The current state of development of the television broadcasting system of the future is described with regard to LSI development. It is no need to say that television broadcasting systems are very huge and require a large number of inexpensive LSI's. Hi-Vision broadcasting has already been started in Japan. In the United States, a digital terrestrial broadcasting system (ATV) will be standardized in the near future. On the other hand, the situation in Europe remains unclear but MPEG-2 is now in the stage of system finarizing. We also hear much about "multimedia" but the concept of multimedia broadcasting still requires a lot of time to be translated into reality. Some important current technical topics and related basic technologies are also described in this paper. They include DCT, Hybrid DCT coding, error correcting coding, coded modulation, and improvement of the MUSE system. Finally, the discussion considers the relationship between system development and VLSI technology and the importance of mutual understanding between VLSI engineers and system designers. Some possible requirements for VLSI development are also stated.

Development of low power MUSE (Multiple Sub-Nyquist Sampling Encoding) chip set through reduction in operating voltage (from 5 V to 3.7 V) is described. This leads to great cost reduction since the chips could be mounted on low cost plastic packages and the necessity for cooling fans to dissipate heat was obviated. To maintain compatibility with standard 5 V analog and digital peripherals such as 4 Mbit DRAMs and an A/D converter, a special voltage-level converter was also developed.

A new approach to implement queues for controlling ATM switch LSI is presented. In many conventional architecture, external FIFOs are provided for each output link and used to manage the address of the buffer in an ATM switch. We reduce the number of FIFOs by using a self-timed queue with a search circuit that finds the earliest entry for each output link. Using this architecture, number of the FIFOs is reduced to 1/N, where N is the switch size. Delay priority and multicasting can be supported without doubling the number of the queues. This new queue can also be utilized as an ATM switch by itself. Evaluation chip was fabricated using 0.5-µm CMOS process technology. Inter-stage transfer speed over 500 MHz and cycle time over 125 MHz was obtained. This performance is enough for a 622-Mbps 1616 ATM Switch.

This paper presents a fully digital high speed (60 Mb/s) Quadrature Phase Shift Keying (QPSK)/Offset QPSK (OQPSK) burst demodulator for radio applications, which has been implemented on a 0.5 µm Complementary Metal Oxide Semiconductor (CMOS) master slice Very Large Scale Integrated circuit (VLSI). The developed demodulator VLSI eliminates analog devices such as mixers, phase-shifters and Voltage Controlled Oscillator (VCO) for bit-timing recovery, which are used by conventional high-speed burst demodulators. In addition to the fully digital implementation, the VLSI achieves fast carrier and bit-timing acquisition in burst modes by employing a reverse-modulation carrier recovery scheme with a wave-forming filter for OQPSK operation, and a bit-timing recovery scheme with bit-timing estimation and interpolation using a pulse-shaping filter. Results of performance evaluation assuming application in Time Division Multiple Access (TDMA) systems show that the developed VLSI achieves excellent bit-error-rate and carrier-slipping-rate performance at high speed (60 Mb/s) with short preamble words (less than 100 symbols) in low E_b/N_o environments.

A single-chip architecture for three-dimensional (3-D) computer graphics (CG) is discussed assuming portable equipment with a 3-D CG interface. Based on a discussion of chip requirements, an architecture utilizing DRAM technology is proposed. A 31-Mbit, on-chip DRAM cell array allows a full-color, 480640-pixel frame with two 3-D frame buffers for double buffering and one 2-D frame buffer for superimposed or background images. The on-chip pixel generator produces R, G, B, and Z data in a triangular polygon with a zigzag-scan interpolation algorithm. The on-chip frame synthesizer combines data from one of the 3-D buffers with that from the 2-D buffer to produce superimposed or background 2-D images within a 3-D CG image. Parallel alpha-blending and Z-comparison circuits attached to the DRAM cell array provide a high data I/O rate. Estimation of the chip performance assuming the 0.35-µm CMOS design rule shows the chip size, the drawing speed, on-chip data I/O rate, and power dissipation would be 1413.5-mm, 0.25 million polygons/s, 1 gigabyte/s, and 590 mW at a voltage of 3.3 V, respectively. Based on circuit simulations, the chip can run on a 1.5-V dry cell with a drawing speed of 0.125 million polygons/s and a power dissipation of 61 mW. A scaled-down version of the chip which has an 1-kbit DRAM cell array with an attached alpha-blending circuit is being fabricated for evaluation.

An Ultra-high-speed (higher than 60 MHz) Viterbi decoder VLSIC with coding rates from one-half to fifteen-sixteenth and a constraint length of seven for forward error correction (FEC) has been developed using 0.8-µm semicustom CMOS LSIC technology and a newly developed high-speed ACS circuit. To reduce power consumption of the one-chip Viterbi decoder, a universal-coding-rate scarce-state-transition (SST) Viterbi decoding scheme and low-power-consumption burst-mode-selection (BMS) path memory have been proposed and employed to the developed VLSIC. In addition, a new maximum-likelihood-decision (MLD) circuit for the SST Viterbi decoder has been developed. The total power consumption of the developed chip is reduced to 75% of the conventional one and the developed Viterbi decodar VLSIC achieves a maximum operation speed of 60 MHz. It achieves near theoretical net coding-gain performance for various coding rates.

Mixed-signal LSIs promise to permit increased levels of integration, not only in voiceband but also in multi-GHz-band applications such as wireless communications and optical data links. This paper reviews the evolution of mixed-signal communications LSIs and discusses some of their design problems, including device noise and crosstalk noise. In the low-power and low-voltage designs emerging as new disciplines, the target supply voltage for voiceband LSIs is around 1 V, and even GHz-band circuits are approaching 2 V. MOS devices are expected to play an important role even in the frequency range over 100 MHz, in the area of wireless or optical communications circuits.

This paper describes the circuit design and experimental results of a video-rate 10-b analog-to-digital converter (ADC) suitable for consumer video products, such as high-definition TV sets. Triple-stage conversion scheme combined with two new conversion methods, "Dynamic Sliding Reference Method" and "Triangular Interpolation Method," and an internal Bi-CMOS Sample/Hold circuit have been developed. These conversion methods require no adjustment circuit to fit reference voltages between conversion stages and realize small active area. As a result, a maximum conversion frequency of 16 MHz, acceptable SNRs of 56 dB and 48 dB for 10 kHz and 8 MHz input frequency respectively and small DNLE of 0.75 LSB have been achieved. This ADC is fabricated with 1.2 µm Bi-CMOS technology and integrates very small number of bipolar transistors of 2 K on a small active area of 2.52.7 mm² and consumes 350 mW.

A multi-stage noise-shaping (MASH) A/D converter combining an RC-integrator and a digital correction technique for high accuracy is described. Using 1.2-µm BiCMOS technology, we developed an A/D converter for digital audio with an S/N ratio of over 100 dB. This paper discusses the principles of MASH technology with an RC-integrator, the technique for correcting RC variation, and the experimental results obtained with a fabricated chip.

There are two approaches to implementing the international standard video coding algorithms such as H.261 and MPEG: a programmable DSP approach and a building block approach. The advantages and disadvantages of each are discussed here in detail, and the video coding algorithms and required throughput are also summarized. For more complex standard such as MPEG-, VLSI architecuture became more sophisticated. The DSP approach incorporates special processing engines and the building block approach integrates general-purpose microprocessors. Both approaches are capable of MPEG- NTSC coding in a single chip. Reduction of power consumption is a key issue for video LSIs. Architectures and circuits that reduce the supply voltage while maintaining throughput are summarized. A 0.25-µm, 3-GOPS, 0.5-W, SIMD-VSP for portable MPEG- systems could be made by using architecture-driven voltage scaling as well as feature-size scaling and SOI devices.

The hybrid coding with motion compensated prediction and discrete cosine transform (MC+DCT) has been recognized as the standard technique in motion picture coding. In this paper, a motion estimation processor compatible with ITU-T H.261 and MPEG standards is described. A half-pel precision processing unit is introduced with an exhaustive block matching unit for integer-pel precision search. The necessary processing power for the exhaustive block matching is implemented with a 1-dimensional array structure utilizing a sub-sampling technique. In comparison with the conventional 2-dimensional array structure, path of the data transfer is so simple that the low power dissipation characteristic is obtained. The problem of communication bandwidth to the frame memory, which is a bottleneck of half-pel precision motion estimation, is solved by introducing a candidate pixel buffer into the inter-processor data transfer. A static latch circuit with conflict free operation is newly developed for reducing the power consumption. This chip is capable of processing NTSC-resolution video in real-time at the 40 MHz operation. The chip integrates about 540 k transistors in the 121 mm² die using 0.8 µm double metal CMOS technology.

Processing elements (PEs) with a dynamically reconfigurable pipeline architecture allow the high-speed calculation of widely used neural model which is multi-layer perceptrons with the backpropagation (BP) learning rule. Its architecture that was proposed for a single chip is extended to multiprocessors' structure. Each PE holds an element of the synaptic weight matrix and the input vector. Multi-local buses, a swapping mechanism of the weight matrix and the input vector, and transfer commands between processor elements allow the implementation of neural networks larger than the physical PE array. Estimated peak performance by the measurement of single processor element is 21.2 MCPS in the evaluation phase and 8.0 MCUPS during the learning phase at a clock frequency of 50 MHz. In the model, multi-layer perceptrons with 768 neurons and 131072 synapses are trained by a BP learning rule. It corresponds to 1357 MCPS and 512 MCUPS with 64 processor elements and 32 neurons in each PE.

High performance I/O circuits for fast memory devices such as Synchronous DRAMs were studied. For a TTL interface, the effect of capacitive loading must increase as I/O speed is increased, and signal termination is required for frequencies over 100 MHz. For this reason, industry-proposed alternative interface approaches such as GTL and CTT were investigated using experimental test devices. The results showed that open-drain type drivers have a problem; as the frequency increases, the high-level output voltage becomes degraded. In contrast, a push-pull driver T-LVTTL (Terminated Low Voltage TTL), developed as an implementation of the CTT interface specification, was found to be suitable for high-speed data transfer. A high-speed bus driver circuit connecting an impedance element in series to the stub is proposed as an application of T-LVTTL. Simulated results showed that this scheme greatly improves the signal integrity of memory bus systems; the operating frequency could very well be the highest among several schemes discussed as candidates for the post-LVTTL standard interface.

A programmable clock generator, based on a phase-locked loop (PLL) circuit, has been developed with 0.5 µm CMOS triple-layer Al interconnection technology for use as an on-chip clock generator in a 300-MHz video signal processor. The PLL-based clock generator generates a clock signal whose frequency ranges from 50 to 350 MHz which is an integral multiple, from 2 to 16, of an external clock frequency. In order to achieve stable operation within this wide range, a voltage controlled oscillator (VCO) with selectable low VCO gain characteristics has been developed. Experimental results show that the clock generator generates a 297-MHz clock with a 27-MHz external clock, with jitter of 180 ps and power dissipation of 120 mW at 3.3-V power supply, and it can also oscillate up to 348 MHz with a 31.7-MHz external clock.

A 3.2 GHz, 50 mW, 1 V, GaAs clock pulse generator (CG) based on a phase-locked loop (PLL) circuit has been designed for use as an on-chip clock generator in future high speed processor LSIs. 0.5 µm GaAs MESFET and DCFL circuit technologies have been used for the CG, which consists of 224 MESFETs. An "enhanced charge-up current" inverter has been specially designed for a low power and high speed voltage controlled oscillator (VCO). In this new inverter, a voltage controlled dMESFET is combined in parallel with the load dMESFET of a conventional DCFL inverter. This voltage controlled dMESFET produces an additional charge-up current resulting in the new VCO obtaining a much higher oscillation frequency than that of a ring oscillator produced with a conventional inverter. With a single 1 V power supply (V_dd), SPICE calculation results showed that the VCO tuning range was 2.25 GHz to 3.65 GHz and that the average VCO gain was approximately 1.4 GHz/V in the range of a control voltage (V_c) from 0 to 1 V. Simulation also indicated that at a V_dd of 1 V the CG locked on a 50 MHz external clock and generated a 3.2 GHz internal clock (=50 MHz64). The jitter and power dissipation of the CG at 3.2 GHz oscillation and a V_dd of 1 V were less than 8.75 psec and 50 mW, respectively. The typical lock range was 2.90 GHz to 3.59 GHz which corresponded to a pull-in range of 45.3 MHz to 56.2 MHz.

A novel effective channel length extraction method has been developed, which utilizes the difference between the local threshold voltage of channel region and that of external region. In this method, the dependence of external resistance on V_g is taken into account, and it is not necessary to extract V_th. It is found that the external resistance can be approximated as the linear function of V_g with V_g around V_th. For a 0.4 µm gate length LDD MOSFET, the accuracy and resolution are estimated to be less than 0.02 µm and 0.003 µm, respectively.

The statistical properties of insertion losses and return losses for optical connectors are investigated theoretically using the probability theory and the Monte Carlo simulation. Our investigation is focused on an orientation method for reducing insertion loss by which a fiber-core center is adjusted in a region of within a certain angle to the positioning key direction. It is demonstrated that the method can significantly improve insertion losses, and that an adjusting operation angle of 90 degrees is sufficient to realize an insertion loss of less than 0.5 dB with 99% cumulative probability. Good agreement was obtained between the theoretical distribution and the experimental results for single-mode fiber connection. Consequently, it is indicated that the statistical distributions of insertion losses and return losses of optical connectors in the field can be predicted theoretically from the values measured in the factory by connection to a master connector.

Electromagnetic plane wave scattering by a loaded trough on a ground plane has been analyzed by Kobayashi and Nomura's method. The field in each region is expressed first in terms of appropriate eigen functions, whose excitation coefficients are determined by the continuity condition across the aperture of the trough. Simple far field expression which is suitable for numerical calculation for small aperture cases has been derived. Scattering far field patterns and radar cross section are calculated and compared with those obtained by other methods. Good agreements have been observed for all incident angles.

A beam adaptive frame for finite-element beam-propagation analysis is proposed. The width of the frame can be adapted itself to either the guiding structure or the propagating beam in optical circuits, so the size of the computational window can be reduced.