρ≥0.7, it is determined to be a far-cooling state, and heating is added based on U_k; when 0.4≤ρ<0.7, it is determined to be a medium heat demand state; when 0≤ρ<0.4, it is determined to be a hot start state, and the initial drive value is close to U_k. The smaller the current hardware DAC value, the greater the heating power; when T_k−T_ref is less than or equal to 0, Formula 3 is not used, and the preset default start drive value is used instead.

[0066] If the historical steady-state driving value U_k is invalid, the difference between the temperature control target temperature and the average start-up temperature is calculated to obtain the start-up temperature difference e[0]. Based on the start-up temperature difference e[0], it is determined whether the isothermal crystal oscillator is in a cold start, a hot start, or an overheat start when it is turned on, and the start-up strategy or temperature control strategy is executed according to the start-up status.

[0067] The startup strategy is used to determine the startup method, including whether to use the historical steady-state drive value U_k as the starting position for power-on heating, or to add some heating power to the historical steady-state drive value U_k as the starting position for heating, or to reduce the heating, etc.

[0068] This embodiment can heat up quickly when the machine is cold, and reduce the heating action in advance when approaching the inflection point. Different processing methods are used for cold start, warm start and overheat start to avoid problems such as temperature overshoot or excessively low temperature when restarting the machine, and to avoid over-adjustment caused by warm-up drift.

[0069] In one embodiment, determining the startup state of the thermostatic crystal oscillator based on the startup temperature difference includes: If the temperature difference at startup exceeds the first temperature threshold, the startup state of the cryogenic crystal oscillator is determined to be a cold start. If the temperature difference at startup is not less than the second temperature threshold and not greater than the first temperature threshold, then the startup state of the isothermal crystal oscillator is determined to be a hot start. If the temperature difference at startup is less than the second temperature threshold, then the startup state of the thermostatic crystal oscillator is determined to be overheat startup. The step of executing the corresponding startup strategy based on the startup status includes: If the startup state is cold start, then the heating drive value will be set within the preset cold start power range for heating. If the startup state is a hot start, heating will begin from the historical steady-state driving value. If the startup state is an overheat start, reduce heating and wait for the temperature to drop before entering control near the inflection point.

[0070] Specifically, in one embodiment, three cases are first distinguished based on the temperature difference at startup: e[0]>15°C is a cold start; -3°C≤e[0]≤15°C is a hot start; e[0]<-3°C indicates that the startup temperature is already much higher than the inflection point, and it enters an overheated start. Wherein, e[0]=T_k-T[0].

[0071] In the case of a cold start, the system is far from the inflection point temperature. In this case, a limited high heating power can be output. For example, a larger heating power can be provided to heat the thermostatic crystal oscillator. More specifically, a preset cold start power range can be used for heating. In the case of a hot start, heating begins from the historical steady-state drive value U_k or its vicinity. In the case of an overheat start, heating is reduced first (i.e., heating begins from a second heating power that is smaller than the first heating power corresponding to the historical steady-state drive value U_k. For example, the DAC is set to the preset overheat start drive value 18000). After the temperature drops, closed-loop control close to the inflection point is entered.

[0072] This embodiment can heat up quickly when the machine is cold, and reduce the heating action in advance when approaching the inflection point. Different processing methods are used for cold start, warm start and overheat start, which can reduce problems such as temperature overshoot or excessively low temperature when restarting the machine.

[0073] In one embodiment, determining whether the sampled temperature within the target window has entered a steady state includes: Calculate the temperature change trend and temperature fluctuation within the most recent third preset time period corresponding to each sampling time in the target window, and calculate the average value of the absolute temperature difference of the most recent preset number of times, where the absolute temperature difference is the absolute value of the difference between the temperature control target temperature and the corresponding average temperature. If the absolute value, temperature fluctuation, and average value of the temperature change trend for the fourth consecutive preset duration within the target window are all less than the corresponding judgment threshold, then the sampling temperature is determined to have entered a steady state.

[0074] Specifically, the sampling temperature entering the target window does not mean that it has stabilized. The controller retains the sampling temperature for the third most recent preset duration (e.g., the most recent 60 seconds) for each sampling moment, calculates the temperature change trend r (in °C / s or °C / min) and temperature fluctuation σT during this period, and calculates the average value e_avg of the absolute temperature difference for the most recent preset number of times (e.g., the most recent 5 times or 5 sampling periods).

[0075] The temperature change trend r can be used to determine whether the temperature rise is too slow, too fast, or a steady-state trend.

[0076] Temperature change trends can be represented by the slope of linear regression, the temperature difference between the first and last temperatures, or the temperature derivative after filtering.

[0077] Temperature fluctuations can be represented by standard deviation, range, or root mean square deviation.

[0078] A sample is considered a valid stable sample only if it simultaneously meets the conditions of Formula 4 at the current sampling time: |r| < first judgment threshold, e_avg < second judgment threshold, σT < third judgment threshold (Formula 4) The first determination threshold is, for example, 0.05°C / min, the second determination threshold is, for example, 0.04°C, and the third determination threshold is, for example, 0.03°C.

[0079] The above conditions must be met continuously for a fourth preset duration (e.g., 30 seconds or 30 1-second sampling periods) before a steady state is confirmed; if the conditions are not met at any intermediate sampling time, the continuous counting restarts. Thus, if the temperature has only passed T_k and is still slowly rising, or if the average value is close but fluctuates significantly, a steady state will not be determined prematurely.

[0080] In this embodiment, the steady-state driving value is learned only after the temperature trend, average error, and fluctuation all meet the requirements. After stabilization, the number of DAC rewrites is reduced, which helps to reduce temperature swings near the inflection point.

[0081] In one embodiment, the dynamic adjustment of heating power based on subsequent temperature differences at different sampling times includes: During the rapid heating phase when the startup state is cold start and the subsequent temperature difference is greater than the third temperature threshold, the temperature change is determined once every second preset time interval. If the temperature change is less than the first change threshold, the heating range is increased according to the temperature rise gap. If the temperature change exceeds the second threshold, then reduce the heating amplitude. Specifically, during a cold start and when the subsequent temperature difference e[n] > the third temperature threshold (e.g., 6°C), the system is in a rapid heating phase. This phase cannot be judged solely by the temperature difference, as the furnace body may have just begun heating or may have already accumulated significant heat at the same temperature difference. The controller compares the temperature every second preset time interval (e.g., 2 seconds): if the temperature rise within the second preset time interval (e.g., 2 seconds) is insufficient to reach the first change threshold (e.g., 0.1°C), the DAC value is reduced based on the temperature rise gap to increase the heating amplitude. In one specific embodiment, the single increase in heating amplitude can be limited to 37 to 224 drive units.

[0082] If the temperature rises beyond the second change threshold (e.g., 0.8°C) within the second preset duration (e.g., 2 seconds), the DAC value is increased to reduce the heating amplitude, thus reducing the heating earlier. For example, the DAC value is increased by 224 drive units to reduce the heating power.

[0083] Once the sampling temperature approaches T_k, the allowable heating rate gradually decreases. Heating is only supplemented if the actual heating rate falls below the target rate for the current stage; if the temperature has already exceeded T_k and is decreasing very slowly, the DAC value is increased further while heating is reduced. The purpose of this is to initiate braking before reaching the inflection point, rather than waiting until the temperature has passed the inflection point before processing.

[0084] In this embodiment, the temperature difference and the actual heating rate are observed simultaneously during the heating process, and the control action is gradually reduced before the inflection point, which can reduce the overshoot caused by thermal inertia.

[0085] In one embodiment, the method further includes: After a cold start, the isothermal crystal oscillator is rapidly heated. The sampling temperature during the rapid heating process is predicted using an individual thermal model. The heating power is adjusted according to the predicted sampling temperature to ensure that the final temperature during the rapid heating phase does not exceed the target inflection point temperature or that the overshoot is minimized.

[0086] In one embodiment, the method further includes: During the factory calibration phase, the isothermal crystal oscillator is heated at least once, and time-series data is acquired during each heating process. The acquired data includes: ambient temperature Ta, furnace temperature T, heating power P, crystal oscillator frequency, and optional voltage control voltage Vtune. The correlation coefficients of the isothermal crystal oscillator are determined based on the collected data during the heating process. These correlation coefficients include thermal delay, thermal inertia, natural heat dissipation coefficient, heating hysteresis, and model coefficients for frequency variation with temperature and rate of temperature change.

[0087] In one embodiment, the method further includes: The sampling temperature at multiple future sampling times is predicted using an individual thermal model. The model parameters of the individual thermal model include the aforementioned correlation coefficients. If it is predicted that the temperature will exceed the inflection point due to thermal inertia after high-power heating is stopped at the first sampling time, the heating power will be reduced in advance at the second sampling time before the current temperature reaches the inflection point temperature. The second sampling time is before the first sampling time, and it is predicted that the temperature will not exceed the inflection point temperature after high-power heating is stopped at the second sampling time due to thermal inertia.

[0088] In one embodiment, the method further includes: In the near-end fine-tuning phase where the absolute value of the subsequent temperature difference is less than the fourth temperature threshold, if the absolute value of the subsequent temperature difference is less than the fifth temperature threshold, the driving value will no longer be updated. The near-end fine-tuning phase occurs before entering steady state and after the rapid heating phase. If the absolute value of the subsequent temperature difference is between the fifth and sixth temperature thresholds, then at most q drive units will be changed in each control cycle. If the absolute value of the subsequent temperature difference is between the sixth and seventh temperature thresholds, the cumulative temperature difference of the control cycle is calculated. If the cumulative temperature difference is greater than the first temperature difference threshold, the drive value is corrected once. The initial value of the driving value is the historical steady-state driving value.

[0089] Specifically, if the absolute value of the subsequent temperature difference e[n], |e[n]|, is less than the fourth temperature threshold (e.g., 1.25°C), then the near-end fine-tuning stage begins. If the absolute value of the subsequent temperature difference e[n], |e[n]|, is less than the fifth temperature threshold (e.g., 0.015°C), then the drive value is no longer rewritten; if the absolute value of the subsequent temperature difference e[n], |e[n]|, is between the fifth and fourth temperature thresholds, then at most one drive unit is changed per control cycle; if the absolute value of the subsequent temperature difference e[n], |e[n]|, is between the sixth (e.g., 0.05°C) and the seventh (e.g., 0.5°C) temperature thresholds, then continuous adjustment according to the sampling time is not required, but the temperature difference is first accumulated according to Formula 5: EΣ[n] = EΣ[n-1] + e[n] Formula 5 When the cumulative temperature difference |EΣ[n]| > the first temperature difference threshold (e.g., 0.5°C·times), the driving value correction Δu is calculated according to Formula 6: Δu = clip(-K_e · EΣ[n],-22,22) Formula 6 EΣ[n] represents the temperature difference accumulated over the control cycle. K_e is a constant. In one specific embodiment, sampling is performed once per second, so when the first temperature difference threshold is 0.5°C·times, it represents the cumulative amount of several temperature difference samples, not the instantaneous temperature difference reaching 0.5°C. Only after the cumulative amount exceeds the first temperature difference threshold is the drive value corrected once according to the drive value correction amount, with a single correction amount not exceeding 22 drive units, and then the cumulative amount is cleared to zero.

[0090] When the temperature is too low, decrease Δu to increase heating; when the temperature is too high, increase Δu to decrease heating. If the temperature is still rapidly approaching T_k, heating should be temporarily suspended even if a positive temperature difference exists.

[0091] Δu is a digital value for heating drive, ranging from 0 to 65535.

[0092] In one embodiment, the method further includes: After reaching steady state, if the subsequent temperature difference is less than the second temperature difference threshold, the driving value remains unchanged; If the sampled temperature is higher than the target temperature for temperature control, the driving value will be reduced by a certain number of driving units based on the magnitude of the subsequent temperature difference. If the sampled temperature is lower than the target temperature for temperature control, the driving value will be increased by several driving units based on the magnitude of the subsequent temperature difference.

[0093] Specifically, after reaching steady state, the second-level control continues to sample the temperature, and can promptly handle situations exceeding the allowable range; under normal circumstances, it switches to a small correction every 300 seconds. For example, if the subsequent temperature difference e[n] is less than the second temperature difference threshold (e.g., 0.0156°C), the driving value remains unchanged. If the subsequent temperature difference |e[n]| reaches or exceeds 0.0156°C, the driving value is slightly corrected within a 300-second inertial adjustment period, depending on whether the temperature is too low or too high.

[0094] When the temperature is too low, the heating amount is increased by, for example, one drive unit; when the temperature is too high, the heating amount is decreased by, for example, one to four drive units, according to the subsequent temperature difference e[n]. This avoids modifying the drive value back and forth every second near the steady state.

[0095] After reaching a steady state, the drive value is adjusted every five preset intervals (e.g., 300 seconds).

[0096] In one embodiment, the method further includes: After obtaining the latest steady-state driving value, the absolute temperature difference is calculated multiple times, and the average absolute temperature difference is obtained. If the mean absolute temperature difference is greater than the absolute temperature difference threshold, the lock on the latest steady-state driving value is released, a new learning window is started, and a new steady-state driving value is selected.

[0097] Specifically, the latest steady-state drive value U_k is not permanently locked. The controller calculates the absolute temperature difference every second and averages, for example, the last 5 sampled values. When the average value exceeds the absolute temperature difference threshold (e.g., 0.03°C), it indicates that the original maintenance drive value is no longer suitable for the current heat dissipation conditions. At this point, the lock is released, and a short learning window of, for example, 3 minutes is initiated. While the near-end control continues to operate, a new steady-state drive value U_k is selected, and a new steady-state drive value U_k is determined after the window ends. In this way, parameters are updated only when there is a continuous deviation, and a single temperature noise will not trigger a rewrite.

[0098] This embodiment relearns steady-state driving values, which can better adapt to environmental changes.

[0099] The method also includes: within the learning window, if the sampled temperature exceeds the temperature control target temperature (inflection point temperature) and the temperature exceeds the third temperature difference threshold (e.g., 2°C), heating is stopped; if the sampled temperature falls back below the temperature control target temperature (inflection point temperature), the learning window continues.

[0100] In one embodiment, the method further includes: If a valid temperature is not obtained after p consecutive retries within the current sampling period at the current sampling time, it is determined that the temperature sampling at the current sampling time has failed, and the valid sampling temperature at the previous sampling time is used as the sampling temperature at the current sampling time. And / or, If the sampling temperature for the sixth consecutive preset duration is higher than the eighth temperature threshold, then the minimum heating strategy is executed, temperature control is stopped, and the power supply to the constant temperature crystal oscillator is cut off.

[0101] Specifically, when temperature readings occasionally fail, the power is not immediately cut off. Instead, the system retryes p times and temporarily uses the previous valid temperature. For example, if a valid temperature is not obtained after three consecutive retries during the current sampling period or sampling time, the control temperature will be maintained at the previous valid temperature. Alternatively, the valid temperature flag can be set to invalid.

[0102] If the eighth temperature threshold (e.g., 100°C) or a higher effective temperature is read for the sixth consecutive preset duration (e.g., 3 seconds), the drive is set to minimum heating (i.e., the DAC output is forced to 65535 to minimize the heating power), temperature control is stopped, and the power supply to the OCXO is cut off.

[0103] If the sampling temperature exceeds T_k by 2°C during the learning process, heating should be stopped first, and then resumed only after the temperature drops back below T_k.

[0104] The following example uses a cold start to illustrate how the various stages are connected. In this embodiment, the temperature is read every second, using a 10-point moving average and a 16-bit reverse heating drive. The output of a cryogenic crystal oscillator (OCXO) serves as the common reference for the PLL, and the subsequent stage generates four independent frequency outputs: Startup classification: Cold start e>15°C; Hot start -3°C≤e≤15°C; Overheat start e<-3°C. Function: Determines initial heating, adaptive trend, or initial cooling.

[0105] Rapid heating zone: e>6°C; Function: Combines the 2-second heating rate to implement positive or negative power correction.

[0106] Proximal control window: |e|<1.25°C; Function: Switches from rate control to dead zone, cumulative error, and small step correction.

[0107] Proximal fine-tuning: |e|<0.015°C; 0.015°C≤|e|≤0.05°C; 0.05°C<|e|<0.5°C; Function: Maintain within dead zone; single-step zone adjusts by a maximum of 1 unit per cycle; cumulative zone is corrected after reaching the threshold.

[0108] Steady-state history: 60 seconds; Function: Calculate regression slope, standard deviation, and temperature range.

[0109] Steady-state duration: 30 seconds; Function: To prevent brief passage through the target temperature from being mistakenly identified as a steady state.

[0110] Inertial adjustment period: 300 seconds; function: to reduce the density of active control after steady state.

[0111] Initial learning window: Enters 180 seconds after startup and lasts for 30 minutes; Function: Selects U_k at the end of the longest continuous stable segment.

[0112] Drift detection: 5-point average absolute error > 0.03°C; Function: Unlock and start 3 minutes of relearning.

[0113] Over-temperature protection: Effective temperature reaches 100°C for 3 consecutive seconds; Function: Minimize heating and cut off related power supply.

[0114] Assuming the startup temperature is more than 15°C lower than T_k, and there is no usable U_k in memory, the device first warms up using a cold start method. If the temperature difference is still greater than 6°C, the actual temperature rise is checked every 2 seconds; as the temperature gradually approaches T_k, the allowable temperature rise rate and the single DAC change decrease together. After entering the 1.25°C range, it enters near-end fine-tuning, first passing through the 0.015°C dead zone, the 0.05°C single-step zone, and the error accumulation zone. The system is considered stable only after the temperature trend, average error, and fluctuation meet the conditions for 30 consecutive seconds. After running for 180 seconds, it begins learning U_k, and at the end of 30 minutes, it saves the drive value corresponding to the longest continuous stable period. Afterwards, if the average error for the most recent 5 seconds continuously exceeds 0.03°C, it relearns using a 3-minute window.

[0115] refer to Figure 2 This application also provides a staged temperature control system for a thermostatic crystal oscillator, the system comprising: a thermostatic crystal oscillator, a temperature sampling device, a heating actuator, and a controller; A temperature sampling device is used to sample the temperature of the thermostatic crystal oscillator and provide the sampled temperature to the controller; A heating actuator is used to heat a thermostatic crystal oscillator under the control of a controller. A controller for executing the staged temperature control method for the isothermal crystal oscillator described in any of the above-mentioned methods.

[0116] Specifically, the working principle of the controller is described in the above description of the staged temperature control method of the isothermal crystal oscillator, and will not be repeated here.

[0117] The heating actuator is digitally driven and controlled by the controller.

[0118] This application stores the inflection point temperature T_k and steady-state drive value U_k for each oven-controlled crystal oscillator (OCXO). The inflection point temperature T_k is used as the temperature control target, and U_k is used for power-on feedforward and subsequent learning. Based on the validity of U_k and the temperature difference between U_k and T_k at power-on, cold start, hot start, overheat start, or start-up from near U_k is selected. When far from the inflection point, heating is adjusted using both the temperature difference and the actual heating rate. As the inflection point approaches, the target heating rate and the single-drive change are gradually reduced. A dead zone, a single-step adjustment zone, and an error accumulation zone are set near the inflection point, and heating is paused when the temperature is still rapidly approaching the target. Simultaneously, the temperature trend, average temperature difference, temperature fluctuation, and continuous satisfaction time are checked. After confirming stability, the adjustment frequency is reduced and a new steady-state drive value is learned. The drive value corresponding to the longest continuous stable time is saved within a limited time. After locking, the short-term average temperature difference determines whether to relearn. This application can reduce inflection point temperature overshoot caused by thermal inertia; avoid miscorrection of temperature rise drift; and meet the requirements of rapid temperature rise, low overshoot, low frequency deviation and steady-state low jitter of isothermal crystal oscillators.

[0119] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a non-volatile computer-readable storage medium. When executed, the program can include the processes of the embodiments described above, specifically including steps such as temperature sampling, sliding averaging, startup state determination, staged DAC adjustment (temperature adjustment or heating adjustment), steady-state detection, steady-state drive value learning, and over-temperature protection. Any references to memory, storage, database, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0120] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0121] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A staged temperature control method for a thermostatic crystal oscillator, characterized in that, The method includes: Obtain the inflection point temperature of the cryogenic crystal oscillator and the historical steady-state drive value actually used when the oscillator was last stable within the target temperature window near the inflection point temperature; Temperature sampling is performed on the isothermal crystal oscillator, and the sampled temperature is averaged at each sampling time to obtain the average temperature at the corresponding sampling time. The average temperature includes the power-on average temperature. The inflection point temperature is used as the target temperature for temperature control. The difference between the target temperature and the average temperature is calculated to obtain the subsequent temperature difference at the corresponding sampling time. The subsequent temperature difference includes the start-up temperature difference. Based on the validity of historical steady-state driving values, inflection point temperature, startup temperature, reference room temperature, average startup temperature, and startup temperature difference, execute the corresponding startup strategy. The heating power is dynamically adjusted based on the subsequent temperature difference at different sampling times. The step of executing a corresponding startup strategy based on the validity of historical steady-state driving values, inflection point temperature, startup temperature, reference room temperature, average startup temperature, and startup temperature difference includes: If the historical steady-state driving values ​​are valid, then the heat demand ratio is calculated based on the inflection point temperature, start-up temperature, and reference room temperature. If the heat demand ratio indicates that the isothermal crystal oscillator is in a cold state, the initial drive is to increase heating based on the historical steady-state drive value. If the heat demand ratio indicates that the isothermal crystal oscillator is in a heat engine state, then the initial drive is the historical steady-state drive value. If the historical steady-state driving value is invalid, the difference between the target temperature and the average start-up temperature is used to obtain the start-up temperature difference. The startup state of the cryogenic crystal oscillator is determined based on the start-up temperature difference. The startup state can be one of three: cold start, warm start, or overheat start. If the startup state is cold start, then the heating drive value will be set within the preset cold start power range for heating. If the startup state is a hot start, heating will begin from the historical steady-state driving value. If the startup state is an overheat start, reduce heating and wait for the temperature to drop before entering control near the inflection point.

2. The method according to claim 1, characterized in that, The method further includes: If the sampled temperature enters the target window, determine whether the sampled temperature has entered a steady state within the target window; If the sampling temperature enters a steady state, the steady-state driving value required to maintain the inflection point temperature is learned within the learning window. The candidate steady-state driving values ​​of the longest stable segment within the entire learning window are retained and updated to the latest steady-state driving values. The longest stable segment is the segment with the longest continuous stable time. If there is no continuous stable segment within the entire learning window, the driving value being used at the end of the learning window is retained as the latest steady-state driving value.

3. The method according to claim 1 or 2, characterized in that, The step of determining the start-up state of the thermostatic crystal oscillator based on the start-up temperature difference includes: If the temperature difference at startup exceeds the first temperature threshold, the startup state of the cryogenic crystal oscillator is determined to be a cold start. If the temperature difference at startup is not less than the second temperature threshold and not greater than the first temperature threshold, then the startup state of the isothermal crystal oscillator is determined to be a hot start. If the temperature difference at startup is less than the second temperature threshold, then the startup state of the thermostatic crystal oscillator is determined to be overheat startup.

4. The method according to claim 2, characterized in that, The determination of whether the sampled temperature within the target window has entered a steady state includes: Calculate the temperature change trend and temperature fluctuation within the most recent third preset time period corresponding to each sampling time in the target window, and calculate the average value of the absolute temperature difference of the most recent preset number of times, where the absolute temperature difference is the absolute value of the difference between the temperature control target temperature and the corresponding average temperature. If the absolute value, temperature fluctuation, and average value of the temperature change trend for the fourth consecutive preset duration within the target window are all less than the corresponding judgment threshold, then the sampling temperature is determined to have entered a steady state.

5. The method according to claim 1, characterized in that, The method of dynamically adjusting the heating power based on the subsequent temperature difference at different sampling times includes: During the rapid heating phase when the startup state is cold start and the subsequent temperature difference is greater than the third temperature threshold, the temperature change is determined once every second preset time interval. If the temperature change is less than the first change threshold, the heating range is increased according to the temperature rise gap. If the temperature change exceeds the second threshold, the heating amplitude should be reduced.

6. The method according to claim 5, characterized in that, The method further includes: In the near-end fine-tuning phase where the absolute value of the subsequent temperature difference is less than the fourth temperature threshold, if the absolute value of the subsequent temperature difference is less than the fifth temperature threshold, the driving value will no longer be updated. The near-end fine-tuning phase occurs before entering steady state and after the rapid heating phase. If the absolute value of the subsequent temperature difference is between the fifth and sixth temperature thresholds, then at most q drive units will be changed in each control cycle. If the absolute value of the subsequent temperature difference is between the sixth and seventh temperature thresholds, the cumulative temperature difference of the control cycle is calculated. If the cumulative temperature difference is greater than the first temperature difference threshold, the drive value is corrected once. The initial value of the driving value is the historical steady-state driving value.

7. The method according to claim 2 or 4, characterized in that, The method further includes: After obtaining the latest steady-state driving value, the absolute temperature difference is calculated multiple times, and the average absolute temperature difference is obtained. If the mean absolute temperature difference is greater than the absolute temperature difference threshold, the lock on the latest steady-state driving value is released, a new learning window is started, and a new steady-state driving value is selected.

8. The method according to claim 1, characterized in that, The method further includes: If a valid temperature is not obtained after p consecutive retries within the current sampling period at the current sampling time, it is determined that the temperature sampling at the current sampling time has failed, and the valid sampling temperature at the previous sampling time is used as the sampling temperature at the current sampling time. And / or, If the sampling temperature for the sixth consecutive preset duration is higher than the eighth temperature threshold, then the minimum heating strategy is executed, temperature control is stopped, and the power supply to the constant temperature crystal oscillator is cut off.

9. A staged temperature control system for a thermostatic crystal oscillator, characterized in that, The system includes: a thermostatic crystal oscillator, a temperature sampling device, a heating actuator, and a controller; A temperature sampling device is used to sample the temperature of the thermostatic crystal oscillator and provide the sampled temperature to the controller; A heating actuator is used to heat a thermostatic crystal oscillator under the control of a controller. A controller for executing the staged temperature control method for the isothermal crystal oscillator according to any one of claims 1-8.

Images