摘要:
An image sensor and method of sensing an image provides a variable amplifier at each pixel, thereby expanding the dynamic range of the sensor. An image sensor core (500) comprises a plurality of pixel sensors (510). Each pixel sensor preferably includes a photosensitive element (310), a variable amplifier (312), and an analog-to-digital converter ("ADC") (314). The ADC (512) includes a comparator (512) coupled to a latch/shift register (514). The comparator (512) receives the output of the variable amplifier (312) and a ramp signal and produces an output when the ramp signal exceeds the amplifier (312) output. The latch/shift register (514) latches a count signal corresponding to the ramp signal when triggered by the comparator (512). In operation, the photosensitive element (310) is exposed to light for a first integration period. The variable amplifier (312) amplifies the resulting photocharge by the maximum possible amount. The ADC (314) converts the photocurrent to a digital signal and latches the five most significant bits in memory (514). These five bits are provided in parallel to the variable amplifier (312) to set the amplification level. Then, the photosensitive element (310) is exposed to light for a second integration period and the resulting photocharge is amplified at the set amplification level. The output of the amplifier (312) is digitized by the ADC (314) and the latch/shift register (514) stores the 12-bit result. The five bits selecting the amplification level and the 12-bit result provide the image sensor with 17 bits of dynamic range.
摘要:
Vertical planar and non-planar insulated gate semiconductor device cells having improved ruggedness under drain avalanche conditions are disclosed. The cells employ high concentration implants which are strategically located in the central cell regions. The implants are effective to concentrate the electric field intensity and avalanche current flow in the central cell region and to prevent current flow into the base of a parasitic bipolar transistor, thereby preventing activation of the transistor. Both surface-peaked and subsurface-peaked implants are disclosed.
摘要:
Near-end echo-reduction is achieved by passing a transmit signal through a first filter bank and feeding a remote signal plus an echo of the transmit signal through a second filter bank, then subtracting a gain-coefficient-compensated (scaled) version of the filtered transmit signal from the filtered-composite-remote signal to obtain an echo-reduced receive signal. The echo-reduced receive signal is suitably employed by tone decoders, voice response mechanisms, and the like, in a telephone system, and avoids false alarms due to near-end echo. In one embodiment, the gain coefficients are fixed for each frequency of interest. In another embodiment, the gain coefficients are dynamically arrived at based on the filtered transmit signal and the echo-reduced receive signal. The two filter banks can be combined in one, multiplexed filter bank.
摘要:
A method of semiconductor fabrication, in which a single aperture is used to define both a thin oxide tunneling region and a drain diffusion region in a self-aligned fashion, produces a device structure suitable for use in an electrically-erasable read-only memory (EEPROM) cell. A gate oxide is grown, then a photoresist mask is formed having a slit for ion implantation into the drain diffusion region. The oxide within the slit is etched away, and ion implantations forms a drain diffusion region. After the mask is stripped away, a healing furnace cycle removes the implant damage. A thin tunnel oxide layer is grown over the drain diffusion region, and then a polysilicon floating gate is formed so that one edge of this gate intersects a portion of the area of tunnel oxidation so as to form a small region of tunnel oxide under the floating gate. The process sequence then reverts to a conventional MOS flow. The self-aligned drain diffusion region and tunnel oxide region can be used in a variety of EEPROM cell designs. One embodiment involves a double-polysilicon, single-metal process in which thre diffusion regions are used to form the EEPROM cell. The second layer of polysilicon overlies the floating gate and forms a control gate word line. The control gate and the floating gate overly the channel between the drain diffusion region and a common source diffusion region. The first layer of polysilicon is also used to form a select gate overlying the channel between the drain diffusion region and a select drain diffusion region. The metal layer provides contact to the select drain diffusion region and forms the bit line.
摘要:
The bit lines in an alternate-metal, virtual-ground (AMG) electrically programmable read-only-memory (EPROM), are formed by utilizing a plurality of field oxide regions and a plurality of pairs of dielectric/floating gate strips, which have the ends of each pair of strips connected together over a field oxide region, as an implant mask. By connecting together the ends of each pair of dielectric/floating gate strips, the width of the strips at the edges of the field oxide regions will remain constant. As a result, the isolation between adjacent bit lines, which is defined by the width of the strips, will also remain constant.
摘要:
A lead frame (32) for use in an integrated circuit package (30) is herein disclosed wherein the lead frame is produced by molding an electrically conductive material into a desired lead frame shape. There is also disclosed several possible arrangements for a lead frame (32) produced using the molding method including protrusions (36) on the lead frame (32) adapted to provide a mechanical connection to the integrated circuit package (30), a heat sink (38) molded as an integral part of the lead frame (32) with heat dissipating characteristics specific to the application in which the lead frame (32) will be used, and an arrangement which provides a heat conducting portion adapted to thermally connect a component attached to the lead frame (32) to an external heat sink.