Multi-Level Class D Audio Amplifier Driver Topology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Class D audio amplifiers face issues with poor power efficiency at low audio input signals due to switching losses and excessive EMI noise, requiring large external inductors and capacitors, which increase costs and reduce reliability.
Innovation Solution
A class D audio amplifier with an improved output driver topology that uses a controller to generate pulse width or pulse density modulated control signals to create 3- or 4-level output signals, reducing common mode ripple voltage and inductor ripple current, and incorporates a feedback loop to suppress quantization noise and non-ideal switching errors, integrated on a CMOS or BCD semiconductor die for compact and efficient operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If traditional H-bridge Class D amplifier topology with 2-level PWM is used, then the amplifier achieves high power efficiency at large audio input signals, but power efficiency deteriorates at low audio input signals due to switching losses and idle power consumption
Solution Approach 1:
The output voltage levels are segmented into multiple discrete levels (3-level, 4-level, or 5-level PWM) instead of traditional 2-level switching. This segmentation allows the amplifier to operate at reduced voltage levels during low audio input signals, thereby reducing switching losses and idle power consumption while maintaining power efficiency across a wide dynamic range
Solution Approach 2:
The amplifier dynamically adjusts the number of output levels and switching frequency based on the audio input signal level. During low signal levels, the system transitions to multi-level PWM with lower switching activity, while during high signal levels, it operates in traditional 2-level mode, optimizing power efficiency adaptively across different operating conditions
2Device complexity
If traditional 2-level PWM switching is used, then the amplifier achieves simple circuit topology, but excessive EMI noise is generated due to high amplitude rectangular pulses with repetition frequencies between 250 kHz-2 MHz
Solution Approach 1:
The single large-amplitude switching pulse is segmented into multiple smaller voltage level transitions (3-level, 4-level, or 5-level PWM). This segmentation reduces the amplitude of individual voltage steps and distributes the spectral energy across multiple lower-frequency transitions, thereby reducing EMI noise while maintaining the same fundamental switching frequency range
Solution Approach 2:
The voltage amplitude parameter of the PWM output is changed from 2-level to multi-level (3, 4, or 5 levels). This parameter change reduces the dv/dt of each switching transition, which directly reduces electromagnetic radiation and EMI noise generation while preserving the audio signal fidelity
3Reliability
If load inductors and capacitors of appropriate size are used for filtering, then the amplifier achieves adequate lowpass filtering of switching frequency components, but the inductors and capacitors become so large that they must be provided as external components, increasing costs and reducing reliability
Solution Approach 1:
The filtering function is segmented and distributed across multiple voltage levels in the multi-level PWM output. By transitioning through intermediate voltage levels rather than direct switching between positive and negative rails, the amplitude of ripple currents through load inductors and capacitors is reduced, allowing for smaller filter component sizes that can be integrated on-chip
Solution Approach 2:
The switching transitions in multi-level PWM are designed to minimize ripple current stress on filter components. The segmented voltage transitions convert the potentially harmful high-amplitude ripple currents into lower-amplitude currents that naturally require smaller filtering components, turning a design constraint into an advantage for integration
Data Source
AI summary
The present invention relates in one aspect to a class D audio amplifier with improved output driver topology supporting multi-level output signals such as 3-level, 4-level or 5-level pulse width or pulse density modulated output signals for application to a loudspeaker load. The present class D audio amplifiers are particularly well-suited for high-volume consumer audio applications and solutions.