An adaptive loop calibration algorithm, control device and storage medium for an industrial high-voltage multiphase Buck PMIC chip

Abstract

The application discloses an industrial high-voltage multi-phase Buck PMIC chip adaptive loop calibration algorithm, a control device and a storage medium. All functions of the scheme are integrated in the PMIC chip, relying on on-chip ADC to collect operation parameters, calculating load characteristics, and then combining a solidified loop model to adaptively set the coefficient of a digital compensator, and synchronously optimizing the multi-phase PWM phase; every 100 ms, a small amplitude step disturbance is injected to detect the loop stability margin online, and when the phase margin is lower than 45 DEG or the gain margin is lower than 10 dB, the self-healing re-calibration is automatically started. The application is based on a hardware chip landing, is different from a traditional fixed parameter compensation scheme, optimizes power efficiency, adapts to industrial high-voltage complex working conditions, reduces performance degradation caused by device temperature drift and process deviation, and meets the energy-saving and environment-friendly application direction.

Description

Technical Field

[0001] This invention relates to the field of power management integrated circuits (PMICs), and more particularly to an adaptive loop calibration algorithm, control device, and storage medium for industrial high-voltage multiphase Buck power management chips. Specifically, this invention aims to provide an on-chip embedded algorithm capable of dynamically adjusting loop compensation parameters according to load changes and environmental conditions to optimize power conversion efficiency and transient response performance. Background Technology

[0002] With the increasing demands for power management integrated circuits (PMICs) in fields such as industrial automation, new energy, and data centers, high-voltage, high-current, and high-efficiency multiphase buck converters have become the mainstream solution. Multiphase buck converters effectively reduce the current stress on individual power stages, decrease output ripple, and improve transient response speed by distributing the total current across multiple parallel power stages. However, in industrial high-voltage applications, PMICs face challenges such as wide-range input voltage fluctuations, drastic dynamic changes in load current, and a wide operating temperature range, making loop stability and performance optimization of multiphase buck converters particularly complex.

[0003] Traditional Buck converter loop compensation designs typically employ analog compensators with fixed parameters (such as Type-II or Type-III), whose compensation parameters are set during the design phase based on worst-case or typical operating conditions. This fixed compensation scheme struggles to maintain optimal loop performance across the entire operating range (e.g., when input voltage, output current, ambient temperature, and other parameters change). When input voltage or output current changes significantly, the converter's gain and phase characteristics alter, potentially leading to insufficient loop bandwidth and slow transient response; or excessive phase margin, causing oscillations or even system instability. Particularly in multiphase Buck converters, the current sharing accuracy and phase interleaving stability between phases are crucial to overall performance, and traditional fixed compensation schemes struggle to effectively handle these dynamic changes.

[0004] Furthermore, the parameters of the PMIC chip and its peripheral components (such as inductors and capacitors) drift due to manufacturing process deviations, temperature changes, and long-term aging. These parameter drifts further deteriorate loop performance, rendering the initial calibration compensation parameters inapplicable, leading to decreased power supply output voltage accuracy, worsened transient response, and even system failure. Currently, mass-produced industrial-grade high-voltage multiphase PMICs all rely on factory-fixed compensation parameters, lacking on-chip online adaptive self-healing calibration capabilities. Post-production manual offline debugging and calibration are required, resulting in high development and maintenance costs. Therefore, there is an urgent need in existing technologies for an adaptive loop calibration algorithm for industrial high-voltage multiphase Buck PMIC chips that can sense the operating status in real time, adaptively adjust loop compensation parameters, and effectively address device parameter drift, ensuring that the power system maintains excellent stability, efficiency, and transient response performance under various complex operating conditions. Summary of the Invention

[0005] This invention aims to solve the technical problems of existing industrial high-voltage multiphase Buck PMIC chips, such as difficulty in adaptively adjusting loop compensation parameters under complex operating conditions, poor transient response performance, and susceptibility to device parameter drift. This invention provides an adaptive loop calibration algorithm, control device, and storage medium for industrial high-voltage multiphase Buck PMIC chips. Through on-chip embedded algorithms, it achieves dynamic and adaptive adjustment of loop compensation parameters, thereby significantly improving power conversion efficiency, transient response speed, and system stability.

[0006] To achieve the above objectives, the present invention provides the following technical solution: An adaptive loop calibration algorithm for an industrial high-voltage multiphase Buck PMIC chip, characterized by the following steps: 1. Real-time status parameter acquisition: Key operating parameters such as input voltage (Vin), output voltage (Vout), inductor current of each phase (ILx), output load current (Iload), and internal chip temperature are acquired in real time within the multi-phase Buck PMIC chip. These parameters are digitized by an on-chip 12-bit high-precision ADC and used as input to the algorithm, without the need for external sampling chips to assist in acquisition.

[0007] 2. Load Characteristic Parameter Calculation: Based on the real-time acquired output load current, the characteristic parameters of the current load are calculated, mainly including the load current change rate (dIload / dt) and the output voltage ripple component. The load current change rate reflects the severity of load abrupt changes, while the output voltage ripple component reflects the loop stability and transient response.

[0008] 3. Dynamic Compensation Coefficient Calculation: Based on the calculated load characteristic parameters and the preset loop model (which is permanently stored in the on-chip non-volatile memory and obtained through simulation or experimentation during the chip design phase), the required compensation coefficients for the on-chip integrated digital PID or zero-pole type digital compensator are dynamically calculated. This calculation process is implemented using lookup tables, linear interpolation, or first-order simplified model predictive control (MPC) algorithms, aiming to achieve optimal gain and phase margins for the loop under the current operating conditions.

[0009] 4. Multi-phase PWM signal adaptive adjustment: Dynamically calculated compensation coefficients are applied to the digital compensator to generate optimized PWM control signals. Simultaneously, based on real-time monitoring of the inductor current in each phase, the phase interleaving angle of each phase PWM signal is automatically adjusted to ensure current sharing accuracy and phase balance during load changes, effectively suppressing output ripple and improving transient response. The entire calculation is completed in the on-chip digital logic of the PMIC, requiring no external MCU.

[0010] 5. Loop Health Monitoring and Self-Healing Calibration: A small step disturbance signal with an amplitude of 1%~3% of the rated output voltage is injected into the loop at a fixed 100ms cycle. This does not affect the normal load operation of the power supply. The output voltage response is analyzed, and the loop's gain margin (GM) and phase margin (PM) are identified online. Preset safety thresholds are set: a lower limit of 45° for phase margin and a lower limit of 10dB for gain margin. When GM or PM is detected to be below these thresholds, a self-healing calibration mechanism is triggered, and steps 3 and 4 are re-executed to correct the compensation coefficient and phase interleaving angle, ensuring the loop remains in a stable and optimal operating state. This calibration process can be automatically triggered during chip startup, load surges, or periodic maintenance.

[0011] This invention also provides a control device for an industrial high-voltage multiphase Buck PMIC chip, characterized in that it comprises: 6. Data Acquisition Unit: Integrated inside the PMIC chip, equipped with a 12-bit on-chip ADC, used to acquire operating parameters such as input voltage, output voltage, inductor current of each phase, output load current, and internal chip temperature in real time.

[0012] 7 Load Characteristic Calculation Unit: Composed of on-chip digital logic circuits, used to calculate the load current change rate and output voltage ripple component based on the acquired load current.

[0013] 8. Compensation Coefficient Calculation Unit: Internally equipped with a fixed lookup table and first-order simplified MPC operation logic, it is used to dynamically calculate the compensation coefficient of the digital compensator based on the load characteristic parameters and the preset loop model. The parameter data table is stored in the on-chip non-volatile memory area.

[0014] 9. PWM Generation and Phase Adjustment Unit: Used to generate PWM control signals based on compensation coefficients and adjust the phase interleaving angle of the PWM signals according to the inductor current of each phase.

[0015] 10. Loop Health Monitoring Unit: Used to inject a step disturbance signal of 1%~3% of the rated output every 100ms, analyze the output response, and identify the loop gain margin and phase margin online.

[0016] 11. Self-healing calibration control unit: used to trigger the recalculation and adjustment of compensation coefficients and phase interleaving angle when the loop phase margin is <45° or the gain margin is <10dB.

[0017] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the above-described adaptive loop calibration algorithm.

[0018] The beneficial effects of this invention are as follows: 12 Significantly improved transient response performance: By adaptively adjusting loop compensation parameters, the PMIC can quickly respond to load changes, effectively suppress output voltage overshoot and undershoot, and ensure output voltage stability.

[0019] 13. Optimize power conversion efficiency: Under different operating conditions, the algorithm can ensure that the loop always works in the optimal state, reduce energy loss, conform to the direction of energy conservation and environmental protection industry, and reduce power consumption of industrial equipment.

[0020] 14. Enhanced system stability and robustness: The self-healing loop health monitoring and calibration mechanism effectively addresses device parameter drift and environmental changes, eliminating the need for manual calibration after delivery and ensuring the stability and reliability of the PMIC during long-term operation.

[0021] 15. Achieving on-chip embedded intelligence: The algorithm is implemented entirely within the PMIC chip, without the need for external MCU intervention, which reduces system complexity and material costs, and improves integration.

[0022] 16. Improve multiphase current sharing accuracy: By adjusting the PWM phase interleaving angle in real time, the current sharing performance of the multiphase Buck converter under dynamic load is effectively improved, further reducing output ripple. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the system architecture of the industrial high-voltage multiphase Buck PMIC chip control device of the present invention; Figure 2 This is a schematic diagram of the adaptive loop calibration algorithm of the present invention; Figure 3 This is a schematic diagram of the digital compensator structure of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the specific embodiments of this invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it.

[0025] 1. System Architecture Please see Figure 1 The present invention provides an adaptive loop calibration control device for an industrial high-voltage multiphase Buck PMIC chip, which mainly includes: a data acquisition unit 101, a load characteristic calculation unit 102, a compensation coefficient calculation unit 103, a PWM generation and phase adjustment unit 104, a loop health monitoring unit 105, and a self-healing calibration control unit 106.

[0026] • Data Acquisition Unit 101: This unit is integrated inside the PMIC chip, equipped with a 12-bit on-chip ADC and a fixed sampling frequency of 50kHz. It is responsible for real-time monitoring and digitizing the key operating parameters of the Buck converter.

[0027] ◦ Input voltage (Vin): Sampled via a high-voltage divider resistor network and an ADC.

[0028] ◦ Output voltage (Vout): Sampled through a voltage divider network and an ADC, serving as the loop feedback signal.

[0029] ◦ Inductor current of each phase (ILx): Sampled by an on-chip integrated current sampling circuit in conjunction with an ADC, used for current sharing control and loop health assessment.

[0030] ◦ Output load current (Iload): Sampled by the total output current sampling circuit and ADC, used for load characteristic calculation.

[0031] • Internal chip temperature: Sampled by an on-chip temperature sensor for temperature compensation.

[0032] • Load characteristic calculation unit 102: Receives real-time output load current data provided by data acquisition unit 101 and processes it to extract load characteristic parameters. The main calculations include: ◦ Load current change rate (dIload / dt): This is calculated by dividing the difference in load current between consecutive sampling points by the sampling time interval. This parameter is a key indicator for judging the degree of load change.

[0033] ◦ Output voltage ripple component: The AC component of the output voltage signal is extracted by high-pass filtering to evaluate loop stability and transient response.

[0034] • Compensation coefficient calculation unit 103: This unit is the core of adaptive loop calibration. It receives the load characteristic parameters output by the load characteristic calculation unit 102 and, combined with the loop model data pre-stored in the on-chip non-volatile memory, dynamically calculates the compensation coefficients required by the digital compensator. The loop model data is obtained through offline chip calibration and describes the frequency response characteristics of the Buck converter under different operating conditions. The compensation coefficients are calculated using one or more of the following methods: ◦ Lookup Table (LUT): Based on the load current change rate and output voltage ripple component, the optimal compensation coefficient is directly retrieved from the preset lookup table.

[0035] ◦ Interpolation algorithm: Perform linear interpolation on the data in the lookup table to obtain refined compensation parameters.

[0036] ◦ First-order simplified model predictive control (MPC) algorithm: Based on the first-order simplified Buck state-space model, predict the output voltage change and solve for the optimal compensation coefficient.

[0037] • PWM Generation and Phase Adjustment Unit 104: This unit generates the PWM control signal required by the multiphase Buck converter based on the compensation coefficient output by the compensation coefficient calculation unit 103. Simultaneously, it is also responsible for adjusting the phase interleaving angle of each phase PWM signal in real time based on the inductor current information of each phase provided by the data acquisition unit 101. For example, by comparing the average value of the inductor current of each phase, if a current sharing deviation is found, the PWM start-up time of the corresponding phase is fine-tuned to ensure the balance of current in each phase, thereby reducing output ripple and improving transient response.

[0038] • Loop Health Monitoring Unit 105: This unit is responsible for online assessment of the health status of the Buck converter control loop. It achieves this through the following methods: ◦ Disturbance injection: A step disturbance signal of 1%~3% of the rated output voltage is periodically injected into the loop every 100ms. The disturbance amplitude is controlled and will not affect the normal power supply of the whole machine.

[0039] ◦ Response Analysis: Analyze the response of the output voltage to the injected disturbance signal and identify the loop gain margin (GM) and phase margin (PM) in real time.

[0040] ◦ Threshold comparison: Preset judgment thresholds: phase margin 45°, gain margin 10dB. If the values ​​are lower than the thresholds, the loop performance is judged to be degraded.

[0041] • Self-healing calibration control unit 106: When the loop health monitoring unit 105 detects that GM and PM are lower than the preset threshold, or when the chip is powered on and initialized or the load changes significantly, it triggers the entire link to recalculate the compensation coefficient and PWM phase to ensure the loop operates optimally for a long time.

[0042] 2. Algorithm Flow Please see Figure 2 This invention provides an adaptive loop calibration algorithm for industrial high-voltage multiphase Buck PMIC chips, which mainly includes the following steps: Step S201: Initialization and parameter setting.

[0043] After the PMIC chip is powered on, the system initializes by setting the ADC sampling frequency to 50kHz, the PWM switching frequency, the pre-stored loop model parameters, the GM / PM threshold, and the self-healing calibration trigger conditions.

[0044] Step S202: Real-time status parameter acquisition.

[0045] The data acquisition unit 101 acquires the input voltage Vin, output voltage Vout, inductor current ILx of each phase, output load current Iload and chip temperature of the Buck converter in real time. The analog signals are converted into digital signals by the on-chip 12-bit ADC.

[0046] Step S203: Calculation of load characteristic parameters.

[0047] The load characteristic calculation unit 102 calculates the current load current change rate dIload / dt and the output voltage ripple component based on the collected output load current Iload.

[0048] Step S204: Calculation of dynamic compensation coefficient.

[0049] The compensation coefficient calculation unit 103, combined with the pre-stored loop model, dynamically solves the digital compensator parameters through any of the following methods: table lookup, interpolation, or first-order simplified MPC.

[0050] Step S205: Generation and phase adjustment of multiphase PWM signals.

[0051] The PWM generation and phase adjustment unit 104 generates a PWM waveform based on the compensation coefficient and corrects the phase crossover angle in real time by combining the inductor current of each phase.

[0052] Step S206: Loop health monitoring.

[0053] The loop health monitoring unit 105 injects a small step disturbance every 100ms, collects the output waveform, and calculates the gain margin and phase margin in real time.

[0054] Step S207: Determine whether self-healing calibration is required.

[0055] The self-healing calibration control unit 106 compares the measured margin with the threshold. If the phase margin is <45° or the gain margin is <10dB, it jumps to S204 to readjust the parameters; otherwise, it continues to collect data in a loop.

[0056] This invention is not limited to the above-described embodiments. Any simple modifications or equivalent substitutions to the technical solutions of this invention fall within the protection scope of this invention.

Claims

1. An adaptive loop calibration algorithm for an industrial high-voltage multiphase Buck PMIC chip, characterized in that, Includes the following steps: a. Real-time status parameter acquisition: Inside the PMIC chip, the input voltage, output voltage, inductor current of each phase, output load current and internal temperature of the multiphase Buck converter are acquired in real time through an on-chip integrated 12-bit ADC. All acquisition devices are integrated inside the chip, and there is no external acquisition hardware. b. Load characteristic parameter calculation: Based on the real-time acquired output load current, calculate the current load current change rate and output voltage ripple component; c. Dynamic compensation coefficient calculation: Based on the calculated load current change rate and output voltage ripple component, as well as the preset loop model stored in the on-chip non-volatile memory, the compensation coefficients required by the on-chip integrated digital PID or zero-pole type digital compensator are dynamically calculated; the coefficients are solved using lookup tables, linear interpolation, or first-order simplified model predictive control (MPC) algorithms; the entire operation is completed in the on-chip logic without the need for an external MCU; d. Multiphase PWM signal adaptive adjustment: The compensation coefficients obtained by dynamic calculation are applied to the digital compensator to generate an optimized PWM control signal; at the same time, the phase interleaving angle of each phase PWM signal is automatically adjusted according to the real-time collected inductor current of each phase. e. Loop health monitoring and self-healing calibration: Inject a step disturbance signal with an amplitude of 1% to 3% of the rated output voltage into the loop every 100ms, analyze the response of the output voltage, and identify the gain margin (GM) and phase margin (PM) of the loop online; when the phase margin is lower than 45° or the gain margin is lower than 10dB, trigger the self-healing calibration mechanism and re-execute steps c and d.

2. The algorithm according to claim 1, characterized in that, The self-healing calibration mechanism is automatically triggered during the PMIC chip startup phase, load mutation phase, or periodic maintenance phase.

3. A control device for an industrial high-voltage multiphase Buck PMIC chip, characterized in that, include: a. Data acquisition unit: Integrated into the PMIC chip, with a built-in 12-bit ADC and a sampling frequency of 50kHz, used to acquire the input voltage, output voltage, inductor current of each phase, output load current, and internal temperature of the multiphase Buck converter in real time. b. Load characteristic calculation unit: composed of on-chip digital logic, used to calculate the current load current change rate and output voltage ripple component based on the real-time acquired output load current; c. Compensation coefficient calculation unit: Internally integrates lookup table, linear interpolation and first-order simplified MPC operation logic. Loop model parameters are fixed in on-chip non-volatile storage area for dynamically calculating the compensation coefficients required by the digital compensator. d. PWM generation and phase adjustment unit: used to apply the dynamically calculated compensation coefficient to the digital compensator to generate an optimized PWM control signal; at the same time, based on the real-time acquired inductor current of each phase, automatically adjust the phase interleaving angle of each phase PWM signal; e. Loop health monitoring unit: configured to inject a step disturbance of 1%~3% of the rated output every 100ms to identify the gain margin (GM) and phase margin (PM) of the loop online. f. Self-healing calibration control unit: used to trigger the compensation coefficient calculation unit and the PWM generation and phase adjustment unit to re-execute the calculation of compensation coefficients and the adjustment of phase interleaving angle when the phase margin is <45° or the gain margin is <10dB.

4. The control device according to claim 3, characterized in that, The digital compensator is an on-chip integrated proportional-integral-derivative (PID) controller or a pole-zero compensator.

5. A computer-readable storage medium, characterized in that, The device contains a computer program that, when executed by a processor, implements the adaptive loop calibration algorithm for the industrial high-voltage multiphase Buck PMIC chip as described in any one of claims 1-2.

6. A computing device, characterized in that, It includes a processor and a memory, the memory storing a computer program that, when executed by the processor, implements the adaptive loop calibration algorithm for the industrial high-voltage multiphase BuckPMIC chip according to any one of claims 1-2.

Images