Image Sensor Dual Sense Node Conversion Gain Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional CMOS image sensors face challenges in achieving high dynamic range and low temporal read noise due to the need for adjustable conversion gain, which often requires additional transistors that increase parasitic capacitance, limiting the minimum achievable noise and conversion gain.

Innovation Solution

The solution involves an image sensor with two reset transistors connected in series between a reset voltage supply and a sense node, allowing for selective operation in high and low conversion gain modes without adding additional transistors to the sense node, using control logic to manage the transistors' switching during readout and non-readout periods.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If additional transistors are added to the sense node to enable adjustable conversion gain, then the pixel can operate in different gain modes, but the parasitic capacitance increases, limiting the minimum achievable noise

Engineering Contradiction:
Improveadjustable conversion gainVSAvoidparasitic capacitance
Core Design Contradiction:
Adaptability or versatilityVSObject-generated harmful factors

Solution Approach 1:

The pixel is divided into two separate sense nodes (first sense node and second sense node) with different capacitance values. The first sense node has lower parasitic capacitance for high conversion gain operation, while the second sense node has higher capacitance for low conversion gain operation. This segmentation allows the pixel to achieve adjustable conversion gain without adding additional transistors to a single sense node, thereby avoiding increased parasitic capacitance.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Instead of adjusting conversion gain by adding components in the same dimension (additional transistors at the sense node), the invention introduces a new dimension by creating multiple sense nodes with different capacitance characteristics. The control logic switches between these nodes to achieve different gain modes, effectively adding a dimensional aspect to the conversion gain adjustment mechanism.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If conversion gain is increased to reduce read noise, then small signals are amplified better, but the charge capacity is reduced, limiting dynamic range

Engineering Contradiction:
Improveread noiseVSAvoidcharge capacity
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The pixel dynamically switches between two sense nodes with different capacitance values based on the required conversion gain. When high conversion gain is needed (for low light conditions), the first sense node with lower capacitance is used to minimize noise. When low conversion gain is needed (for high light conditions), the second sense node with higher capacitance is used to maintain charge capacity. This dynamic switching allows the system to optimize both read noise and charge capacity according to operating conditions.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The invention changes the effective capacitance parameter of the sense node by switching between two different sense nodes. The first sense node has a first capacitance value optimized for high conversion gain, while the second sense node has a second capacitance value optimized for low conversion gain. This parameter change allows the pixel to adapt its electrical characteristics to different operating conditions, achieving both low noise and adequate charge capacity as needed.

Inventive Principle:
Principle #35Parameter changes

3Quantity of substance

If conversion gain is decreased to increase charge capacity, then dynamic range is improved, but read noise increases, reducing measurement precision

Engineering Contradiction:
Improvecharge capacityVSAvoidread noise
Core Design Contradiction:
Quantity of substanceVSMeasurement precision

Solution Approach 1:

The pixel dynamically selects the appropriate sense node based on lighting conditions. In low light conditions where high conversion gain is needed, the first sense node is used to minimize read noise. In high light conditions where large charge packets are expected, the second sense node is used to provide adequate charge capacity while maintaining acceptable noise levels. This dynamic adaptation resolves the contradiction between charge capacity and read noise.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The invention changes the capacitance parameter by switching between two sense nodes with different capacitance values. The first sense node has lower capacitance for high conversion gain operation, while the second sense node has higher capacitance for low conversion gain operation. This parameter change allows the system to optimize the trade-off between charge capacity and read noise according to the specific operating conditions.

Inventive Principle:
Principle #35Parameter changes

4Adaptability or versatility

If reset transistor switching is implemented for gain mode selection, then conversion gain is adjustable, but power consumption increases during switching operations

Engineering Contradiction:
Improveconversion gain modeVSAvoidpower consumption
Core Design Contradiction:
Adaptability or versatilityVSUse of energy by moving object

Solution Approach 1:

The control logic implements periodic switching of the reset transistor to alternate between high gain mode and low gain mode based on the readout cycle. During the readout period, one sense node is active while the other is reset in preparation. During non-readout periods, the reset transistor can be switched to prepare the alternate sense node. This periodic action allows gain mode adjustment while managing power consumption by coordinating switching with the natural readout rhythm of the image sensor.

Inventive Principle:
Principle #19Periodic action

Solution Approach 2:

The invention maintains continuous operation by ensuring that while one sense node is being read out, the other sense node is being reset in parallel. This continuous preparation and switching between nodes allows gain mode changes without interrupting the overall readout process, maintaining useful action continuity and reducing the impact of switching operations on power consumption.

Inventive Principle:
Principle #20Continuity of useful action

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach reduces power consumption, minimizes interference, and allows for high conversion gain without increasing parasitic capacitance, resulting in improved dynamic range and reduced noise across various light conditions.

Implementation Method 1

a pinned photodiode

Methodology Applied
Scientific EffectPhotoelectric effect: Photoelectric Effect

Data Source

PatentUS10170514B2Image sensor
Publication Date: 2019.01.01 CMOSIS
  • US10170514B2 patent drawing
  • US10170514B2 patent drawing
  • US10170514B2 patent drawing

AI summary

An image sensor comprises an array of pixels comprising: a pinned photodiode; a first sense node A; a second sense node B; a transfer gate TX connected between the pinned photodiode and the first sense node A; a first reset transistor M3 connected between a voltage reference line Vrst and the second sense node B; a second reset transistor M4 connected between the first sense node A and the second sense node B; and a buffer amplifier M1 having an input connected to the first sense node A. The control logic is arranged to operate the pixels in a low conversion gain mode and in a high conversion gain mode. In each of the conversion gain modes the control logic is arranged to operate one of a first reset control line RS1 and a second reset control line RS2 to continuously switch on one of the first reset transistor M3 and the second reset transistor M4 during a readout period of an operational cycle of the pixels.