A 16-Mb 400-MHz loadless CMOS four-transistor SRAM macro

We have used a 5-metal 0.18-μm CMOS logic process to develop a 16-Mb 400-MHz loadless CMOS four-transistor SRAM macro. The macro contains: (1) end-point dual-pulse drivers for accurate timing control; (2) a wordline-voltage-level compensation circuit for stable data retention; and (3) an all-adjoining twisted bitline scheme for reduced bitline coupling capacitance. The macro is capable of 400-MHz high-speed access at 1.8-V supply voltage and is 66% the size of a conventional six-transistor SRAM macro. We have also developed a higher-performance 500-MHz loadless four-transistor SRAM macro in a CMOS process using 0.13-μm gate length     

A 220-MHz pipelined 16-Mb BiCMOS SRAM with PLL proportionalself-timing generator

This 512 Kw×8 b×3 way synchronous BiCMOS SRAM uses a 2-stage wave-pipeline scheme, a PLL self-timing generator and a 0.4-μm BiCMOS process to achieve 220 MHz fully-random read/write operations with a GTL I/O interface. Newly developed circuit technologies include: 1) a zig-zag double word-line scheme, 2) a centered bit-line load layout scheme, and 3) a phase-locked-loop (PLL) with a multistage-tapped ring oscillator which generates a clock cycle proportional pulse (CCPP) and a clock edge lookahead pulse (CELP)     

A 60-MHz 240-mW MPEG-4 videophone LSI with 16-Mb embedded DRAM

A 240-mW single-chip MPEG-4 videophone LSI with a 16-Mb embedded DRAM is fabricated utilizing a 0.25-μm CMOS triple-well quad-metal technology. The videophone LSI is applied to the 3GPP 3G-324M video-telephony standard for IMT-2000, and implements the MPEG-4 video SPL1 codec, the AMR speech codec, and the ITU-T H.223 Annex B multiplexing/demultiplexing at the same time. Three 16-bit multimedia-extended RISC processors, dedicated hardware accelerators, and a 16-Mb embedded DRAM are integrated on a 10.84 mm×10.84 mm die. It also integrates camera, display, audio, and network interfaces required for a mobile video-phone terminal. In addition to conventional low-power techniques, such as clock gating and parallel operation, some new low-power techniques are also employed. These include an embedded DRAM with optimized configuration, a low-power motion estimator, and the adoption of the variable-threshold voltage CMOS (VT-CMOS). The MPEG-4 videophone LSI consumes 240 mW at 60 MHz, which is only 22% of that for a conventional multichip design. Variable threshold voltage CMOS reduces standby leakage current to 26 μA, which is only 17% of that for the conventional CMOS design     

100-MHz serial access architecture for 4-Mb field memory

A 4-Mb field memory with a 100-MHz serial access rate has been developed. A new architecture that significantly improves serial I/O operation speed, reduces layout area, and offers simple control is proposed. To accomplish this task, a new architectural data shifter and high-speed redundancy circuit have been developed. The field memory has a 568-line×960 pixel×8-b (4,362,240 b) memory cell array designed for high-definition television (HDTV) screens. A 1.0 μm CMOS process technology is used to produce a die size of 12.94 mm×25.9 mm. The write-read cycle time is 9 ns, the access time is 8 ns, and the active current is 170 mA at a 50-MHz cycle rate with a standby current of about 3 mA     

A 5-V-only 16-Mb flash memory with sector erase mode

A 5-V-only 16-Mb CMOS flash memory with sector erase mode is described. An optimized memory cell with diffusion self-aligned drain structure and channel erase are keys to achieving 5-V-only operation. By adopting this erase method and row decoders to apply negative bias, 512-word sector erase can be realized. The auto chip erase time of 4 s has been achieved by adopting 64-b simultaneous operation and improved erase sequence. The cell size is 1.7 μm×2.0 μm and the chip size is 6.3 mm×18.5 mm using 0.6-μm double-layer metal triple-well CMOS technology     

A 1.8-V 128-Mb 125-MHz multilevel cell flash memory with flexible read while write

Application of multilevel cell (MLC) technology to a flexible read-while-write flash memory has been achieved through the use of a highly optimized sensing architecture. The goal of this implementation is to provide performance on par with single-bit-per-cell implementations while significantly reducing the overall die size. In order to achieve the required high-speed operation using MLC structures, all offsets to the sense amplifier were minimized and the column load and local sense amplifier were optimized to provide ample differential gain. Through the use of these optimization techniques, a 1.8-V MLC-based flexible read-while-write memory with 125-MHz continuous burst and 40-ns random read access time has been manufactured. Using a 0.13-/spl mu/m technology, this new device provides a die size that is 25% of the size of the equivalent single-bit-per-cell device manufactured on a 0.18-/spl mu/m technology.     

A 110-MHz/1-Mb synchronous TagRAM

A 4-way set associative TagRAM with 1.189-Mb capacity has been developed which can handle a secondary cache system of up to 16 Mbytes. A 9-ns cycle operation and clock to D_out of 4.7 ns are achieved by use of circuit techniques such as a pipelined decoding scheme, a single PMOS load BiCMOS main decoder, a BiCMOS sense-amplifying comparator, doubly placed self-timed write circuits, and highly linear VCO for a PLL. The device is successfully implemented with 0.7-μm double polysilicon double-metal BiCMOS technology     

A 100-MHz 4-Mb cache DRAM with fast copy-back scheme

A 4-Mb cache dynamic random access memory (CDRAM), which integrates 16-kb SRAM as a cache memory and 4-Mb DRAM into a monolithic circuit, is described. This CDRAM has a 100-MHz operating cache, newly proposed fast copy-back (FCB) scheme that realizes a three times faster miss access time over with the conventional copy-back method, and maximized mapping flexibility. The process technology is a quad-polysilicon double-metal 0.7-μm CMOS process, which is the same as used in a conventional 4-Mb DRAM. The chip size of 82.9 mm² is only a 7% increase over the conventional 4-Mb DRAM. The simulated system performance indicated better performance than a conventional cache system with eight times the cache capacity     

A 312-MHz 16-Mb random-cycle embedded DRAM macro with a power-down data retention mode for mobile applications

An embedded DRAM macro with a self-adjustable timing control (STC) scheme, a negative edge transmission scheme (NET), and a power-down data retention (PDDR) mode is developed. A 13.98-mm/sup 2/ 16-Mb embedded DRAM macro is fabricated in 0.13 /spl mu/m logic-based embedded DRAM process. Co-salicide word lines and MIM capacitors are used for high-speed array operation. The delay timing variation of 36 % for an RC delay can be reduced to 3.8% by using the STC scheme. The NET scheme transfers array control signals to local array blocks with high accuracy. Thereby, the test chip achieves 1.2-V 312-MHz random cycle operation even in the low-power process. 73-/spl mu/W data retention power is realized by using the PDDR mode, which is 5% of conventional schemes.     

A 150-MHz graphics rendering processor with 256-Mb embedded DRAM

A 150-MHz graphics rendering processor with an integrated 256-Mb embedded DRAM, delivering a rendering rate of 75 M polygons/s, is presented, 287.5 M transistors are integrated on a 21.3×21.7 mm ² die in a 0.18-μm embedded DRAM CMOS process with six layers of metal. Design methodologies for hierarchical electrical and physical design of this very large-scale IC, including power distribution, fully hierarchical timing design, and verification utilizing a newly developed nonlinear model, clock design, propagation delay, and crosstalk noise management in multi-millimeter RC transmission lines, are presented     

A 3.84 GIPS integrated memory array processor with 64 processingelements and a 2-Mb SRAM

An integrated memory array processor (IMAP) ULSI with 64 processing elements and a 2-Mb SRAM has been developed for image processing. The chip attains a 3.84 GIPS peak performance through the use of SIMD parallel processing and a 1.28 GByte/s on-chip processor-memory bandwidth. The IMAP is capable of parallel indirect addressing, which increases applications for parallel algorithms. Large power consumption with the wide memory bandwidth is avoided by reducing the number of active sense amplifiers and adopting dynamic power control. Fabricated with a 0.55-μm BiCMOS double layer metal process technology, the IMAP contains 11 million transistors in a 15.1×15.6 mm² die area     

A 90-nm FPGA I/O buffer design with 1.6-Gb/s data rate for source-synchronous system and 300-MHz clock rate for external memory interface

As FPGAs integrate into high-speed systems, performance and signal integrity become more important in I/O design. This paper describes the development of an FPGA design to support 1.6 Gb/s differential source-synchronous standards and 300 MHz external memory interfaces. Speed and performance were achieved using circuits such as differential level-shifters with voltage and temperature compensated current sources, on-chip decoupling capacitors, and floating-well output buffers. Programmable drive strength, output impedance matching, hot-socketing compliance, and 3.3-V voltage tolerance are features of the I/O buffer. In addition, DLLs and programmable phase-offset circuits were used to obtain precise timing control. The chip was manufactured on a 90-nm CMOS process.     

A 550-ps access 900-MHz 1-Mb ECL-CMOS SRAM

An ultrahigh-speed 1-Mb emitter-coupled logic (ECL)-CMOS SRAM with 550-ps clock-access time, 900-MHz operating frequency, and 12-μm² memory cells has been developed using 0.2-μm BiCMOS technology. Three key techniques for achieving the ultrahigh speed are a BiCMOS word decoder/driver with an nMOS level-shift circuit, a sense amplifier with a voltage-clamp circuit, and a BiCMOS write circuit with a variable-impedance bitline load. The proposed word decoder/driver and sense amplifier can reduce the delay times of the circuits to 54% and 53% of those of conventional circuits. The BiCMOS write circuit can reduce the power dissipation of the circuit by 74% without sacrificing writing speed. These techniques are especially useful for realizing ultrahigh-spaced high-density SRAMs, which will be used as cache and control memories in mainframe computers     

Cosmic ray soft error rates of 16-Mb DRAM memory chips

Manufactured 16-Mb DRAM memory chips use three different cell technologies for bit storage: stacked capacitors, trenches with internal charge, and trenches with external charge. We have measured the soft fail probability of 26 different 16-Mb chips produced by nine vendors to evaluate whether the different cell technologies have an impact on the chip soft error rate. This testing involved irradiation with neutrons, protons, and pions, the principle hadrons of terrestrial cosmic rays. The results show clear differences in soft-fail sensitivity, which appears to be related to the cell structure     

A 300-MHz 4-Mb wave-pipeline CMOS SRAM using a multiphase PLL

A 4-Mb (64 k×64) synchronous wave-pipeline CMOS SRAM is fabricated by 0.25-μm CMOS technology. Multiphase active pulse control (MPAC) enables fully random 300 MHz operation at 2.5 V, resulting in a bandwidth of 2.4 GB/s. The pulse is generated by multiphase PLL (MPPLL) using an array oscillator with current consumption of 7.5 mA     

A 900-MHz RF front-end with integrated discrete-time filtering

Discrete-time analog filters, rather than off-chip components, have been used to perform frequency selection and down conversion in the integrated front-end for a 900-MHz RF receiver. The first stage of frequency down conversion is implemented with a subsampling switched-capacitor sample-and-hold circuit clocked at 78 MHz. Subsequent stages of discrete-time filtering are realized using switched-capacitor biquadratic filters. An experimental prototype of the front-end had been integrated in a 0.6-μm BiCMOS technology. The circuit provides a system gain of 36 dB and 32 dB suppression of interfering channels over a 40 MHz bandwidth. Referred to the system input, the third-order intercept-point is -16 dBm, and the spot input-referred noise is -82 dBm over a 30 kHz bandwidth. The experimental circuit dissipates 90 mW from a 3.3-V supply and occupies an active area of 1.9×1.9 mm²      

A 1.8-ns access, 550-MHz, 4.5-Mb CMOS SRAM

An ultrahigh-speed 4.5-Mb CMOS SRAM with 1.8-ns clock-access time, 1.8-ns cycle time, and 9.84-μm² memory cells has been developed using 0.25-μm CMOS technology. Three key circuit techniques for achieving this high speed are a decoder using source-coupled-logic (SCL) circuits combined with reset circuits, a sense amplifier with nMOS source followers, and a sense-amplifier activation-pulse generator that uses a duplicate memory-cell array. The proposed decoder can reduce the delay time between the address input and the word-line signal of the 4.5-Mb SRAM to 68% of that of an SRAM with conventional circuits. The sense amplifier with nMOS source followers can reduce not only the delay time of the sense amplifier but also the power dissipation. In the SRAM, the sense-amplifier activation pulse must be input into the sense amplifier after the signal from the memory cell is input into the sense amplifier. A large timing margin required between these signals results in a large access time in the conventional SRAM. The sense-amplifier activation pulse generator that uses a duplicate memory-cell array can reduce the required timing margin to less than half of the conventional margin. These three techniques are especially useful for realizing ultrahigh-speed SRAM's, which will be used as on-chip or off-chip cache memories in processor systems     

A 2.5-ns clock access, 250-MHz, 256-Mb SDRAM with synchronousmirror delay

A 256-Mb SDRAM (245.7 mm²) has been developed using (1) a high cell occupation ratio (60.2%) array design for chip size reduction and a high yield, (2) a prefetched pipeline scheme (PPS) using a first-in first-out (FIFO) buffer with parallel serial converter for 250-MHz clock frequency operation, and (3) a synchronous mirror delay (SMD) circuit for 2.5-ns clock access and low standby current     

A 12.5-ns 16-Mb CMOS SRAM with common-centroid-geometry-layoutsense amplifiers

A 16-Mb CMOS SRAM using 0.4-μm CMOS technology has been developed. This SRAM features common-centroid-geometry (CCG) layout sense amplifiers which shorten the access time by 2.4 ns. A flexible redundancy technique achieves high efficiency without any access penalty. A memory cell with stacked capacitors is fabricated for high soft-error immunity. A 16-Mb SRAM with a chip size of 215 mm² is fabricated and an address access time of 12.5 ns has been achieved     

A 16-ns 1-Mb CMOS EPROM

A 16-ns 1-Mb CMOS EPROM has been developed utilizing high-speed circuit technology and a double-metal process. In order to achieve the fast access time, a differential sensing scheme with address transition detection (ATD) is used. A double-word-line structure is used to reduce word-line delay. High noise immunity is obtained by a bit-line bias circuit and data-latch circuit. Sufficient threshold voltage shift (indispensable for fast access time) is guaranteed by a threshold monitoring program (TMP) scheme. The array is organized as 64 K×16 b, which is suitable for 32-b high-performance microprocessors. The active power is 425 mW, the programming time is 100 μs, and the chip size is 4.94×15.64 mm²