Image Sensor ADC Bit-Split Conversion for Higher Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Analog-to-digital converters in image sensors face challenges in increasing resolution without increasing complexity or processing time, particularly due to noise dependencies on signal amplitude, which affects the signal-to-noise ratio and requires efficient noise management.
Innovation Solution
The method involves comparing the analog signal amplitude to a threshold representing half the full-scale signal value, performing conversion over a reduced number of bits for most significant or least significant bits based on the comparison, and adjusting the full-scale value and quantization levels accordingly to achieve higher resolution without excessive complexity or time increase.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of conversion bits is increased to improve resolution, then the measurement precision is improved, but the device complexity and processing time increase
Solution Approach 1:
The patent divides the conversion process into two independent stages: a first converter producing most significant bits (MSBs) and a second converter producing least significant bits (LSBs). This segmentation allows each converter to operate at lower resolution, reducing individual converter complexity while achieving high overall resolution when results are combined.
Solution Approach 2:
The patent introduces a temporal dimension by performing conversions at different times with different full-scale ranges. The first conversion occurs before the second, with the second conversion operating on a reduced full-scale range based on the MSB result. This dimensional approach allows high precision without requiring a single complex high-resolution converter.
2Measurement precision
If the number of conversion bits is increased to improve resolution, then the measurement precision is improved, but the processing time increases
Solution Approach 1:
The conversion process is segmented into two sequential stages with the second stage operating on a reduced full-scale range. Because the second converter only needs to resolve a smaller portion of the total signal range (determined by the MSBs), it can complete its conversion faster than a single high-resolution converter would require, reducing total processing time.
Solution Approach 2:
The patent performs a first conversion with a full-scale range that is larger than strictly necessary for the final precision requirement, then performs a second conversion with a reduced full-scale range. This partial action approach allows the system to quickly establish the MSBs and then spend less time resolving the LSBs, optimizing overall conversion time.
3Productivity
If a variable ramp slope is used to decrease conversion time, then the processing time is reduced, but quantization errors increase
Solution Approach 1:
The patent segments the conversion into two stages, with the second stage using a reduced full-scale range. This allows the use of faster conversion methods (such as variable ramp slopes) in the second stage without compromising overall precision, because the reduced range means fewer quantization levels are needed and the impact of variable slope errors is minimized.
Solution Approach 2:
The patent applies different conversion strategies to different parts of the signal range. The first converter handles the full-scale range with high accuracy requirements for the MSBs, while the second converter handles only a portion of the range (determined by the MSBs) with relaxed accuracy requirements for the LSBs. This local quality approach allows optimization of conversion speed in the second stage without significantly impacting overall precision.
Data Source
AI summary
A method of analog-to-digital conversion over n bits of an analog signal, including the steps of: comparing the amplitude of the analog signal with a threshold representing the amplitude of the full-scale analog signal divided by 2k, where k is an integer smaller than n; performing an analog-to-digital conversion of the analog signal over n−k bits to obtain the n−k most significant bits of a binary word over n bits if the result of the comparison step indicates that the amplitude of the input signal is greater than the threshold, and the n−k least significant bits of this binary word otherwise. An analog-to-digital converter and its application to image sensors.