Abstract: A compact, multi-state field effect transistor (FET) cell having a gate with edge portions of a different conductivity type than a central portion of the gate. Both the edge portions and the central portion extend from the source to the drain of the multi-state FET device. This device would have two different threshold voltages (V.sub.T), one where the central portion would turn on first, followed by the edges for the entire gate width to be active to give a second level of current flow. Such devices would be useful in building very compact or high density multi-state read-only-memories (ROMs).

Abstract: A system for interfacing a Processor to a Coprocessor using standard bus cycles. The Processor, upon encountering in its instruction stream an instruction having a particular Operation word format, will transfer a Command word following the Operation word to a particular Coprocessor designated by a Coprocessor Identity field in the Operation word. Upon decoding the Command word, the Coprocessor will respond with any of a set of response primitives which define functions which the Coprocessor requires to Processor to perform in support of the Command by the Coprocessor. The interface provides for all functions which the Coprocessor may require, including selective vectoring to appropriate exception handlers.

Abstract: A process of forming a conductive material layer in at least two steps by forming a conductive material layer from a plurality of thin layers of conductive material. The use of a two-step formation process for the conductive material layer permits process versatility in incorporating implantation steps and patterning steps between formation of the thin layers of conductive material. Direct transfer from dielectric layer formation to conductive material layer formation steps, and performing the intermediate process steps in the same piece of equipment as the thin conductive layer formation assists in adhesion of the thin layers to each other to form the total conductive material layer. The use of in situ doped semiconductor material, such as in situ doped polycrystalline silicon and in situ doped amorphous silicon reduces the exposure of other dopants that may be present to thermal cycles of high temperature, greater than 900.degree. C., that causes these dopants to migrate undesirably.

Abstract: A bulge well structure for trench devices in wells of a conductivity type opposite to that of the substrate where the bottom of the trench has localized, extra doping. The additional doping into the bottom of the trench prior to device formation may be implanted while the photoresist mask for the trench formation is still in place. In one embodiment of the method, the trenches and the bulge or well extension formations at their bottoms are created before isolation regions are formed. The structure and method permit increased doping only where needed and are compatible with thin epitaxial layers and sharp transition interfaces of epitaxy with substrate for optimum latchup protection. No extra masks are required and the tight packing allowed by trench technology is not altered. Protection against soft errors and junction leakage by forming DRAM trench capacitors in a well of opposite conductivity type from the substrate may be provided.

Abstract: Status flag handling method and apparatus for use in a digital data processing system provide error-resistant operation and simplicity. Two storage elements comprising one bit of a status register are operated such that: a reset places both elements in first predetermined states; a set flag operation places both elements in second predetermined states; a read flag operation alters the state of the second storage element; and a clear flag alters the state of the first storage element if and only if the state of the second storage element has previously been altered by a read flag operation. The flag output corresponds to the state of the first storage element. When implemented with single instructions, inadvertant flag negation and errors due to intervening interrupts are avoided. The read flag operation temporarily disables the set flag mechanism, protecting against setting the flag during a read operation. The flag is always read as asserted prior to being negated.

Abstract: This disclosure relates to a process for etching polycrystalline silicon in preference to single crystal silicon. Polycrystalline silicon is anisotropically etched in a plasma which inclues a noncarbonaceus silicon etching compound such as chlorine together with about 0.4-1.5 percent by volume of oxygen. The process is used to fabricate semiconductor devices which require the etching of polycrystalline silicon in the presence of exposed monocrystalline silicon.

Abstract: A pipelined multiplier which serially receives a signed input multiplicand and a signed multiplier to generate a signed serial output product is provided. The multiplier utilizes a technique which simplifies the addition of partial product bits by creating a uniform partial product array. Columns of partial product bits are sequentially added in a pipelined structure. Carry bits which are generated during the column addition of partial product bits are delayed in the pipeline and coupled back to the input of the pipeline at the appropriate time for another addition of column bits as product bits are serially outputted. By minimizing delays in the pipeline, multiplication of signed operands of large bit length may be quickly performed.

Abstract: A switched capacitor filter includes an amplifier in a feedback or feed-forward path thereof. The amplifier (A5, C6, S7,) has a gain dependent on the ratio of stray capacitances to the capacitors used in the filter so as to compensate for errors in filter characteristics which said stray capacitances would otherwise introduce.

Abstract: A write protect mechanism for a programmable read-only memory prevents writes to the PROM unless a protect register contains predetermined information. The protect register is itself a write protected control register. The predetermined information cannot be written into the protect register except during a short, predetermined period after the occurrence of an event such as a reset. The protect register may be written to with information other than the predetermined information at any time. The preferred embodiment comprises a single-chip microcomputer with on-board electrically-erasable programmable read-only memory which is write protected in several, separate blocks.

Abstract: A single-chip microcomputer comprises a CPU (1), a RAM (2), an EPROM (3), a timer (4), serial I/O communication logic (5), four I/O ports (11-14) and an input pin for providing a RESET signal or programming potential V.sub.PP (23). An EPROM control register (53) in the CPU is loadable under the control of a computer program stored in an external memory (40). Responsive to a first bit in the EPROM control register an address buffer/latch 61 and a data latch 62 temporarily latch address and data information during a write operation to the EPROM. Responsive to a second bit in the EPROM control register the programming potential is applied to the EPROM for a predetermined time to program the data information into the EPROM at its associated address. The accuracy of the programming operation may be verified using the CPU under the control of the external computer program, to compare the address and data information programmed into the EPROM with the original source of such information.

Abstract: A DRAM memory cell has a trench capacitor and a transistor. The trench of the trench capacitor penetrates to a buried layer which acts as the primary portion of one of the plates of the capacitor. When the buried layer is the same conductivity type as the transistor of the memory cell, the buried layer is biased to a voltage selected to reduce the maximum voltage across the capacitor. This allows for a reduction in the thickness of the dielectric which coats the trench which increases the capacitance of the capacitor. When the buried layer is of the opposite conductivity type from the transistor type of the memory cell, there is no parasitic MOS transistor formed between the primary portion of the capacitor plate and the source of the transistor of the memory cell.

Abstract: A method for directly providing a conversion of an analog input signal to a digital signal in two's complement code with a sampled data converter. Positive and negative reference voltages and an analog ground voltage are required. After a sign bit determination of the input signal is made, the data converter is coupled between either a first pair of reference voltages or a second pair of reference voltages depending upon the sign bit. The first pair of reference voltages comprises the positive reference and ground reference, and the second pair of reference voltages comprises the ground reference and a negative reference. By selectively coupling the chosen reference voltages to the converter, a converter may directly output two's complement code.

Abstract: A memory has an address buffer which receives a row address and a column address and outputs these buffered address signals to a predecoder. A row decoder and column decoder use predecoded signals provided by the predecoder to select a row and a column from a main array. A redundant row is provided to replace a defective row from the main array. A programmable redundant decoder is programmable to select the redundant row in response to the predecoder signals which select the defective row.

Abstract: A process for making hermetic, low cost pin grid array (PGA) semiconductor die packages. The process involves die bonding a semiconductor die or integrated circuit chip to a substrate having an interconnect or metallization pattern thereon. The die is electrically connected to the pattern and then the die and the inner bonds are hermetically sealed inside a cap that is smaller than the substrate so that the ends of the metallization pattern are exposed. The leads are then electrically connected, such as by solder or other technique to the exposed ends of the pattern.

Abstract: A CMOS sense amplifier for use in a memory comprises two CMOS differential amplifiers. Each differential amplifier receives the same two signals generated from a selected bit line pair and each provides a different one of a complementary pair of signals. Each differential amplifier has a current mirror for loads. Each differential amplifier uses a transistor current source. A transistor will operate as a more ideal current source if it is in saturation. The transistor current source is biased by the current mirror of the differential amplifier of which it is a part. The resulting differential amplifier thus has a transistor current source which is biased closer to saturation than if biased by a normal clock signal which is either at the high or low power supply voltage. The self-biasing aspect avoids the problems associated with generating a special reference voltage for the differential amplifier.

Abstract: Adjusting field effect transistor (FET) threshold voltage (V.sub.T) by diffusing impurities in polysilicon gates through a refractory metal silicide. Dopants of different conductivities may be cross-diffused. This adjustment can be made relatively late in the fabrication of the wafers to provide a quick turn around time of custom circuits, gate arrays and application specific integrated circuits (ASICs). A masking step selectively provides blocking elements to prevent the diffusion from occurring in certain of the FETs.

Abstract: A filter structure uses multiple discrete filter circuits which are cascaded to provide a multiple tap filter of programmable tap length. In one form, an FIR filter may be implemented wherein each circuit generates partial sum operands which must be added to provide a filter output. The cascaded circuits perform partial addition operations near simultaneously by using a serial addition which is synchronized with a start bit. The number of taps in the filter structure implemented by the cascaded discrete filter circuits is variable and may be programmed with a programmable storage register in each discrete circuit which stores operand data fixing the tap length of each discrete circuit. The multiple filter circuits provide a single filter structure with a large tap length and high sampling rate.

Abstract: A CMOS circuit having a differential input stage which provides a single output is provided. An output stage has a capacitor which is used as a Miller integrator coupled thereto for frequency stabilization. A cascode portion is coupled to the Miller integrator to maintain one of the capacitor's electrodes at a predetermined voltage potential. A compensation portion is coupled to the cascode portion to compensate for power supply induced errors created when either an N-channel transistor in an N-well process or a P-channel transistor in a P-well process is used in the cascode portion.

Abstract: A data interface circuit for use when interfacing between two communication links communicating frames of digital data in PCM and ADPCM formats is provided. The data interface circuit provides control information for selecting one of a plurality of algorithms to control the transformation of data between a plurality of PCM and ADPCM formats. A single encoded control signal is utilized to establish frame boundaries and to select a predetermined one of the plurality of algorithms to use in converting between PCM and ADPCM data.

Abstract: An analog two pole filter is provided which uses a single amplifier to implement a predetermined transfer function. The filter has a differential input and converts the two inputs to a single output utilizing the same amplifier which performs the filtering function. By coupling a capacitor across the differential input and utilizing the differential aspect of the input signals, the capacitor may be implemented with half the capacitance otherwise required to implement the predetermined transfer function, thereby minimizing circuit area.