摘要:
A semiconductor memory device includes a plurality of sense amplifiers for amplifying current changes which occur in corresponding bit line pairs in accordance with binary signals stored in activated memory cells. Each of the sense amplifiers includes first and second current mirror circuits for generating currents of the magnitudes respectively corresponding to currents flowing through a corresponding bit line pair, a storing circuit, responsive to a signal selecting a memory cell, for storing the currents generated by the first and second current mirror circuits before activation of the memory cell, or a difference between these currents, and a current supplying circuit, responsive to activation of the memory cell and based on the amount stored in the storing circuit, for supplying, to the first and second current mirror circuits, currents having a predetermined relationship with the currents having been generated by the first and second current mirror circuits before activation of the memory cell. A current change corresponding to data stored in the selected memory cell and not affected by an offset occurs in a connection node between the first current mirror circuit and the current supplying circuit, or a connection node between the second current mirror circuit and the current supplying circuit.
摘要:
A method of manufacturing a sealed electronic component, which can seal a housing in a high-vacuum state while preventing enclosure of a gas within the housing, as well as achieving the improvement in manufacturing efficiency. According to the method, after forming an unwelded section by a primary welding process step, including a first beam irradiation process step and a second beam irradiation process step, annealing treatment is performed in an annealing process step by irradiating an electron beam to a predetermined portion on a locus of the electron beam formed in the first beam irradiation process step. The locus may be on a housing or a lid.
摘要:
An internal address signal corresponding to data to be written into a memory cell is held in a latch circuit. The held internal address signal is selected by a multiplexer in the next writing operation and applied to a decoder. Write data is taken in and held by the latch circuit during the period in which data is not being read out from the memory cell array. A comparator compares the held internal address signal and an internal address signal for reading data. If a matching is found between them, the multiplexer outputs data from the latch circuit for external output. Accordingly, delay of a writing operation following a reading operation can be eliminated without increasing chip cost, package cost, and system cost, as a result high speed operation of cache memories is achieved and the speed performance of computers of various levels such as supercomputers, large size calculators, work stations and personal computers can be improved.
摘要:
Sense circuits are provided correspondingly to bit line pairs provided corresponding to memory cell columns, respectively. The sense circuit senses, amplifies and latches storage data of the selected memory cell, and information latched by the sense amplifier is rewritten into the selected memory cell after selection of the memory cell. Thereby, destruction of storage information of the memory cell is prevented.
摘要:
A pair of driving bipolar transistors of a lateral type T1 and T2 have emitters coupled to a ground potential, collectors connected to a pair of highly resistive elements R1 and R2. Highly resistive elements R1 and R2 have respective other ends coupled to power supply potential V.sub.CC, and bases and collectors of transistors T1 and T2 are cross-connected to each other, thereby forming a flipflop circuit. Access MOS transistors Q3 and Q4 having a gate potential controlled by word line WL are each connected to form a conduction path between one of storage nodes A and B and one of the pair of bit lines BL and /BL.
摘要:
In the present invention, a row address processing circuit and a column address processing circuit operate in synchronism with an externally applied synchronous signal in a semiconductor memory device. The row address processing circuit and the column address processing circuit each include an address buffer and a decoder. The address buffer or decoder operates in synchronism with the synchronous signal.
摘要:
An internal address signal corresponding to data to be written into a memory cell is held in a latch circuit. The held internal address signal is selected by a multiplexer in the next writing operation and applied to a decoder. Write data is taken in and held by the latch circuit during the period in which data is not being read out from the memory cell array. A comparator compares the held internal address signal and an internal address signal for reading data. If a matching is found between them, the multiplexer outputs data from the latch circuit for externally output. Accordingly, delay of a writing operation following a reading operation can be eliminated without increasing chip cost, package cost, and system cost, as a result high speed operation of cache memories is achieved and the speed performance of computers of various levels such as supercomputer, large size calculator, work station and personal computer can be improved.
摘要:
A cache memory device includes a plurality of memory cell arrays each including a plurality of memory cell rows, a plurality of first fuse elements each provided corresponding to each memory cell row and disconnected when the corresponding memory cell row is defective, and a plurality of second fuse elements each provided corresponding to each memory cell array and disconnected when the corresponding memory cell array is defective. As a result, the cache memory device can indicate that, when a bit line of a certain memory cell array is defective, the memory cell array is defective by disconnecting a second fuse element corresponding to the memory cell array.
摘要:
A plurality of bit line signal IO lines L1, /L1, . . . Ln and /Ln are arranged on a memory cell array. These bit line signal IO lines are arranged to cross respective bit lines BL1, /BL1, . . . BLn and /BLn, and are connected to the corresponding bit lines, respectively. Each bit line signal IO line has an end extended to an end, in a direction perpendicular to the bit line, of a memory cell array, and is coupled at the end to a bit line peripheral circuit. Although bit line peripheral circuits could be arranged only at upper and lower ends of the bit lines in the prior art, the bit line peripheral circuits can be arranged also at the ends of the bit line signal IO lines in the invention. This can increase a degree of freedom in a layout for the bit line peripheral circuits, and thus the bit line peripheral circuits can be dispersedly arranged in a larger area.
摘要:
In a static random access memory, when address signals change, one-shot pulses are responsively generated. A detection signal obtained by ORing the one-shot pulses is employed as an equalize signal. Potentials of a bit line pair is equalized in response to the equalize signal. A write inhibiting signal having a pulse width larger than that of the equalize signal is generated by a pulse width increasing circuit. A write operation of data is inhibited in response to the write inhibiting signal.