ADC Noise Cancellation With Smaller Sampling Capacitors

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional analog-to-digital converter (ADC) architectures face challenges in reducing kT/C sampling noise, noise coupling, and amplifier thermal noise, which limits the use of larger sampling capacitors due to driving difficulties and significant die area occupation.

Innovation Solution

The proposed solution involves various techniques that significantly reduce noise sources in ADC circuits, allowing for smaller sampling capacitors while maintaining improved noise performance and power efficiency. These techniques include canceling or reducing kT/C sampling noise, noise coupling, and amplifier thermal noise, enabling smaller capacitors and reduced die area and power consumption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If larger sampling capacitors are used to reduce kT/C sampling noise, then noise performance is improved, but die area occupation increases significantly

Engineering Contradiction:
Improvenoise performanceVSAvoiddie area
Core Design Contradiction:
Measurement precisionVSArea of stationary object

Solution Approach 1:

The patent applies correlated double sampling to convert the harmful kT/C sampling noise into a measurable and cancelable component. By taking two samples (one with the signal plus noise, one with noise only) and subtracting them, the noise that would normally require large capacitors to suppress is instead actively canceled through correlation processing, achieving low noise with small capacitors

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent introduces an intermediary correlation processing stage between the sampling capacitors and the final digital output. This intermediary process computes the correlation between two samples and uses it to eliminate noise components, allowing small capacitors to achieve the noise performance that would otherwise require large capacitors

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If larger sampling capacitors are used to reduce kT/C sampling noise, then noise performance is improved, but driving difficulty increases

Engineering Contradiction:
Improvenoise performanceVSAvoiddriving ease
Core Design Contradiction:
Measurement precisionVSEase of operation

Solution Approach 1:

The patent transforms the noise problem into a solvable correlation problem. Instead of trying to drive large capacitors to suppress noise, the system uses small capacitors and relies on correlation processing to identify and eliminate noise, making the driving task much easier while maintaining noise performance

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent changes the approach from passive noise suppression through capacitor sizing to active noise cancellation through correlation processing. By changing the fundamental parameter from capacitor size to correlation computation, the system achieves noise reduction without the driving difficulties associated with large capacitors

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If conventional ADC architectures are used, then kT/C sampling noise can be reduced with larger capacitors, but power consumption increases

Engineering Contradiction:
Improvenoise performanceVSAvoidpower consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent converts the power-consuming approach of using large capacitors into a computationally-efficient correlation process. The correlated double sampling algorithm processes two samples and subtracts their correlation, actively canceling noise without requiring the continuous power consumption needed to drive large capacitor arrays

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent replaces expensive, power-hungry large capacitors with cheap, low-power small capacitors. The noise suppression function is transferred from the physical capacitor size to a computational correlation process that consumes significantly less power, achieving the same noise performance with much lower energy usage

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

Data Source

PatentEP3793092B1Low noise analog-to-digital converter
Publication Date: 2025.04.23 ANALOG DEVICES INT UNLTD CO
  • EP3793092B1 patent drawingFigure 1~2
  • EP3793092B1 patent drawingFigure 3~4
  • EP3793092B1 patent drawingFigure 5~6

AI summary

Noise sources in an ADC circuit can include kT/C noise of a sampling capacitor, noise coupling on to sampling capacitors from digital circuits, and amplifier noise. Also, charge injection from mismatch in sample switches can cause offsets. These various noise sources can be largely canceled or reduced using described techniques. As a result, the size of the sampling capacitors can be greatly reduced, while still achieving significantly improved noise performance and power efficiency for the overall converter.