DPWM DAC Noise Shaping for Low-Distortion Analog Output
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital-to-analog converters face issues such as performance degradation due to unmatched rise/fall times, clock jitter limitations, intersymbol interference, and high harmonic distortion, especially at high data rates, and require complex Dynamic Element Matching schemes to mitigate errors.
Innovation Solution
A digital-to-analog converter system incorporating an interpolation filter, a Digital Pulse Width Modulation (DPWM) Noise Shaper to suppress in-band quantization errors, and a hybrid Finite Impulse Response (FIR) filter coupled with a reconstruction amplifier to generate an accurate analog output, effectively addressing the limitations of existing converters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multi-bit truncation DACs are used, then conversion efficiency is improved, but static nonlinearity conversion errors increase requiring complex Dynamic Element Matching schemes
Solution Approach 1:
The patent extracts and addresses only the dominant error component (dynamic nonlinearity) rather than attempting to correct all error types. By focusing on suppressing dynamic nonlinearity through noise shaping and filtering, the system achieves high conversion efficiency without requiring complex Dynamic Element Matching schemes for static nonlinearity correction
Solution Approach 2:
The patent changes the approach from correcting static nonlinearity errors to shaping and filtering dynamic error spectra. By using noise shaping to push quantization noise to higher frequencies and applying anti-imaging filters, the system transforms the error profile to achieve both high conversion efficiency and low distortion
2Area of stationary object
If continuous time implementations are used, then circuit area is reduced, but distortion and in-band noise increase due to unmatched rise/fall times and intersymbol interference
Solution Approach 1:
The patent applies preliminary digital filtering and noise shaping before the analog conversion stage. By pre-shaping the quantization noise spectrum and filtering out-of-band components in the digital domain, the system prevents distortion and in-band noise problems from occurring in the continuous-time analog stage
Solution Approach 2:
The patent replaces complex analog continuous-time filtering mechanisms with digital signal processing techniques. By performing noise shaping, filtering, and error correction in the digital domain before conversion, the system achieves precise control without requiring complex analog continuous-time filter circuits
3Manufacturing precision
If discrete time implementations with switched capacitors are used, then manufacturing precision is maintained, but dynamic range is limited by thermal noise and large bandwidth is required
Solution Approach 1:
The patent extracts and removes out-of-band quantization noise components through digital filtering before conversion. By separating and eliminating high-frequency noise components in the digital domain, the system extends the effective dynamic range without requiring excessively wide analog bandwidth
Solution Approach 2:
The patent changes the noise spectrum distribution through digital noise shaping, concentrating quantization noise in high-frequency bands that are subsequently filtered. This parameter transformation allows the system to achieve high dynamic range in the signal band while using narrower overall bandwidth
Data Source
AI summary
A system and method is disclosed for a digital to analog converter which includes an interpolation filter to up-sample a digital signal, a noise shaping modulator to suppress in-band quantization errors due to digital pulse width modulation and truncation errors, and a hybrid finite impulse response filter/digital to analog converter coupled to a reconstruction filter which outputs the analog signal.