Differential readout from pixels in CMOS sensor

Abstract

The present invention provides an improved pixel readout circuit that compensates for common mode noise during a read out operation. This is accomplished by using a differential readout of the signal and reset value from the desired pixel compared with the reset value from a reference pixel. In this manner common mode noise can be offset and therefore minimized. In one embodiment of the invention, the reference pixel is the nearest neighbor pixel in the same row. In another embodiment, the reference pixel is the nearest neighboring pixel in a different row.

Description

FIELD OF THE INVENTION[0001]The invention relates generally to improved semiconductor imaging devices and in particular to an imaging device which can be fabricated using a standard CMOS process. Particularly, the invention relates to a CMOS imager having an array of image sensing cells and to the driving signals which operate the cells.BACKGROUND OF THE INVENTION[0002]There is a current interest in CMOS active pixel imagers for possible use as low cost imaging devices. An exemplary pixel circuit of a CMOS active pixel sensor (APS) is described below with reference to FIG. 1. Active pixel sensors can have one or more active transistors within the pixel unit cell, can be made compatible with CMOS technologies and promise higher readout rates compared to passive pixel sensors. The FIG. 1 exemplary pixel cell 10 is a 4T APS, where the 4T is commonly used in the art to designate use of four transistors to operate the pixel. A 4T pixel has a photodiode 162, a reset transistor 184, a tran...

AI Technical Summary

Benefits of technology

[0013]The present invention provides an improved pixel readout circuit and method of operation which minimizes substrate and other common mode noise during a read out operation. The circuit improves the consistency of the pixel to pixel output of the pixel array and increases the dynamic range of the pixel output. This is accomplished by obtaining a differential readout of the reset signal and integrated charge signal from a desired pixel along with a reset signal and a comparison signal from a different reference pixel, where the comparison signal is a reset signal taken in the reference pixel at the same time that the charge accumulated signal is taken in the desired pixel. Thus, a reset value is taken twice from the reference pixel; once during reset sampling phase, e.g., the reset value, and once during signal sampling phase e.g., the comparison value,. In this manner common mode noise can be minimized by a combination of signals from the desired and reference pixels. In one exemplary embodiment of the invention, the reference pixel is the nearest neighbor pixel in the same row. In another exemplary embodiment, the reference pixel is the nearest neighboring pixel in a different row of the same column.

Problems solved by technology

This increases substrate noise coupling to a pixel, which can compromise the signal to noise ratio of the image sensor core.

The substrate noise occurs when spurious noise signals are injected locally into the substrate through ohmic or capacitive coupling, thereby breaking the equipotentiality of the substrate.

Since Vcm2 does not equal Vcm1, they do not cancel out in the differential signal, and instead adversely affect the pixel signal ultimately produced.

But this will cause a decreased fill factor for the pixels, and for some architectures will cause an increase in KTC (thermal) noise.

Images