Intelligent management method and system for power amplifier

Abstract

The application is suitable for the field of power electronics, and provides an intelligent management method and system of a power amplifier. The method comprises: acquiring an output voltage and current of the power amplifier, and determining a real-time load impedance; predicting a junction temperature of the power amplifier according to a coupling thermal model which associates an effective thermal resistance with a load mismatch degree; and cooperatively determining and adjusting a stability compensation parameter for maintaining stable operation and a pre-distortion parameter for compensating nonlinear distortion based on the predicted junction temperature and the real-time load impedance. Through the cooperative intelligent management, the safety and reliability can be improved, the performance maximization can be realized, and the high fidelity in all working conditions can be ensured.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power electronics, and particularly relates to an intelligent management method and system of a power amplifier. BACKGROUND

[0002] High-performance underwater acoustic systems, such as those used for underwater detection and communication, require the core component, the power amplifier, to drive loads with complex and time-varying impedance characteristics, such as piezoelectric transducers, while maintaining high power, high fidelity, and high reliability. In the prior art, in order to manage the heat of the power amplifier, a thermal model is usually used to estimate the junction temperature of the power device. However, these thermal models often treat the thermal resistance of the device as a fixed value, ignoring the key physical effect that load impedance mismatch can significantly worsen heat dissipation efficiency, resulting in increased equivalent thermal resistance. Therefore, when the load impedance changes dramatically, the temperature prediction of the existing thermal model is not accurate enough, which makes the thermal protection mechanism lagging behind and may cause the power device to be damaged due to overheating.

[0003] In addition, for the control of the working stability of the amplifier and the compensation of nonlinear distortion, the prior art usually adopts isolated solutions. For example, a compensation network with fixed parameters is used, or a simple feedback adjustment is made based on temperature only. These solutions do not take into account the real-time changes in load impedance, especially the reactance characteristics of the load, and their comprehensive effects on system stability and nonlinear characteristics. This isolated and uncoupled control method of each control link leads to the following defects in the existing amplifier management system: first, the risk of thermal breakdown is high, as the additional temperature rise caused by load mismatch cannot be accurately predicted, and protection is not timely; second, the performance is limited, as the fixed power limiting strategy adopted for safety is usually too conservative, resulting in the inability to fully utilize the performance of the system under benign load conditions; and finally, the signal fidelity is reduced, as the fixed compensation parameters cannot adapt to the system characteristic drift caused by temperature and load changes, leading to the deterioration of nonlinear distortion. SUMMARY

[0004] The embodiments of the present application provide an intelligent management method and system of a power amplifier, aiming to solve the technical problems of high risk of thermal breakdown, inability to maximize performance, and reduced signal fidelity of the power amplifier due to the inability to cooperatively handle the coupling relationship among load impedance, thermal state, and nonlinear distortion in the prior art.

[0005] In a first aspect, the embodiments of the present application provide an intelligent management method of a power amplifier, comprising the following steps:

[0006] obtaining the output voltage and the output current of the power amplifier, and determining the real-time load impedance of the load driven by the power amplifier based on the output voltage and the output current;

[0007] predict a junction temperature of the power amplifier according to a coupled thermal model, wherein the coupled thermal model relates an effective thermal resistance of the power amplifier to a mismatch degree between the real-time load impedance and a preset optimal load impedance;

[0008] determine and adjust one or more stability compensation parameters for maintaining the power amplifier to operate stably based on the junction temperature of the power amplifier and the real-time load impedance;

[0009] determine and adjust one or more pre-distortion parameters for compensating for non-linear distortion of the power amplifier based on one or more reactance characteristics of the junction temperature of the power amplifier and the real-time load impedance.

[0010] In a possible implementation manner of the first aspect, in the coupled thermal model, the effective thermal resistance of the power amplifier is modeled to increase with an increase of the mismatch degree between the real-time load impedance and the preset optimal load impedance.

[0011] In a possible implementation manner of the first aspect, the predicting the junction temperature of the power amplifier according to the coupled thermal model comprises:

[0012] calculating the effective thermal resistance by using a formula containing a deterioration item proportional to a square of the mismatch degree;

[0013] determining the junction temperature of the power amplifier based on the calculated effective thermal resistance and a dissipated power of the power amplifier.

[0014] In a possible implementation manner of the first aspect, the predicting the junction temperature of the power amplifier according to the coupled thermal model comprises:

[0015] obtaining the effective thermal resistance from a pre-stored lookup table according to the mismatch degree;

[0016] determining the junction temperature of the power amplifier based on the obtained effective thermal resistance and a dissipated power of the power amplifier.

[0017] In a possible implementation manner of the first aspect, the stability compensation parameters comprise one of:

[0018] an incremental value for adjusting a variable capacitance of a physical compensation network of the power amplifier, a coefficient for adjusting a digital filter in a digital signal processing path.

[0019] In a possible implementation manner of the first aspect, the determining and adjusting one or more stability compensation parameters for maintaining the power amplifier to operate stably based on the junction temperature of the power amplifier and the real-time load impedance comprises:

[0020] calculate a dynamic stability degradation index based on the junction temperature and a phase angle of the real-time load impedance;

[0021] determine and adjust the stability compensation parameter to maintain a phase margin of the power amplifier according to the dynamic stability degradation index.

[0022] In a possible implementation manner of the first aspect, the reactive characteristic includes an imaginary part of the real-time load impedance.

[0023] The determining and adjusting of the one or more pre-distortion parameters for compensating for the non-linear distortion of the power amplifier based on the junction temperature of the power amplifier and the one or more reactive characteristics of the real-time load impedance includes:

[0024] The determining and adjusting of the pre-distortion parameter based on the junction temperature, the imaginary part of the real-time load impedance, and a real-time output power of the power amplifier.

[0025] In a possible implementation manner of the first aspect, the pre-distortion parameter is a set of coefficients of a memory polynomial pre-distortion model; and the reactive characteristic includes a phase angle of the real-time load impedance.

[0026] The determining and adjusting of the one or more pre-distortion parameters for compensating for the non-linear distortion of the power amplifier based on the junction temperature of the power amplifier and the one or more reactive characteristics of the real-time load impedance includes:

[0027] querying and obtaining the set of coefficients in a preset multi-dimensional lookup table with the junction temperature and the phase angle of the real-time load impedance as coordinates.

[0028] In a possible implementation manner of the first aspect, the method further includes:

[0029] generating a thermal management signal based on the junction temperature to perform thermal protection on the power amplifier.

[0030] In a second aspect, an embodiment of the present application provides an intelligent management system of a power amplifier, including:

[0031] a sensing device configured to acquire an output voltage and an output current of the power amplifier;

[0032] a cooperative controller connected to the sensing device, the cooperative controller being configured to execute the method according to the first aspect.

[0033] The embodiments of the present application can more accurately predict the junction temperature of the power device by establishing and utilizing the coupled thermal model that associates the load impedance mismatch with the effective thermal resistance, thereby realizing active and early thermal risk warning and protection, and effectively avoiding thermal breakdown caused by load mismatch. Further, based on the accurate and real-time evaluation of the system thermal state and the load state, the performance limit can be dynamically adjusted, avoiding the excessive conservatism of the traditional fixed limiting strategy, and the system performance can be dynamically released under the premise of ensuring safety, thereby realizing the maximization of output capacity. Further, by simultaneously incorporating the junction temperature and the load reactance characteristics into the stability and distortion compensation model, the system stability and nonlinearity are cooperatively and adaptively controlled, ensuring that the output signal quality remains stable and excellent under various working conditions, and realizing high-fidelity in all working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0035] Figure 1 A structural schematic diagram of a power amplifier intelligent management system provided by the embodiments of the present application;

[0036] Figure 2 A flowchart of a power amplifier intelligent management method provided by the embodiments of the present application;

[0037] Figure 3 A schematic diagram of the impedance-thermal coupling observer principle in the embodiments of the present application;

[0038] Figure 4 A schematic diagram of the dynamic stability degradation index model principle in the embodiments of the present application;

[0039] Figure 5 A schematic diagram of the multi-dimensional coupled pre-distortion coefficient model principle in the embodiments of the present application;

[0040] Figure 6 A principle schematic diagram of the lookup table implementation method in the embodiments of the present application. DETAILED DESCRIPTION

[0041] In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, technologies, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

[0042] It is to be understood that the terminology "includes", "has", "holds", "contains" and / or "comprising", "including", "having" and their conjugates used in the application and the appended claims mean "including but not limited to", and are not intended to (and not to be interpreted to) exclude other means, methodologies, ingredients, components, elements, and / or steps.

[0043] It is also to be understood that the terminology "and / or" used in the application and the appended claims means and includes any and all combinations of one or more of the associated listed items and is interpreted to be equivalent to "one or more of the listed items and / or any and all combinations of one or more of the listed items".

[0044] As used in this application and the appended claims, the terminology "if' can be interpreted as meaning "when" or "once" or "in response to a determination" or "in response to a detection" depending on the context. Similarly, the phrase "if determined" or "if detected" can be interpreted as meaning "once determined" or "in response to a determination" or "once detected" or "in response to a detection", depending on the context.

[0045] In addition, in the description of the application and the appended claims, the terms "first", "second", "third", etc. are used merely as labels for convenience, and are not intended to signify or imply relative importance.

[0046] Reference in the specification to "one embodiment" or "some embodiments" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. The appearances of the phrase "in one embodiment" or "in some embodiments" in various places in the specification are not necessarily all referring to the same embodiment, although it can. The terms "including", "containing", "having" and variations thereof are meant to encompass the terms "including but not limited to".

[0047] Embodiment 1

[0048] The embodiment provides a collaborative control scheme based on an innovative coupling model, which is preferably implemented in a complete power amplifier intelligent management system. Figure 1An overall structure schematic diagram of a power amplifier intelligent management system 100 provided by an embodiment of the present application. The intelligent management system 100 is particularly suitable for fine management of power amplifiers with harsh working environment and complex and changeable load characteristics, for example but not limited to, a hydroacoustic power amplifier in a deep water environment for ocean exploration or underwater communication, a radio frequency power amplifier for plasma generation, or a power amplifier for driving an antenna in a mobile communication base station, and the common feature of these scenarios is that the load impedance will deviate significantly due to environmental changes or working state changes.

[0049] Referring to Figure 1 The intelligent management system 100 mainly includes a cooperative controller 10, a power amplifier 20 and a sensing device 40 in structure. The power amplifier 20 is the executive main body of the system, which receives a low-power input signal (not shown in the figure) and amplifies the power to drive the load 30. The power amplifier 20 can be any type of power amplifier, such as class A, class AB, class D or class E amplifier. In the hydroacoustic application scenario, the load 30 is usually a piezoelectric transducer array or other acoustic emission device, whose impedance characteristics not only present a sharp reactance change near the resonance frequency, but also its resonance point itself will drift significantly with factors such as working frequency, environmental temperature and hydrostatic pressure. The sensing device 40 is tightly coupled with the output end of the power amplifier 20, aiming to collect the output voltage signal and output current signal of the power amplifier 20 in real time and high precision. The sensing device 40 can be composed of a voltage sensor and a current sensor (for example, a Hall effect-based current sensor, a high-frequency current transformer or a precision shunt resistor) with high enough bandwidth, which needs to be sufficient to capture the fundamental wave and key harmonic components of the signal to ensure the accuracy of subsequent calculation. The cooperative controller 10 is the brain of the whole intelligent management system 100, which can be physically realized as a high-performance digital signal processor (DSP), a field programmable gate array (FPGA), an application-specific integrated circuit (ASIC) or a system on chip (SoC) containing these processing units. The cooperative controller 10 is connected with the sensing device 40 through a high-speed analog-to-digital converter (ADC) interface to receive the sensing signals collected and digitized by the sensing device 40, which contain real-time voltage and current information; at the same time, the cooperative controller 10 is also connected with the control port of the power amplifier 20 through a digital interface (such as SPI bus) or a digital-to-analog converter (DAC) to send a series of cooperatively calculated control signals to it, so as to dynamically and cooperatively adjust its working parameters, such as bias point, power supply voltage, pre-distortion coefficient or compensation network parameters, etc.

[0050] It should be noted that the intelligent management method provided by the embodiment is executed in the form of firmware or software in the cooperative controller 10. Figure 2This application provides a detailed flowchart of a power amplifier intelligent management method according to an embodiment. The method is executed in a continuous, high-speed control loop, where the period Δt of each loop is sufficiently short to ensure timely response to changes in system state. The method specifically includes the following steps:

[0051] Step S201: Real-time Status Acquisition. At the beginning of each control cycle, the co-controller 10 acquires the instantaneous output voltage sequence Vm(t) and instantaneous output current sequence Im(t) of the power amplifier 20 at the current moment through the sensing device 40. Based on this raw time-domain sampled data, the co-controller 10 then performs a series of digital signal processing operations to determine the macroscopic operating state of the system. Specifically, the co-controller 10 can calculate the effective value voltage Vrms and effective value current Irms at the current operating frequency, as well as the phase difference φ between them, by applying Fast Fourier Transform (FFT) or other frequency domain analysis algorithms (such as the Gossel algorithm for single-frequency signals) to the voltage and current data within one or more complete signal cycles. Based on these basic electrical quantities, the real-time load impedance ZL(n) of the load 30 driven by the power amplifier 20 can be determined, where n represents the index of the current control cycle. The real-time load impedance ZL(n) is a complex number, i.e., ZL(n) = (Vrms / Irms)×(cos(φ) + jsin(φ)), which contains information in two dimensions: the magnitude |ZL(n)| and the phase angle φL(n). Simultaneously, the co-controller 10 also calculates the real-time dissipation power Pdiss(n) and real-time output power Pout(n) of the power amplifier 20. The output power Pout(n) refers to the effective power successfully transferred to the load 30, which can be calculated as Pout(n) = Vrms×Irms×cos(φ). The dissipation power Pdiss(n) refers to the power lost as heat by the internal power devices (such as MOSFETs or IGBTs) of the power amplifier 20 due to non-ideal characteristics such as conduction losses and switching losses. It can be accurately calculated by measuring the amplifier's DC input power Pin and subtracting the output power Pout(n), i.e., Pdiss(n) = Pin(n) - Pout(n). These real-time state parameters, especially the real-time load impedance ZL(n) which is the core state variable, form the data foundation for all subsequent adaptive control algorithms, providing the system with an accurate snapshot of its current operating environment.

[0052] Step S202: Coupled Thermal Model Prediction. After acquiring the real-time operating status, the co-controller 10 then executes a core prediction algorithm that predicts the junction temperature Tjvirt(n) of key power devices (e.g., output stage transistors) inside the power amplifier 20 based on an innovative impedance-thermal coupling model. This junction temperature is a virtual or estimated value because it is not directly measured by physical sensors but calculated through a model. Unlike simple first- or second-order thermal models in the prior art that treat the thermal resistance from the device to the environment as a fixed value, the coupled thermal model proposed in this application reveals and utilizes a key, but often overlooked, physical phenomenon: the effective thermal resistance of the power amplifier 20 is not a constant value but increases significantly with the increase of the mismatch between its actual driven real-time load impedance ZL(n) and a preset optimal load impedance Zopt. Understandably, the optimal load impedance Zopt is the ideal load value matched by the power amplifier 20 during its design to achieve a certain optimal performance index (such as maximum efficiency, maximum undistorted output power, etc.). It is usually a pure resistance value and is stored as a preset constant in the non-volatile memory of the co-controller 10. When the actual load ZL(n) deviates from Zopt for various reasons, a mismatch occurs, which will cause distortion of the voltage and / or current waveforms of the output stage inside the power amplifier, resulting in greater voltage or current stress. This will generate additional local hot spots inside the device or worsen the overall heat dissipation conditions, and its macroscopic effect is equivalent to an increase in its effective thermal resistance.

[0053] Figure 3 This is a schematic diagram of the impedance-thermal coupling observer 300 used for junction temperature prediction in this embodiment. The observer 300 is implemented within the co-controller 10 as a software module or hardware logic, receiving the calculation results from step S201 and several preset parameters. Specifically, the impedance calculation module 310 and the power dissipation calculation module 320 are responsible for calculating the real-time load impedance ZL(n) and power dissipation Pdiss(n), respectively. As the core of this observer, the coupling thermal resistance calculation module 330 receives the real-time load impedance ZL(n) and the preset optimal load impedance Zopt, calculates the mismatch between them, and determines a dynamically changing effective thermal resistance Rtheff(n) accordingly. Finally, the junction temperature prediction module 340 integrates the dynamic effective thermal resistance Rtheff(n), the real-time power dissipation Pdiss(n), and the ambient temperature Ta sensed by an external temperature sensor, and iteratively calculates the predicted junction temperature Tjvirt(n) for the current period by solving a discrete-time thermal model differential equation.

[0054] In a preferred embodiment of this application, the above-mentioned coupling relationship is modeled using a specific mathematical formula. The junction temperature prediction module 340 updates the predicted junction temperature Tjvirt(n) using a formula that includes a deterioration term proportional to the square of the mismatch. This formula is a discrete-time first-order thermal model, and its specific form is as follows:

[0055] .

[0056] The physical meanings of each parameter are as follows: Tjvirt(n) is the virtual junction temperature of the power device predicted in the nth control cycle, which is the final output of this step; Tjvirt(n-1) is the predicted junction temperature value of the previous control cycle, used as the initial state for this calculation, representing the memory of the thermal system; Δt is the duration of the control cycle, i.e., the time interval between each execution of this calculation, which is a system design constant; Cth is the equivalent heat capacity of the power module, characterizing its temperature change inertia, which is a preset constant obtained through experimental calibration or by consulting the device datasheet; Pdiss(n) is the value of the power device's virtual junction temperature predicted in the current cycle by step S. The real-time power dissipation calculated by 201 is the source of heat input; Ta is the ambient temperature or heatsink temperature of the power amplifier. For deep-water applications, this value can be obtained by a temperature sensor installed on the amplifier casing or heatsink, and is usually stable in the low single digits of Celsius; Rth is the reference thermal resistance of the power module when driving the optimal load Zopt (i.e., in a fully matched state), which is also a preset constant calibrated experimentally; γ is a key, dimensionless load mismatch thermal resistance degradation coefficient, which characterizes the degree of deterioration of heat dissipation efficiency with load mismatch. The larger the value, the more significant the impact of load mismatch on thermal resistance. This coefficient needs to be experimentally calibrated by conducting a series of thermal tests on a specific power amplifier under different loads, for example, by combining a load-pulling system and an infrared thermal imager; |ZL(n) - Zopt| 2 The square of the Euclidean distance between the real-time load impedance and the optimal load impedance in the complex plane is used to quantitatively describe the current load mismatch. This metric is equally sensitive to both amplitude and phase mismatches.

[0057] The core innovation of this formula lies in the dynamically deteriorating term in the denominator. This step dynamically amplifies the reference thermal resistance Rth to obtain the effective thermal resistance that varies with operating conditions in real time. When the load is perfectly matched, ZL(n) = Zopt, which is 1, and the effective thermal resistance equals its minimum value, i.e., the reference thermal resistance. As the load mismatch increases, the value of this term is greater than 1 and increases non-linearly with the square of the mismatch, thus increasing the effective thermal resistance and simulating a deterioration in heat dissipation capacity. This reduces the heat dissipation term in the thermal model (i.e., the part to the right of the minus sign in the formula), resulting in a faster rise rate and a higher steady-state value for the junction temperature prediction Tjvirt(n).

[0058] Step S203: Adaptive Stability Control. Based on the junction temperature Tjvirt(n) predicted in step S202 and the real-time load impedance ZL(n) determined in step S201, the co-controller 10 further determines and adjusts one or more stability compensation parameters to actively maintain the stable operation of the power amplifier 20. As a high-gain feedback system, the stability of the power amplifier (usually measured by phase margin and gain margin) is extremely sensitive to operating conditions. Stability is not only affected by device temperature (increased temperature usually changes transistor parameters and reduces phase margin), but also severely affected by load reactance characteristics (capacitive or inductive loads can introduce additional phase shifts, potentially turning negative feedback into positive feedback, thus causing system self-oscillation). This embodiment uses a novel coupling model to simultaneously consider these two key, independent degradation factors, achieving dynamic and adaptive maintenance of stability.

[0059] Specifically, this embodiment introduces the concept of Dynamic Stability Degradation Index (DSDI) to quantify risk. Figure 4 This is a schematic diagram of the dynamic stability degradation exponential model 400. This model 400 receives the predicted junction temperature Tjvirt and the phase angle φL(n) of the real-time load impedance as input, and calculates a dimensionless exponential DSDI(n), the formula of which is:

[0060] .

[0061] The meanings of each parameter are as follows: DSDI(n) is the dynamic stability degradation index, used to quantify the risk of the system becoming unstable under the current operating conditions. Under ideal operating conditions (such as room temperature Tj0, pure resistive load φL(n)=0), this index is close to 1; the first term... The effect of temperature on stability is characterized. Here, α is the temperature influence weighting coefficient, Tj0 is the reference junction temperature (e.g., 25°C), and Ts is the temperature sensitivity scaling factor; all three are preset constants calibrated through stability analysis and experiments. This term indicates that as the predicted junction temperature Tjvirt(n) increases, the stability degradation increases exponentially, which is consistent with the physical law of nonlinear changes in semiconductor device parameters with temperature; the second term... This characterizes the effect of load reactance on stability. β is the load phase influence weighting coefficient, a preset constant. φL(n) is the phase angle of the real-time load impedance. When the load is purely resistive, φL(n) is 0, and this term is 1, producing no additional effect. When the load exhibits strong reactance (capacitive or inductive), φL(n) approaches ±90°, and |sin(φL(n))| approaches 1, causing this term to increase significantly. This closely matches the physical reality that highly reactive loads are most prone to causing stability problems, as purely reactive loads introduce a phase shift of up to 90 degrees.

[0062] After calculating DSDI(n), the co-controller 10 adjusts the stability compensation parameter in real time based on this exponent. In this embodiment, it is assumed that the control loop of the power amplifier 20 includes a digital compensation filter (e.g., a digital PID controller or a lead-lag compensator). The stability compensation parameter can be a critical pole frequency fp(n) of the digital filter, and its adjustment logic can be a simple proportional relationship: .

[0063] Here, fpbase is the pole frequency under the baseline operating condition, which is a preset constant. By increasing this pole frequency in real time and proportionally, the frequency response characteristics of the compensation network can be dynamically adjusted. This proactively increases the phase margin of the system when facing high-risk operating conditions such as high temperature or highly reactive loads, effectively preventing parasitic oscillations and maintaining system stability. The digital filter coefficients here (indirectly determined by the pole frequency fp(n)) are specific stability compensation parameters.

[0064] Step S204: Adaptive Distortion Compensation. In parallel with stability control, the co-controller 10 also performs another crucial adaptive adjustment based on the predicted junction temperature Tjvirt(n) and real-time load information to determine and adjust one or more pre-distortion parameters for compensating for the nonlinear distortion of the power amplifier 20. The nonlinear distortion characteristics of the power amplifier (such as AM-AM and AM-PM distortion) not only drift with temperature but are also closely related to the reactance characteristics of the load (causing memory effects) and the output power level (causing gain compression). This embodiment uses a multi-dimensional coupled model to comprehensively compensate for these intertwined influencing factors.

[0065] In this embodiment, the input signal of the power amplifier 20 passes through a digital predistortion (DPD) module controlled by the co-controller 10 before entering the amplifier. This module preferably uses a memory polynomial model to process the input signal to generate an anti-distortion signal, thereby canceling the nonlinear distortion that the amplifier will generate. The purpose of this step is to adjust the complex coefficients in the memory polynomial model in real time. Figure 5This is a schematic diagram of the multidimensional coupled predistortion coefficient model 500. Model 500 receives the predicted junction temperature Tjvirt, the imaginary part of the load impedance Im{ZL(n)}, and the real-time output power Pout(n) as inputs, and calculates and outputs a set of real-time adjusted predistortion coefficients. A key third-order memory coefficient c in the model is used... 31 Taking (n) as an example, its calculation formula can be designed as follows:

[0066] .

[0067] This formula consists of four multiplied terms, each compensating for a specific physical effect: c 31ref The reference coefficient value under the reference operating condition (e.g., Tj0, purely resistive load, small-signal output) is a preset complex constant obtained through model identification; the first term The first term is for temperature drift compensation, where κ is a preset temperature drift sensitivity coefficient. This term simulates the physical effect that the gain of semiconductor devices typically decreases with increasing temperature, thereby dynamically adjusting the pre-distortion compensation strength; the second term... This is a load-related memory effect compensation term, where λ is a preset load reactance sensitivity coefficient, and Im{ZL(n)} is the imaginary part of the real-time load impedance. This term introduces the load's reactance characteristics (i.e., the ratio of the imaginary part of the impedance to its magnitude, which is related to the tangent of the phase angle) into the model. When the load is purely resistive, the imaginary part is 0, and this term is 1; when the load is inductive (imaginary part is positive) or capacitive (imaginary part is negative), this term will adjust the amplitude and phase of the compensation coefficient accordingly. This is crucial for accurately compensating for loads like piezoelectric transducers, where the imaginary part of the impedance changes drastically near the resonant point. Here, the imaginary part of the load impedance is a specific reactance characteristic; the third term... This is the power compression compensation term, where Pout(n) is the real-time output power, Psut is the amplifier's saturated output power (a preset constant), and δ is the power compression effect factor (a preset constant). This term characterizes the fact that when the output power Pout(n) approaches the saturated power Psut, the amplifier's gain compression effect will increase sharply, requiring stronger nonlinear compensation (i.e., a larger coefficient), thus achieving accurate feedforward compensation for the power compression effect.

[0068] By calculating and updating c in real time 31 (n) and other similar coefficients in the memory polynomial model, the predistortion module can accurately predict and cancel the complex nonlinear distortion caused by the combined effects of temperature, load reactance and power level, thus ensuring the high fidelity of the output signal of the power amplifier 20 under various dynamic operating conditions.

[0069] As an optional implementation, the method may also include a thermal management step in step S205. This step can be considered as an advanced protection layer based on accurate junction temperature prediction. The co-controller 10 compares the junction temperature Tjvirt(n) predicted in step S202 with one or more preset temperature thresholds (such as the warning threshold Twarn and the shutdown threshold Tshutdown) in real time. When Tjvirt(n) exceeds the warning threshold, the co-controller 10 can generate a thermal management signal that triggers a power limiting mechanism (e.g., by reducing the input signal amplitude or adjusting the amplifier supply voltage) to actively reduce the output power of the power amplifier 20 to prevent the temperature from continuing to rise. If Tjvirt(n) reaches or exceeds the shutdown threshold, the thermal management signal can trigger emergency protection to immediately shut down the output of the power amplifier 20. Since this protection mechanism is based on a rapidly responding predicted junction temperature rather than a slow-responding measured case temperature, it achieves active thermal protection, effectively avoiding thermal breakdown of power devices due to severe load mismatch, and significantly improving the reliability of the system.

[0070] In summary, this embodiment uses a series of tightly coupled mathematical models to link the four key elements of load impedance, thermal state, stability, and nonlinear distortion, and enables them to be managed collaboratively and intelligently in a unified collaborative controller 10. This achieves comprehensive and precise control of the power amplifier 20, significantly improving its safety, performance, and signal fidelity under complex and variable operating conditions.

[0071] Example 2

[0072] This embodiment is a variation of Embodiment 1, with the main difference being the specific implementation of the coupled thermal model prediction step (i.e., step S202). In Embodiment 1, the coupling relationship between load mismatch and effective thermal resistance is described by an analytical formula based on a quadratic relationship. However, in certain power amplifiers, this coupling relationship may exhibit a more complex and non-analytical form, making it difficult to accurately fit using simple mathematical formulas. Therefore, this embodiment provides an implementation based on a look-up table (LUT) to adapt to coupling relationships of arbitrary shapes, thereby improving the accuracy of the model.

[0073] Figure 6 This is a schematic diagram illustrating the principle of the table lookup method. In this embodiment, the working process of step S202 is adjusted as follows:

[0074] First, similar to Example 1, the cooperative controller 10 calculates the real-time load impedance ZL(n) in step S201, and further calculates the load mismatch degree as the lookup table address. .

[0075] Subsequently, the co-controller 10 no longer calculates the effective thermal resistance using a formula. Instead, it uses the calculated mismatch degree d as an address or index to look up a one-dimensional lookup table 610 pre-stored in its internal non-volatile memory (such as Flash or ROM). This lookup table 610 stores a series of discrete mismatch degree d values ​​and their corresponding effective thermal resistance Rtheff values. These data were obtained during the product development phase through extensive experimental testing and calibration of the power amplifier. For example, the amplifier was operated under different mismatch degrees using a load-pull system, and its steady-state junction temperature was measured using equipment such as an infrared thermal imager, thereby calculating the effective thermal resistance for each mismatch degree. If the calculated mismatch degree d value falls exactly on a sampling point in the lookup table 610, the corresponding Rtheff value is read directly; more commonly, if the d value falls between two sampling points, the co-controller 10 can execute an interpolation algorithm (such as calculating a simple linear interpolation, or a smoother higher-order spline interpolation) to calculate a more accurate Rtheff(n) value.

[0076] After obtaining the effective thermal resistance Rtheff(n) that dynamically changes with the load mismatch by looking up tables and interpolation, the co-controller 10 uses the following standard form of a first-order thermal model formula to calculate the junction temperature:

[0077] .

[0078] This formula has a similar structure to the formula in Example 1, but the effective thermal resistance Rtheff(n) in its denominator is obtained directly from lookup table 610 and interpolated, rather than being calculated by an analytical formula containing a deterioration term.

[0079] Understandably, the advantages of this embodiment lie in its high flexibility and potentially higher accuracy. It does not rely on any specific mathematical model assumptions (such as quadratic relationships) and can faithfully reproduce experimentally measured, arbitrarily complex coupling curves, thus potentially obtaining more accurate junction temperature predictions than the formula method. Furthermore, on certain hardware platforms (such as FPGAs), lookup operations implemented using on-chip memory (BRAM) may execute faster than complex floating-point multiplication and addition operations, thus offering a potential performance advantage. Subsequent steps such as adaptive stability control (step S203) and adaptive distortion compensation (step S204) can continue to use the potentially more accurate junction temperature Tjvirt(n) predicted in this embodiment as input, and the process is exactly the same as in Embodiment 1.

[0080] Example 3

[0081] This embodiment is another variation of Embodiment 1, the main difference being the specific implementation of the adaptive stability control step (i.e., step S203). Embodiment 1 describes achieving stability control in the digital domain by adjusting the coefficients of a digital filter. This method is well-suited for modern power amplifier systems where the entire signal link and control loop are implemented in the digital domain. However, for some traditional pure analog power amplifiers that do not have a digital signal processing core, this embodiment provides a way to achieve adaptive stability control in the analog domain, thereby broadening the applicability of the present invention.

[0082] In this embodiment, the feedback compensation network of the power amplifier 20 includes a physical variable capacitor element. This element can be a voltage-controlled varactor diode driven by a control voltage, or a digital stepper capacitor array driven by a digital control signal. By changing the capacitance value, the position of a critical pole or zero in the amplifier's open-loop frequency response can be directly adjusted, thereby changing its phase margin.

[0083] Accordingly, the working process of step S203 is adjusted as follows:

[0084] First, the co-controller 10 also assesses the system's stability risk based on the junction temperature Tjvirt(n) predicted in step S202 and the real-time load impedance ZL(n) determined in step S201. However, it may not calculate the DSDI index, but instead directly calculate the target compensation capacitor increment ΔCe(n) required to maintain sufficient phase margin according to a simplified, more direct coupling formula. This formula can be designed according to the specific amplifier topology, for example:

[0085] .

[0086] In this formula: ΔCe(n) is the incremental value of the compensation capacitor that needs to be adjusted, which is the core output of this step; kcc is a compensation coefficient, which is a preset constant experimentally calibrated by performing stability boundary tests on the amplifier, used to scale the sensitivity of the control; the (Tjvirt(n)- Tj0) term reflects the negative impact of temperature rise on stability, that is, the higher the temperature, the larger the required capacitor increment, which can introduce more phase lag to stabilize the system; This term reflects the impact of load weight on stability. For example, in some amplifier topologies, a lighter load (i.e., a larger |ZL(n)|) may be more likely to cause instability, thus requiring a larger compensation capacitor. Here, the load magnitude |ZL(n)| is a key characteristic of the real-time load impedance and is used as a control input.

[0087] After calculating the required target compensation capacitance increment value ΔCe(n), the co-controller 10 generates a corresponding control signal to adjust the variable capacitor element. If the element is a voltage-controlled varactor diode, the co-controller 10 will output a DC control voltage that has a specific functional relationship with ΔCe(n) through its digital-to-analog converter (DAC); if the element is a digital stepper capacitor array, the co-controller 10 will output a digital code through its digital I / O port or SPI bus to set the internal switching combination of the capacitor array so that its total capacitance value reaches the sum of the reference capacitance and ΔCe(n).

[0088] In this way, even in a purely analog power amplifier, the intelligent computing capabilities of the cooperative controller 10 can be utilized to achieve adaptive, closed-loop control of stability, combining digital intelligence with analog actuators. The incremental value ΔCe(n) of the compensation capacitor or the control voltage / digital code driving it is a specific stability compensation parameter. This embodiment demonstrates the broad platform adaptability of the cooperative control concept described in this application.

[0089] Example 4

[0090] This embodiment is another variation of Embodiment 1, the main difference being the specific implementation of the adaptive distortion compensation step (i.e., step S204). Embodiment 1 uses a multi-dimensional coupled analytical formula to calculate the predistortion coefficients in real time. However, for some power amplifier systems with extremely complex nonlinear behavior that are difficult to describe precisely using analytical formulas (e.g., broadband amplifiers with strong memory effects), this embodiment proposes an adaptive predistortion method based on a multi-dimensional lookup table (LUT).

[0091] In this embodiment, the digital predistortion module also preferably employs a memory polynomial model, but the model coefficients are no longer calculated in real time using analytical formulas, but are obtained by querying a pre-established multidimensional lookup table. This lookup table is also stored in the non-volatile memory of the co-controller 10.

[0092] Please refer to again Figure 6 The same principle applies to the multidimensional table lookup process in this embodiment. The working process of step S204 is modified as follows:

[0093] First, the co-controller 10 prepares the coordinate vectors for table lookup. Unlike the formula method in Embodiment 1, which relies on three dimensions—junction temperature, imaginary part of the load, and output power—the table lookup method can select different or more dimensions based on the significance of the actual impact. As an optional implementation, this embodiment selects the predicted junction temperature Tjvirt(n) and the phase angle φL(n) of the real-time load impedance as two key coordinates, because they represent the main sources of thermal effects and load memory effects, respectively.

[0094] Next, the co-controller 10 uses the pair of values ​​[Tjvirt(n), φL(n)] as input coordinates to perform a lookup in a pre-defined multidimensional lookup table 610. This lookup table is a two-dimensional (or higher-dimensional) table, where each grid point [Tji, φLj] stores a complete set of optimal memory polynomial predistortion coefficients {ckq} for that specific operating condition (i.e., junction temperature Tji and load phase angle φLj). These optimal coefficient vectors are obtained during the product development phase by scanning tests and nonlinear model identification of the power amplifier at various operating points. This process is usually automated; at each grid point, the system iteratively adjusts the predistortion coefficients until the distortion of the output signal (such as adjacent channel power ratio ACPR or error vector magnitude EVM) is minimized.

[0095] Since the real-time operating point [Tjvirt(n),φL(n)] usually does not fall precisely on the grid points of the lookup table, the cooperative controller 10 needs to perform a multidimensional interpolation algorithm (e.g., for a two-dimensional table, bilinear interpolation, which balances computational efficiency and performance, can be used) to calculate the optimal set of interpolated coefficients {ckq(n)} for the current operating condition based on the set of coefficients stored on four or more grid points around the point.

[0096] Finally, the co-controller 10 loads the newly calculated coefficients {ckq(n)} into the coefficient register of the digital predistortion module, replacing the original coefficients. In this way, the predistortion module can process the input signal in a manner best suited to the current temperature and load reactance characteristics, thereby achieving optimal compensation for the nonlinear distortion of the power amplifier.

[0097] The innovation of this embodiment lies in its transformation of the complex and analytically difficult-to-model nonlinear compensation problem into a data-driven lookup and interpolation problem. Its advantages include high adaptability; it eliminates the need for precise physical models or complex analytical formulas, and can compensate for arbitrarily complex nonlinear distortion behaviors, including long-range memory effects. Here, the entire set of coefficients {ckq(n)} of the memory polynomial predistortion model represents a specific predistortion parameter. This method provides a powerful and flexible adaptive distortion compensation implementation path for systems whose nonlinear characteristics are difficult to describe analytically, at the cost of requiring larger storage space and more arduous offline calibration work.

[0098] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. For example, more state variables can be introduced into the coupling model, or different mathematical formulas or machine learning models (such as neural networks) can be used to characterize the coupling relationship. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A smart management method for a power amplifier, characterized in that, Includes the following steps: The output voltage and output current of the power amplifier are obtained, and the real-time load impedance of the load driven by the power amplifier is determined based on the output voltage and the output current. The junction temperature of the power amplifier is predicted based on a coupled thermal model; wherein the coupled thermal model correlates the effective thermal resistance of the power amplifier with the mismatch between the real-time load impedance and the preset optimal load impedance. Based on the junction temperature of the power amplifier and the real-time load impedance, determine and adjust one or more stability compensation parameters to maintain the stable operation of the power amplifier. Based on the junction temperature of the power amplifier and one or more reactance characteristics of the real-time load impedance, determine and adjust one or more predistortion parameters for compensating for nonlinear distortion of the power amplifier; The optimal load impedance is the ideal load value matched by the power amplifier to achieve the highest efficiency or maximum undistorted output power; the mismatch is the square of the Euclidean distance between the real-time load impedance and the optimal load impedance in the complex plane. In the coupled thermal model, the effective thermal resistance of the power amplifier is modeled as increasing with the increase of the mismatch between the real-time load impedance and the preset optimal load impedance. The step of predicting the junction temperature of the power amplifier based on the coupled thermal model includes: The effective thermal resistance is calculated using a formula that includes a deterioration term proportional to the square of the mismatch. Based on the calculated effective thermal resistance and the power dissipation of the power amplifier, the junction temperature of the power amplifier is determined; The formula containing the deterioration term proportional to the square of the mismatch is as follows: ； Where Tjvirt(n) is the predicted junction temperature of the power device obtained in the nth control cycle, Tjvirt(n-1) is the predicted junction temperature of the previous control cycle, Δt is the duration of the control cycle, Cth is the equivalent thermal capacity of the power module, Pdiss(n) is the real-time power dissipation, Ta is the ambient temperature or heat sink temperature of the power amplifier, and |ZL(n) - Zopt| 2 Rth is the square of the Euclidean distance between the real-time load impedance and the optimal load impedance in the complex plane, used to describe the load mismatch. Rth is the reference thermal resistance when driving the optimal load impedance, and γ is the load mismatch thermal resistance degradation coefficient. Among them, based on the deterioration item The reference thermal resistance Rth is dynamically amplified to obtain the effective thermal resistance that varies with operating conditions in real time. .

2. The method according to claim 1, characterized in that, The step of predicting the junction temperature of the power amplifier based on the coupled thermal model includes: The effective thermal resistance is obtained by looking up the mismatch in a pre-stored lookup table. Based on the effective thermal resistance obtained from the query and the power dissipation of the power amplifier, the junction temperature of the power amplifier is determined.

3. The method according to claim 1, characterized in that, The stability compensation parameter includes one of the following: The incremental value of a variable capacitor used to adjust the physical compensation network of the power amplifier, and the coefficients of a digital filter in the digital signal processing path.

4. The method according to claim 1, characterized in that, The determination and adjustment of one or more stability compensation parameters for maintaining stable operation of the power amplifier, based on the junction temperature of the power amplifier and the real-time load impedance, includes: Calculate the dynamic stability degradation index based on the junction temperature and the phase angle of the real-time load impedance; Based on the dynamic stability degradation index, the stability compensation parameters are determined and adjusted to maintain the phase margin of the power amplifier; The formula for calculating the dynamic stability degradation index is as follows: Where DSDI(n) is the dynamic stability degradation index, α is the temperature influence weighting coefficient, Tjvirt(n) is the predicted junction temperature of the power amplifier, Tj0 is the room temperature, Ts is the temperature sensitivity scaling factor, β is the load phase influence weighting coefficient, and φL(n) is the phase angle of the real-time load impedance.

5. The method according to claim 1, characterized in that, The reactance characteristic includes the imaginary part of the real-time load impedance; The determination and adjustment of one or more predistortion parameters for compensating for nonlinear distortion of the power amplifier, based on the junction temperature of the power amplifier and the reactance characteristics of the real-time load impedance, includes: The predistortion parameters are determined and adjusted based on the junction temperature, the imaginary part of the real-time load impedance, and the real-time output power of the power amplifier.

6. The method according to claim 1, characterized in that, The predistortion parameters are a set of coefficients of the memory polynomial predistortion model; the reactance characteristics include the phase angle of the real-time load impedance; The determination and adjustment of one or more predistortion parameters for compensating for nonlinear distortion of the power amplifier, based on the junction temperature of the power amplifier and the reactance characteristics of the real-time load impedance, includes: Using the junction temperature and the phase angle of the real-time load impedance as coordinates, the set of coefficients is retrieved from a preset multidimensional lookup table.

7. The method according to claim 1, characterized in that, The method further includes: Based on the junction temperature, a thermal management signal is generated to perform thermal protection on the power amplifier.

8. An intelligent management system for a power amplifier, characterized in that, include: A sensing device is used to acquire the output voltage and output current of the power amplifier; A cooperative controller, connected to the sensing device, is configured to perform the method as described in any one of claims 1 to 7.

Images