International Business Machines Corporation

Hyun J Shin

Sang Hoo Dhong

Peter Hong Xiao

Hon-Sum Philip Wong

Jungwook Yang

Sudhir Muniswamy Gowda

Wei Hwang

Louis L Hsu

Somnuk Ratanaphanyarat

Toshiaki Kirihata

Jente Benedict Kuang

Ching-Te Kent Chuang

Mary Joseph Saccamango

Taqi Nasser Buti

Chih-Liang Chen

Yoichi Taira

Toshio Sunaga

Lewis M Terman

Disclosed is an image sensor including one or more dummy pixels that produce a reference signal which is used to compensate for errors within the devices of the main pixel cells. In one embodiment, at least one dummy pixel is used in conjunction with other circuitry to correct for nonlinearities in the transfer characteristic of a source follower transistor within each pixel. In another embodiment, an array of dummy pixels is used to correct for leakage current within the pixels during an electronic shutter mode of operation. The two techniques can be combined whereby both threshold voltage mismatch and leakage current are compensated for.

Image sensor with dummy pixel or dummy pixel array

The present invention relates to an automatic gain control circuit in which the automatic gain control function is performed entirely in the digital domain. In an illustrative embodiment, the digital automatic gain control circuit for an image sensor having associated therewith an analog-to digital (A/D) converter for converting analog electrical signals from the image sensor to corresponding digital codes, includes a min/max detector for determining minimum and maximum electrical signal values from the digital codes of the A/D converter for each frame of image. A filter coupled to the min/max detector dampens instantaneous changes of the minimum and maximum values by filtering to provide filtered minimum and maximum values. A digital-to-analog (D/A) converter coupled to the filter generates minimum and maximum analog reference voltages corresponding to the respective minimum and maximum filtered values, the reference voltages being applied to the A/D converter to control associated amplitudes of the digital codes provided thereby.

Digital automatic gain control circuit for image system

A CMOS image sensor is provided in which correlated double sampling is performed entirely in the digital domain. In an exemplary embodiment, the image sensor includes a plurality of imager cells arranged in rows and columns, where the imager cells of a particular column are coupled to a column data line of that column. Each active imager cell is capable of selectively providing a first output on an associated column data line indicative of a reset level during a first sampling interval. During a second sampling interval, each active imager cell provides a signal output on the associated column data line indicative of an amount of light incident upon that imager cell. At least one analog to digital (A/D) converter is coupled to the column data lines and converts the first and signal outputs on each column data line to first and second digital codes, respectively, to complete a correlated double sampling operation. The invention eliminates the need for analog capacitors to store the reset and signal levels.

Image sensor with direct digital correlated sampling

Disclosed is an image sensor having A/D conversion circuitry coupled to column data lines of an image sensor array. The A/D conversion circuitry digitizes analog signals on the column data lines, each representing intensity of light incident upon an active imager cell. Higher resolution is provided for darker light levels than for bright light levels, such that a high resolution image is obtained with less storage data than would otherwise be required. In one embodiment, the A/D conversion circuitry includes a plurality of comparators, each having a first input coupled to one or more column data lines and a second input coupled to receive a time-varying reference signal, and a plurality of n-bit counters coupled to the comparator outputs. An n-bit to m-bit converter nonlinearly maps n-bit codes to m-bit codes and provides the m-bit codes to an m-bit D/A converter which produces the time-varying reference signal. In another embodiment, the A/D conversion circuitry is comprised of a non-uniform successive approximation A/D converter.

Image sensor employing non-uniform A/D conversion

Disclosed is an image sensing device having a reduced number of transistors within each imager cell as compared to prior art devices. Each imager cell includes a photosensitive element providing a photocharge responsive to incoming light, and first, second and third transistors. The first transistor is coupled to an activation line, e.g., a row select line, that carries an activation signal to a first plurality of imager cells to selectively activate cells for image data readout. This transistor transfers the photocharge towards a reference circuit node within the image cell in response to the activation signal. The second transistor is operably coupled to the first transistor, and is operative to selectively set a voltage level at the reference node. The third transistor has a control terminal coupled to the reference node, and an output terminal coupled to an output data bus common to a second plurality of image cells, e.g., a column of cells. The third transistor providing an output signal on the data line related to the reference node voltage, which is indicative of an amount of light incident upon the photosensitive element.

Image sensor pixel circuit

Disclosed is a circuit for performing correlated double sampling entirely in the digital domain. In an exemplary embodiment, the circuit includes a plurality of comparators, each having a first input coupled to an associated data line for receiving first and second signals in first and second sampling intervals, respectively. A time varying reference signal is applied to the second input of each comparator. A plurality of up/down counters are coupled to respective ones of the comparators, and each is operable to count in a first direction during the first sampling interval and in an opposite direction during the second sampling interval. Each up/down counter is caused to stop counting when the amplitude of the variable reference signal substantially equals the amplitude of the respective first or second signal. As a result, each up/down counter provides an output representing a subtraction of one of said first or second signals from the other. The invention has particular utility when used in conjunction with a CMOS image sensor.

Correlated double sampling with up/down counter

An apparatus for controlling the ramp-up rate of a charge pump having an output providing an output voltage and an output current. In one embodiment, the apparatus comprises a current bleeder circuit having an input, an output adapted for connection to ground potential and at least one transistor having a gate, source, drain and body and defining at least one current path between the source and drain to form a current path between the input and output. The body is adapted for connection to the charge pump output. The apparatus further comprises a control circuit having an input adapted for connection to the charge pump output and an output connected to the bleeder circuit input. The control circuit provides a voltage potential to the input of the current bleeder circuit to control the gate-to-source voltage of the current bleeder circuit transistor. The flow of current through the current path of the current bleeder path is a function of the magnitude of the charge pump output and the gate-to-source voltage of the bleeder circuit transistor. Other embodiments of the apparatus of the present invention are described herein.

Ramp-up rate control circuit for flash memory charge pump

A discharge circuit for a semiconductor memory includes a first node, a second node for receiving a control signal having first and second states, and a circuit connected between the first node and ground potential and to the second node. The circuit couples the first node to ground potential when the control signal has the first state and substantially isolates the first node from ground potential when the control signal has the second state. The circuit includes a first subcircuit for defining a current path between the first node and ground potential. The first subcircuit includes a plurality of transistors connected in series, each of which having a gate, source and drain. The circuit further includes a second subcircuit for effecting predetermined gate-to-source, and drain-to-source voltages of the transistors of the first subcircuit when the control signal has the second state.

Discharge circuit in a semiconductor memory

High performance on-chip voltage regulator designs are disclosed which have settling times which are fast enough to meet today's microprocessor/microcontroller requirements when they are entering an active mode from a passive mode. A first preferred embodiment provides a circuit in which a single pulse control signal is required to instantly raise Vy when the microprocessor is in the wake-up period. The circuit includes a charge pump, a differential amplifier, and a microprocessor connected to the power supply through a voltage regulating device. A second embodiment provides a circuit to stimulate Vint prior to CPU wake-up. The principle of operation of this embodiment is to stimulate the voltage regulating device prior to CPU wake-up. By stimulating (pulling down) the Vint node, the voltage regulating device will raise Vy and ready the microprocessor to draw a large current.

High performance on-chip voltage regulator designs

A self-biased, fully differential, complementary receiver apparatus and method is presented. The receiver accepts differential inputs that can vary over the full rail-to-rail common mode voltage range. It produces double-ended complementary outputs swinging rail-to-rail useful in signal level conversion and comparator applications. The receiver includes a dual, fully complementary and mirror-symmetrical arrangement of a differential input stage, a biasing stage and an output stage. A self biasing voltage is generated with a balanced voltage divider coupled between the outputs of the biasing stages. This frees both biasing outputs for use as analogous but complementary receiver outputs while providing the receiver with all the advantages of self bias. For small signal differential inputs, the input and biasing stages operate in their linear region useful for amplifier applications. Whereas the circuit is most advantageously implemented using both p-type and n-type CMOS transistors, it can similarly be advantageously implemented with bipolar transistors. Alternate circuit configurations are described.

Fully differential self-biased signal receiver

A voltage regulator is provided for controlling an on-chip voltage generator which produces a boost voltage across a charge reservoir for supply to one input of a plurality of word line drivers in a memory array. The regulator is configured such that the charge reservoir voltage will track the power supply voltage and the difference between the power supply voltage and the charge reservoir voltage will be maintained substantially constant over a predefined power supply range. The voltage regulator includes a bandgap reference generator, a first differential circuit for producing a transition voltage from the reference voltage and the power supply voltage, a first transistor for comparing the power supply voltage with the boost voltage, a second transistor for comparing the transition voltage with the reference voltage and a latching comparator for equating the signal outputs from the first and second transistors so as to define a control signal for the on-chip voltage generator. Along with further specific details of the voltage regulator, a preferred bandgap reference generator is described.

Bandgap voltage reference generator

A pull-down circuit for a TTL compatible data output buffer uses NMOS devices. The pull-down circuit comprising two NMOS stages. Namely, a diode configuration stage where the gate and drain electrodes are shorted together during pull-down and a common-source stage. Both PMOS and NMOS devices are used for shorting the gate and drain electrodes.

Data output buffer pull-down circuit for TTL interface

Power supply tracking regulator for a memory array

High speed, low power signal switching logic implementable in bipolar or BiCMOS technology is described. Signal switching is between a first prescribed state and a second prescribed state and is accomplished using a conventional active signal pull-up circuit in combination with a novel self-biased, feedback-controlled active signal pull-down circuit. The active signal pull-down circuit is driven by a feedback signal obtained from the active pull-up circuitry such that only a single input connection to the signal switching circuit is required. Preferably, the active signal pull-up circuit comprises an emitter follower coupled transistor and the drive signal for the active signal pull-down circuit is taken from the collector thereof via a dc-coupling level shifter. Various signal switching circuit embodiments are described.

Self-biased feedback-controlled active pull-down signal switching

A sensing technique uses a variable precharge voltage sensing with a single bitline swing in a DRAM cell or array of DRAM cells so that the power dissipation is reduced. The bitline precharge voltage varies from one RAS cycle to the next RAS cycle depending upon the level of the data in the accessed cells. Such an arrangement eliminates the need for a reference voltage generator since the precharge voltage is not the same voltage for each RAS cycle.

Growing community of inventors

Mahopac, NY, United States of America

Hyun J Shin