Class-D Amplifier Feedback Loops for Low-Power Noise Shaping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Class D amplifiers in RF transceivers face challenges in minimizing noise and distortion due to limitations in time resolution, output voltage modulation by power supply, non-linear output impedance, and load-dependent frequency response, which affect the quality of audio signal reproduction.
Innovation Solution
The implementation of a multiple feedback architecture with two integrators, a triangle signal generator, comparator, and a feedback lowpass filter in the Class-D amplifier network, along with Dynamic Element Matching (DEM) and Natural Sampling Pulse Width Modulation (PWM), to create a negative feedback loop that suppresses non-idealities and attenuates quantization noise, ensuring improved noise reduction at low power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a simple integrator feedback loop is used in Class-D amplifier, then the device complexity is reduced, but the noise and distortion suppression capability deteriorates
Solution Approach 1:
The patent implements a multiple-loop feedback architecture where the output signal is fed back through a feedback filter to multiple integrators. This multi-loop structure provides stronger noise and distortion suppression compared to simple single-loop feedback, while maintaining practical circuit complexity through systematic design of the feedback path.
Solution Approach 2:
The patent introduces a feedback filter as an intermediary component in the feedback path. This filter shapes the feedback signal to optimize noise suppression in the audio band while allowing the multiple integrators to work effectively. The feedback filter acts as a mediator that conditions the feedback signal before it reaches the integrators.
2Device complexity
If practical clock frequencies with limited time resolution are used, then the device complexity is reduced, but the quantization distortion increases
Solution Approach 1:
The patent employs periodic action through the multiple integrators that process the feedback signal at each switching period. The integrators accumulate error signals over time, effectively distributing quantization errors across multiple periods and reducing their impact on any single output sample. This periodic integration process mitigates quantization distortion without requiring higher clock frequencies.
3Device complexity
If power supply voltage variations are allowed to directly modulate output voltage, then the device complexity is reduced, but the audio signal quality deteriorates
Solution Approach 1:
The multiple feedback loops continuously monitor the output voltage and compare it with the expected signal. When power supply variations cause output voltage deviations, the feedback mechanism detects these errors and generates corrective signals through the integrators, automatically compensating for power supply-induced noise without requiring complex external regulation.
Solution Approach 2:
The amplifier system performs self-correction of power supply noise through its internal multiple integrator feedback structure. The integrators accumulate error signals caused by power supply variations and generate compensating control signals that automatically adjust the output to cancel out the noise, enabling the system to service its own power supply issues without external intervention.
4Device complexity
If dead-time errors are not compensated, then the device complexity is reduced, but the output impedance non-linearity increases
Solution Approach 1:
The feedback loops capture the actual output signal including dead-time effects and feed it back through the integrators. This allows the system to detect and compensate for dead-time induced non-linearities in real-time, maintaining output impedance linearity without requiring separate dead-time compensation circuitry or complex control logic.
Data Source
AI summary
An audio amplifier system may include an audio CODEC/output (AOP) path featuring analog class-D amplifiers, and using Natural Sampling Pulse Width Modulation (PWM) to convert an analog input into a series of Rail-to-Rail pulses. The audio signal may be encoded in the average value of the PWM pulse train and may be recovered from the PWM signal by analog low pass filtering. The Class-D amplifiers may be designed with a negative feedback loop/network to compare the output signal with the input signal and suppress non-idealities introduced by the Class-D switching stage. Furthermore, operation of the AOP may be designed according to a separate signal transfer function and a separate noise transfer function, and 2nd order noise shaping may be performed at low power, with an optimized filter included in the feedback loop to achieve the best noise reduction at low power. Operation of the amplifier feedback network may be similar to that of a continuous time, low-pass delta-sigma modulator, but with a PWM loop wrapped around the class-D power amplifier.