Abstract: A PMOS non-volatile memory array using SONOS transistors having program and erase threshold voltages for representing digital logic states of zero and one and selected to optimize read disturb characteristics. The threshold voltages are linearly convergent and separated by at least 0.5 volts for a charge retention time of at least 10 years, with the threshold voltages defining a window wherein a read voltage for selected memory transistors can be held flat and not intersect the threshold voltages. The lower threshold is selected to be at a zero charge state for one of the two logic levels of the memory.

Abstract: A PMOS non-volatile memory array using SONOS transistors having program and erase threshold voltages for representing digital logic states of zero and one and selected to optimize read disturb characteristics. The threshold voltages are linearly convergent and separated by at least 0.5 volts for a charge retention time of at least 10 years, with the threshold voltages defining a window wherein a read voltage for selected memory transistors can be held flat and not intersect the threshold voltages. The lower threshold is selected to be at a zero charge state for one of the two logic levels of the memory.

Abstract: A two device floating gate MOS nonvolatile memory cell is disclosed including a floating gate memory device coupled to a select device wherein a thin tunnel dielectric region of insulation material between the substrate and floating gate of the memory device is located in an area above the channel of the memory device in the substrate and wherein an implanted region in the substrate to facilitate the tunneling of carriers in and out of the floating gate extends appreciably underneath the edges of the field oxide regions forming the periphery of the sides of the channel of the memory device. A select device is located in series with the memory device. A process for fabricating this memory cell is also disclosed wherein the doped tunnelling region in the substrate is defined and implanted prior to definition of the field regions.

Abstract: Individual CMOS floating-gate memory cells capable of storing data are arranged in an array structure and selected with horizontal and vertical access lines. Current flow through the array cells is measured, amplified, and then compared with an unprogrammed cell using the sense amplifier of the present invention. The sense amplifier tolerates increased variation in the characteristics of programmed or unprogrammed cells and therefore increases the manufacturing yields of the arrays. It additionally achieves fast accessing and sensing of the stored data.

Abstract: A two device floating gate MOS nonvolatile memory cell is disclosed including a floating gate memory device coupled to a select device wherein a thin tunnel dielectric region of insulation material between the substrate and floating gate of the memory device is located in an area above the channel of the memory device in the substrate and wherein an implanted region in the substrate to facilitate the tunneling of carriers in and out of the floating gate extends appreciably underneath the edges of the field oxide regions forming the periphery of the sides of the channel of the memory device. A select device is located in series with the memory device. A process for fabricating this memory cell is also disclosed wherein the doped tunnelling region in the substrate is defined and implanted prior to definition of the field regions.

Abstract: A semiconductor memory circuit having primary and redundant arrays with the capability of substituting the redundant arrays for defective primary arrays by address location.

Abstract: An on-chip checking and correction circuit extends the program/erase endurance of a semi-conductor memory array and catches data retention failures in individual memory cells of the array. For each data byte in the memory array, four parity bits are computed using a modified Hamming Code and stored in additional bit cells associated with that data byte. During read and write operations, check bits are calculated using the same Hamming Code and used to correct single-bit errors; error checking and correction is repeated if necessary up to a predetermined number of tries.