Multi-Bit DAC Element Rotation for Mismatch Noise Shaping

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Multi-bit delta/sigma modulators in digital-to-analog converters face nonlinear distortion due to differences in nominally identical circuit elements caused by fabrication variations, leading to spectral noise and mismatch errors.

Innovation Solution

A multi-bit data weighted averaging circuit with a thermometer encoder and rotational dynamic element matching circuit is employed, using a binary to thermometer code conversion array to select and rotate DAC elements, applying first-order mismatch noise shaping to mitigate these errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If conventional analog-to-digital and digital-to-analog conversion techniques are used with low sampling rates, then the conversion can be implemented with simpler digital signal processing, but very precise analog components are required which increases manufacturing complexity and cost

Engineering Contradiction:
Improveanalog component precisionVSAvoiddigital signal processing complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent changes the sampling rate parameter from Nyquist rate to oversampling rate (e.g., 64x or 128x oversampling). This parameter change allows the system to use simpler, higher-tolerance analog components while moving the complexity to the digital domain where it can be more easily managed through software-based filtering and processing.

Inventive Principle:
Principle #35Parameter changes

2Ease of manufacture

If oversampling is used with simple high-tolerance components, then high resolution can be achieved with simpler analog components, but fast and complex digital signal processing techniques are required for filtering aliasing frequencies

Engineering Contradiction:
Improveanalog component toleranceVSAvoiddigital signal processing speed
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The patent applies preliminary dithering to the digital signal before oversampling and filtering. This preliminary action randomizes quantization errors and prepares the signal for more effective filtering, reducing the computational burden during the filtering stage while maintaining high resolution and allowing the use of simpler analog components.

Inventive Principle:
Principle #10Preliminary action

3Object-affected harmful factors

If dynamic element matching is used to compensate for DAC element mismatches, then spectral noise is reduced, but the system complexity increases due to element rotation and selection mechanisms

Engineering Contradiction:
Improvespectral noiseVSAvoidelement selection mechanism complexity
Core Design Contradiction:
Object-affected harmful factorsVSDevice complexity

Solution Approach 1:

The patent implements dynamic element matching by continuously rotating the assignment of DAC elements to code words based on their measured characteristics. This dynamic reassignment compensates for element mismatches and reduces spectral noise. The system adapts to component variations by dynamically adjusting which physical DAC element responds to which digital code, thereby equalizing the effective weight of all elements.

Inventive Principle:
Principle #15Dynamics

4Measurement precision

If the number of DAC elements is increased to improve precision, then conversion accuracy improves, but the complexity of the summation circuit and element selection increases significantly

Engineering Contradiction:
Improveconversion accuracyVSAvoidsummation circuit complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the DAC element pool into multiple subsets or groups, each handling a portion of the dynamic range or code space. Instead of all elements being actively selected for every conversion, the system divides the element pool and rotates through different segments. This segmentation reduces the number of elements that need to be simultaneously managed and summed, thereby reducing circuit complexity while maintaining overall precision through the combined contribution of all segments over time.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS9214953B1Generalized data weighted averaging method for equally weighted multi-bit D/A elements
Publication Date: 2015.12.15 DIALOG SEMICONDUCTOR (UK) LTD
  • US9214953B1 patent drawing
  • US9214953B1 patent drawing
  • US9214953B1 patent drawing

AI summary

A multi-level data weighted averaging circuit for inclusion in a multi-bit DAC circuit has a thermometer encoder that includes a binary to thermometer code conversion array configured for retaining or generating a multi-bit code representing an amplitude to be developed by each of the DAC element of the DAC circuit. The thermometer encoder is in communication with a multi-bit delta/sigma modulator to receive an oversampled binary coding representing an amplitude of a sampling of an analog signal. The oversampled binary coding is applied to an element selector to select the elements retaining or generating the element binary strings of bits of the multi-bit thermometer code. The multi-bit thermometer code is transferred to a rotational dynamic element matching circuit that rotationally selects order that the DAC elements are to receive each of the element binary strings of bits of the multi-bit thermometer code.