Abstract: A dynamic memory obtains reduced leakage currents through the access transistors by preventing the low-going column conductors from reaching zero volts for at least a majority of the duration of the active portion of a memory cycle. The low-going conductors are allowed to reach zero volts during the refresh operation. One advantage is a possible increase in the data storage time between required refresh operations. An increase in the refresh interval is especially useful for memory operations wherein a multiplicity of columns are selected for a given row selection. The present technique also addresses the tendency toward increased sub-threshold leakage as field effect transistor thresholds decrease.

Abstract: A memory having multiplexed address inputs uses a column decoder which is deactivated during row address time and becomes activated during column address time. Access time and power dissipation are reduced since the column decoder need not be fully recovered after row address information has terminated and column address information is available.

Abstract: The present invention relates to the inclusion of an additional sense amplifier (100) on each column of a dynamic random access memory (RAM). The second sense amplifier is located near the input/output (DQ) line and functions to increase the rate of discharge of the selected column pair (C.sub.n, C.sub.n) thereby improving the transfer of logic information from a selected memory cell (M) to the input/output line associated therewith. The second sense amplifier in the column of the selected memory cell is activated by the same pulse (CCDQ) which connects the selected column to the input/output line, where only the second sense amplifier associated with the accessed column is activated during a single read/write cycle.

Abstract: A memory of rows and columns of memory cells uses a multiplexed input address buffer having output row-column address lines which are coupled to a multiplexer and to column decoders. The multiplexer is coupled to row address decoders and serves to selectively couple the address lines to the row decoders. The address lines typically first carry row address information and then column address information. The use of a common portion of the address lines to couple the address buffer to the column decoders and multiplexer tends to reduce the overall size of the memory and thereby increases yield and reduces cost.

Abstract: A folded bit line configured DRAM, with standard even and odd rows of memory cells, also includes spare even and odd rows of memory cells which can be substituted for standard rows found to have defective cells or interconnections. Each of the decoders associated with a standard row includes provision for being disconnected if found to be associated with a defective row. One common spare decoder is associated with one spare even and one spare odd row of memory cells. Each spare decoder is designed normally to be deselected for any address but to be able to assume the address of any disconnected standard row. Disconnection of a standard decoder and substitution of a spare decoder with the appropriate even or odd row are made possible by appropriate inclusion of fusible links which are selectively opened by laser irradiation. The use of one spare decoder with both an even and odd row serves to reduce the number of needed spare decoders and thus reduces overall chip size.

Abstract: A memory in which each cell comprises an MOS transistor merged with a storage capacitor and in which the cells are arranged to permit adjacent pairs of transistors in a common column to share a common source and the transistors in a common row to share a common gate electrode conductor. The memory uses a first polycrystalline silicon layer which is patterned to provide interconnected storage electrodes and a second polycrystalline silicon layer which is patterned to provide a plurality of stripes to serve as the bit sense lines and a plurality of gate electrodes.

Abstract: An MOS dynamic random access memory (RAM) includes an array of memory cells arranged in rows and columns. The array is divided into two or more sub-arrays. In an operating cycle where a cell is being accessed for reading and/or writing, only the sub-array containing the accessed cell is fully selected while the other sub-arrays are partially selected. A fully selected sub-array is one in which both a row and a column are selected, whereas in a partially selected sub-array, only a row is selected. In the partially selected sub-array where only refreshing of the cells in the selected row takes place, the column decoders and drivers remain inactive throughout the memory cycle.

Abstract: A memory is provided with standard rows and columns and spare rows and columns for substitution for standard rows and columns found to have defective cells. Each of the decoders associated with a standard row and/or column includes provision for being disconnected if found to be associated with a defective row or column. Each of the decoders associated with a spare row and/or column is designed normally to be deselected for any address but to be able to assume the address of any disconnected row or column. Disconnection of the standard decoders and substitution of the spare decoders are made possible by appropriate inclusion of fusible links which can be selectively opened by laser irradiation.

Abstract: An MOS dynamic random access memory (RAM) includes an array of memory cells arranged in rows and columns. The array is divided into two or more sub-arrays. During an operating cycle latching of the sense amplifiers in the sub-arrays is staggered to avoid coincidence of current peaks each arising when the sense amplifiers in one of the sub-arrays are simultaneously latched. Latching takes place first in a sub-array in which a cell is selected. Recovery of the column conductors in the sub-arrays is also staggered to avoid coincidence of current peaks each occurring when one of the sub-arrays is recovered. The sub-array in which a cell is selected is recovered last.