Abstract: A non-volatile memory cell with increased charge retention is fabricated on the same substrate as logic devices using a single-gate conventional logic process. A silicide-blocking dielectric structure is formed over a floating gate of the NVM cell, thereby preventing silicide formation over the floating gate, while allowing silicide formation over the logic devices. Silicide spiking and bridging are prevented in the NVM cell, as silicide-blocking dielectric structure prevents silicide metal from coming in contact with the floating gate or adjacent sidewall spacers. The silicide-blocking dielectric layer may expose portions of the active regions of the NVM cell, away from the floating gate and adjacent sidewall spacers, thereby enabling silicide formation on these portions. Alternately, the silicide-blocking dielectric layer may cover the active regions of the NVM cell during silicide formation. In this case, silicide-blocking dielectric layer may be thinned or removed after silicide formation.

Abstract: A method and apparatus for selectively restoring data in a non-volatile memory array based on failure type. Weakened data and erroneous data are identified by performing two readings of a specific memory section. Alternatively, an error correction code is used after a first reading of data to identify erroneous data. The manner in which data is restored will depend on whether the data changed because of an erase failure or a program failure. If only a program failure occurred then the data will be reprogrammed without an intervening erase step. If the data experienced an erase failure, then the data will be erased prior to being programmed with correct data.

Abstract: A computer-based method of converting an analog integrated circuit design from a source technology to a target technology, by providing a computer readable source schematic file and a computer readable source layout file for the analog integrated circuit design in the source technology, providing a computer readable technology transfer file that includes conversion tables between properties of the source technology and properties of the target technology, converting the source schematic file in the source technology to a target schematic file in the target technology with the computer using the technology transfer file, and converting the source layout file in the source technology to a target layout file in the target technology with the computer using the technology transfer file.

Abstract: Through the use of a multi-step sweep, the present invention is capable of increasing the speed and improving the error resistance of a signal alignment. In a specific embodiment of the invention, a method for the signal alignment of a target signal and an adjustable signal is disclosed. The target signal is sampled using three or more phase shifted versions of the adjustable signal to obtain a group of target signal state values. Next, through reference to the group of target signal state values it is determined that an edge of the target signal lies between a first phase shifted version and a second consecutive phase shifted version. In response, the first phase shifted version is selected as the starting point for a second sweep. During the second sweep, the phase of the first phase shifted version is sequentially adjusted in relatively small incremental steps to minimize the phase difference relative to the target signal.

Abstract: A memory system including a plurality of ports and a memory core having a plurality of memory banks. Access requests received at the ports are broadcast to the memory banks. Multiple memory banks may be concurrently accessed in response to access requests received on different ports. A memory controller provides the access requests to the memory system, and ensures that a single memory bank is not concurrently accessed by different ports. All access requests are processed with the same latency. If the memory banks include memory cells that must be periodically refreshed, then the memory controller also provides refresh requests to the memory banks. Because multiple memory banks may be concurrently accessed in response to access requests provided on different ports, the memory banks can operate at a lower frequency than the ports.

Abstract: A method and apparatus for scaling an embedded DRAM array from a first process to a second process, wherein the scaling involves reducing the linear dimensions of features by a constant scale factor. From the first process to the second process, DRAM cell capacitor layout area is reduced by the square of the scale factor, while cell capacitance is reduced by the scale factor. The voltage used to supply the logic transistors is scaled down from the first process to the second process. However, the voltage used to supply the sense amplifiers remains constant in both processes. Thus, in an embedded DRAM array of the second process, sense amplifiers are supplied by a greater voltage than the logic transistors. This allows the sensing voltage of DRAM cells to be maintained from one process generation to another, while allowing memory size to scale with the square of the process scale factor.

Abstract: A memory circuit for holding a single binary value. A first bit cell holds one of a logical high value and a logical low value, and a second bit cell also holds one of a logical high value and a logical low value. Circuitry is provided for placing a logical high value in the first bit cell when the binary value in the memory circuit is to be a logical high value, and circuitry is provided for placing a logical high value in the second bit cell when the binary value in the memory circuit is to be a logical low value. In this manner, a logical high value exists within the memory circuit, whether the single binary value within the memory circuit is a logical high value or a logical low value. The difference between the two values of the binary value is which of the two bit cells holds the logical high value. Thus, this memory circuit can be sensed without the use of a sense amplifier.

Abstract: A method for selectively refreshing data in a nonvolatile memory array based on failure type detected by an error correction code. If the page is determined to be error-free, no refresh operation takes place. Otherwise, if single-error words on a page contain erased and programmed bit errors, then a refresh operation, consisting of an erase and program, takes place. The erase operation is skipped if single-error words on a page solely contain a program failure.

Abstract: A method and apparatus for selectively restoring data in a non-volatile memory array based on failure type. Weakened data and erroneous data are identified by performing two readings of a specific memory section. Alternatively, an error correction code is used after a first reading of data to identify erroneous data. The manner in which data is restored will depend on whether the data changed because of an erase failure or a program failure. If only a program failure occurred then the data will be reprogrammed without an intervening erase step. If the data experienced an erase failure, then the data will be erased prior to being programmed with correct data.

Abstract: A reference current generation circuit generates a first branch current that varies by a first percentage in response to variations in a first supply voltage and variations in transistor threshold voltage. The first branch current is mirrored to create a corresponding second branch current. A first portion (sub-current) of the second branch current is supplied through a first transistor, which exhibits the transistor threshold voltage wherein the first sub-current varies by a second percentage in response to the variations in the first supply voltage and variations in transistor threshold voltage, wherein the second percentage is greater than the first percentage. A second portion (sub-current) of the second branch current is supplied through a second transistor. The second portion of the second branch current is mirrored to create a reference current (IREF).

Abstract: A process flow for creating a non-volatile memory cell, the process flow including the steps of forming a doped well in a semiconducting portion of a substrate, forming a gate dielectric layer on top of the substrate, depositing a first polysilicon layer on top of the gate dielectric layer, patterning and etching the first polysilicon layer, selectively oxidizing the first polysilicon layer, implanting lightly-doped source/drain regions into the well, forming sidewall spacers adjacent the first polysilicon layer, implanting source/drain regions into the well, thereby forming a channel area, depositing a dielectric layer on top of the first polysilicon layer, depositing a second polysilicon layer on top of the dielectric layer, forming a masking layer on the second polysilicon layer, and etching both the second polysilicon layer and the dielectric layer using the masking layer.

Abstract: Comparator chain total offset, static and dynamic, is reduced by injecting a compensation quantity in at least one point in the chain of comparator components. The compensation quantity is determined by providing the comparator chain with calibration signals having equal values and evaluating the output states of the comparator chain. The compensation quantity is adjusted until the probabilities of high and low output states are substantially equal and a calibrated value for the compensation is determined.

Abstract: A method of forming an embedded DRAM cell having multiple-thickness gate dielectrics. An oxidation-enhancing dopant is selectively implanted into a well region in an area that is exposed by a first mask. A thermal oxidation step simultaneously produces the field dielectric for two distinct devices each having a different oxide thickness. The method is applicable to quad-density DRAM cells using fewer oxidation steps. The method is also applicable to planar DRAM cells, and does not require increasing the number of masks during the fabrication of planar DRAM cells.

Abstract: An embedded memory system that includes DRAM cells and logic transistors. The capacitor of the embedded memory responds to a positive bias voltage of ½ Vdd. The wordline driver of a p-channel access transistor applying the positive power supply voltage when the p-channel access FET is not being accessed and a voltage lower than the threshold voltage of the p-channel access FET is being accessed. For DRAM cells containing an n-channel access FET, the wordline driver applies either a negative voltage or the ground voltage to the n-channel access FET when the DRAM cell is not being accessed. A second voltage composed of Vdd and a boosted voltage is applied to the n-channel FET when the DRAM cell is being accessed.

Abstract: A method and apparatus for scaling an embedded DRAM array from a first process to a second process, wherein the scaling involves reducing the linear dimensions of features by a constant scale factor. From the first process to the second process, DRAM cell capacitor layout area is reduced by the square of the scale factor, while cell capacitance is reduced by the scale factor. The voltage used to supply the logic transistors is scaled down from the first process to the second process. However, the voltage used to supply the sense amplifiers remains constant in both processes. Thus, in an embedded DRAM array of the second process, sense amplifiers are supplied by a greater voltage than the logic transistors. This allows the sensing voltage of DRAM cells to be maintained from one process generation to another, while allowing memory size to scale with the square of the process scale factor.

Abstract: A method, apparatus, and system in which an embedded memory fabricated in accordance with a conventional logic process includes one or more electrically-alterable non-volatile memory cells, each having a programming transistor, a read transistor and a control capacitor, which share a common floating gate electrode. The under-diffusion of the source/drain regions of the programming transistor and control capacitor are maximized. In one embodiment, the source/drain regions of the programming transistor are electrically shored by transistor punch-through (or direct contact).

Abstract: A method for error detection and correction (EDC) includes: generating a complete EDC code in response to a data packet; distributing the complete EDC code among the data packet to create a plurality of bytes, each including a data portion from the data packet and an EDC code portion from the complete EDC code; storing the bytes in a memory module; retrieving the bytes from the memory module; forwarding the data portions of the bytes retrieved from the memory module to a requesting device; providing the data portions of the bytes retrieved from the memory module to an EDC functional block; providing the EDC code portions of the bytes retrieved from the memory module to the EDC functional block; and performing error checking and correction in the EDC functional block upon receiving the complete EDC code from the provided EDC code portions.

Abstract: A non-volatile memory system including an array of cells, each having an access transistor and a capacitor sharing a floating gate. The access transistors in each row are fabricated in separate well regions, which are independently biased. Within each row, the source of each access transistor is coupled to a corresponding virtual ground line, and each capacitor structure is coupled to a corresponding word line. Alternately, the source of each access transistor in a column is coupled to a corresponding virtual ground line. Within each column, the drain of each access transistor is coupled to a corresponding bit line. Select memory cells in each row are programmed by band-to-band tunneling. Bit line biasing prevents programming of non-selected cells of the row. Programming is prevented in non-selected rows by controlling the well region voltages of these rows. Sector erase operations are implemented by Fowler-Nordheim tunneling.

Abstract: A non-volatile memory system including an array of cells, each having an access transistor and a capacitor sharing a floating gate. The access transistors in each row are fabricated in separate well regions, which are independently biased. Within each row, the source of each access transistor is coupled to a corresponding virtual ground line, and each capacitor structure is coupled to a corresponding word line. Alternately, the source of each access transistor in a column is coupled to a corresponding virtual ground line. Within each column, the drain of each access transistor is coupled to a corresponding bit line. Select memory cells in each row are programmed by band-to-band tunneling. Bit line biasing prevents programming of non-selected cells of the row. Programming is prevented in non-selected rows by controlling the well region voltages of these rows. Sector erase operations are implemented by Fowler-Nordheim tunneling.

Abstract: A dual-port memory system is implemented using single-port memory cells. An access arbiter having a synchronization circuit is used to prioritize and synchronize the access requests associated with the two ports. The access arbiter can also prioritize and synchronize refresh requests, in the case where the single-port memory cells require refresh. Access requests on the two ports and the refresh requests can be asynchronous. The access arbiter synchronizes the various requests by latching the requests into first-stage registers when a row access signal (RAS) is activated, and subsequently latching the contents of the first-stage registers into second-stage registers after a selected delay.