Abstract: The present invention reduces the area on a die required for rows and columns of redundant memory cells by sharing compare circuitry with banks of redundant memory cells based on division of the primary memory array into two or more "planes." Pass gates or multiplexers coupled between at least two banks of fuses and one compare circuit selectively couple the appropriate fuse bank to the compare circuit. Preferably, a bit in the address (e.g., address bit RA9 in a row address word having address bits A0-RA9) is received by and controls the multiplexer to select between the two banks of fuses. Additionally, the planes span blocks of memory in the memory array, where each block is divided by shared sense amplifiers. As a result, while eight lines are coupled to 16 rows or columns, only eight rows or columns will be active at any one time because isolation gates will enable only eight of the 16 rows or columns within two planes of memory.

Abstract: A block write circuit in a memory device having a wide internal data path performs block write and data masking functions. The memory device includes a plurality of data terminals adapted to receive respective data signals, and a plurality of array groups each including a plurality of arrays and each array includes a plurality of memory cells. A plurality of input/output line groups each include a plurality of input/output lines coupled to the arrays of an associated array group. The block write circuit comprises a plurality of write driver groups, each write driver group including a plurality of write driver circuits having outputs coupled to respective data lines in an associated data line group. Each write driver circuit includes an input and develops a data signal on its output responsive to a data signal applied on its input. A multiplexer circuit includes a plurality of inputs coupled to respective data terminals, and a plurality of output subgroups.

Abstract: In a semiconductor memory device, a die architecture is provided that arranges memory arrays into a long, narrow configuration. Bond pads may then be placed along a long side of a correspondingly shaped die. As a result, this architecture is compatible with short lead frame "fingers" for use with wide data busses as part of high speed, multiple band memory integrated circuits.

Abstract: The invention is a dynamic random access memory (DRAM) device having an electronic test key fabricated on board and is a method for testing the DRAM. The electronic test key generates a signal which effects a variation in a period of an internal control signal to stress the DRAM during a test mode.

Abstract: A one-shot circuit comprises a pulse generation circuit having an external clock terminal adapted to receive an external clock signal, a reset terminal coupled to receive a reset signal, and an internal clock terminal on which the pulse generation circuit develops an internal clock signal. The pulse generation circuit operates in a first mode to drive the internal clock signal from a first level to a second level in response to a transition of the external clock signal from a first level to a second level. The pulse generation circuit operates in a second mode to drive the internal clock signal from the second level to the first level in response to the reset signal going active. A delay circuit has input and output terminals coupled to the internal clock terminal and reset terminal, respectively, of the pulse generation circuit. The delay circuit operates to drive the reset signal active a delay time after the pulse generation circuit drives the internal clock signal from the first level to the second level.

Abstract: A voltage generator is provided in a DRAM to generate the appropriate voltage for programming antifuses. In one embodiment of the present invention, a high voltage charge pump is provided on a DRAM product to generate the high voltage necessary for programming antifuses. In one embodiment, the high voltage charge pump includes a plurality of transistors in series to produce a programming voltage that is some multiple of the supply voltage to the memory less.

Abstract: A method and circuit for self-latching data read lines that are used to transfer data that is read from a memory array of a memory device to a data output register of the memory device, wherein a self-latching latch circuit is connected to each data read line. The latch circuits are located physically near the output of the memory array, for latching data that is read from the memory array as soon as the data is applied to the data read lines, and prior to the data being latched in the data output register, thereby minimizing the effects of propagation delay so that the memory cycle time can be decreased. In one embodiment wherein the memory is organized in a "x4" configuration, different groups of the data read lines are selected in alternate read cycles, and the data read lines of the non-selected data read group are equilibrated automatically during the read cycle using the conventional test circuits of the memory device.

Abstract: A maskable data output buffer includes an output stage receiving data signals from a data coder. The signals output from the data coder are normally complementary data signals corresponding to complementary data input signals. However, in response to receiving a mask signal, the data coder forces the output signals to be other than complementary. The output stage normally generates a data output signal corresponding to the complementary data input signals. However, when the data input signals are other than complementary, the output of the output stage assumes a high impedance condition. Since the timing of the high impedance condition is determined from the data signals themselves, the timing of the mask operation is inherently properly timed to the output of the data from the data output buffer.

Abstract: During compression mode testing of a semiconductor memory device, a memory address is compressed to free up 2 or more bits in the address (e.g., an 11-bit address is compressed to 9-bits, freeing up 2 bits). Redundant element enable circuitry is coupled to one or more pins on a packaged chip that are unused during the compression mode testing. The circuitry receives control signals from external testing circuitry to select between the primary memory array in the chip, and redundant rows and columns of memory in the chip. As a result, during compressed address mode testing of the chip, a full 11-bit word is input to test the circuitry, but where 2 of the 11 bits allow the external circuitry to toggle between, and thereby selectively access, the rows and columns of primary and redundant memory in the chip. Alternatively, the circuitry can also be coupled to a non-connected pin on the packaged chip so as to operate during a non-compression mode testing.

Abstract: An architecture for a multi-bank DRAM is described which utilizes banks which are staggered in order to increase the amount of data which can be accessed at any one time. The banks are staggered such that a portion of each bank is provided on opposite sides of a data path so that a single address can simultaneously specify both portions of the bank so that twice the amount of data can be accessed.

Abstract: A memory device is described which equilibrates and precharges the input/output lines synchronously during a write cycle and asynchronously during a read cycle. During a read cycle, the timing of equilibration and precharge functions are decoupled from the clock signal and asynchronously initiate the equilibrate and precharge functions of the I/O lines in response to a latch signal. The invention initiates the equilibration and precharging of the I/O lines earlier in the access cycle thereby increasing memory access speed.

Abstract: A semiconductor integrated circuit includes a plurality of programmable elements, each having a first terminal connected to a first power supply potential, and a second terminal. Each of a plurality of first semiconductor switching elements has a first terminal respectively connected to the second terminal of a corresponding one of the plurality of programmable elements and has a second terminal. Each of a plurality of second semiconductor switching elements has a first terminal connected in common to selected ones of the second terminals of the plurality of first semiconductor switching elements and has a second terminal connected to a second power supply potential.

Abstract: Selected transistors in a charge pump circuit have their associated well regions tied to a capacitor electrode. As a result, the body effect in these devices is reduced, and, consequently, the threshold voltage is reduced as well. With a lower threshold voltage, these transistors allow the charge pump to quickly generate a voltage higher than the positive power supply voltage or a negative substrate bias voltage. In addition, the metal-insulator-semiconductor (MIS) capacitors in the charge pump preferably have their source/drain regions tied to an associated well region, thereby shorting the source/drain/well region junction. Thus, parasitic capacitances associated with these MIS capacitors is significantly reduced, further increasing the speed of the charge pump circuit.

Abstract: A voltage level translator is disclosed which translates a CMOS input signal into a CMOS output signal where the low voltage level of the output signal is equal to the high voltage level of the input signal. The voltage level translator is described in an integrated circuit such as memory circuits, including DRAMs. Specifically, the voltage level translator produces an output signal which can be used as a gate voltage on a precharge transistor for a booted circuit where the gate voltage need only drop to the high voltage level of the input signal to shut the transistor off. The voltage level translator described, therefore, reduces the time and power required to translate an input signal by limiting the voltage swing of the output signal.

Abstract: In a semiconductor memory device, a die architecture is provided that arranges memory arrays into a long, narrow configuration. Bond pads may then be placed along a long side of a correspondingly shaped die. As a result, this architecture is compatible with short lead frame "fingers" for use with wide data busses as part of high speed, multiple band memory integrated circuits.

Abstract: An integrated circuit clock buffer is described which includes output circuits for generating internal clock signals in response to an externally provided clock signal. The clock buffer includes a latch circuit having a delayed feedback to provide a pulsed signal in response to a transition in the external clock signal. The output circuits have a trip point which is skewed in one direction to detect a rising transition of the pulsed signal, and are skewed in a second direction to detect a falling transition of the pulsed signal. The buffer generates two non-skewed internal clock signals which have sharp rising and falling transitions.

Abstract: A system and method for forming a memory having at least 16 megabits (2.sup.24 bits) and only a single deposition layer of highly conductive interconnects. The resulting semiconductor die or chip fits within existing industry-standard packages with little or no speed loss over previous double metal deposition layered DRAM physical architectures. This is accomplished using a die orientation that allows for a fast single metal speed path. The architecture can be easily replicated to provide larger size memory devices. In addition, a method is described for reducing parasitic resistance in an n-sense amplifier.

Abstract: A semiconductor device includes a plurality of conductive layers that are formed on the substrate. Two electrically intercoupled sections of a read-only storage element, such as a fuse element, which together compose the storage element, are each formed in a different one of the conductive layers. The storage element has a storage state, and each section has a conductivity. One can change the storage state of the storage element by changing the conductivity of one of the sections. Additionally, multiple storage elements may be coupled in parallel to form a storage module. Each of the storage elements of the storage module may include multiple storage sections that are each formed in a different conductive layer. The storage elements may store the version number of the mask set used to form the semiconductor device. Alternatively, a conductive layer is formed on a substrate, and one or more read-only storage elements are formed in the conductive layer.

Abstract: A memory chip containing a dual bank memory system is arranged to be mounted cross-wise in its package with the major axis of the memory chip extending along the minor axis of the package. The data output register and the chip bond pads are located between the two memory banks so that the data read/write lines extend only through a portion of the memory chip to the data output register. All of the address chip bond pads are located in one row and all of the data chip bond pads are located in another row that extends in parallel with the row of address chip bond pads. Also, the column decoder circuit for each memory bank is located at the center of the memory bank. This allows the column select lines to be segmented into two groups, with one group of column select lines extending from the center of the memory array outwardly toward one side of the memory array and with the other group of column select lines extending from the center of the memory array outwardly toward opposite side of the memory array.

Abstract: DRAM read accessing circuitry having two parallel connected control lines, one of which includes a level translator stage and the other of which includes an enable gate. Both the level translator stage and the enable gate are connected to receive 0.0 to 3 volt small logic swings from output buffer logic utilized in reading data out of the DRAM. Output signals from the level translator stage are applied to a pull up output transistor in an output driver stage, and output signals from the enable gate are connected to a pull down output transistor in the output driver stage. A feedback connection is provided between the output of the level translator stage and one input to the enable gate to ensure that the enable gate does not generate an enabling output signal for turning on the pull down output transistor until the pull up output transistor is completely turned off.