Neuron MOS binary-logic integrated circuits. II. Simplifyingtechniques of circuit configuration and their practicalapplications

For pt.I see ibid., vol.40, no.3, p.570-6 (March 1993). The fundamental circuit ideas developed by the authors in Part I are applied to practical circuits, and the impact of neuron MOSFET on the implementation of binary-logic circuits is examined. For this purpose, two techniques are presented to simplify the circuit configurations. It is shown that the input-stage D/A converter circuit in the basic configuration can be eliminated without any major problems, resulting in improved noise margins and speed performance. Then a design technique for symmetric functions, which is especially important when the number of input variables increases, is presented. The νMOS logic design is characterized by a large reduction in the number of transistors as well as of interconnections. However, the decrease in transistor count comes at a cost in process tolerance due to the multivalued nature of the device operation. Test circuits were fabricated by a typical double-polysilicon CMOS process, and the measurement results are presented     

采用多输入浮栅MOS器件的四值编-译码电路设计

提出一种采用多输入浮栅MOS管设计具有可控阈值功能的电压型多值逻辑电路的方法.对每个浮栅MOS管的逻辑功能均采用传输开关运算予以表示以实现有效综合。在此基础上提出了一种新的电压型多输入浮栅MOS四值编码器和译码器设计。所提出的电路在结构上得到了非常明显的简化,并可采用标准的双层多晶硅CMOS工艺予以实现。此外,这些电路具有逻辑摆幅完整、延迟小等特点。采用TSMC0.35μm双层多晶硅CMOS工艺参数的HSPICE模拟结果验证了所提出设计方案的正确性。     

On the application of the Neuron MOS transistor principle formodern VLSI design

In this paper, the speed performance, power consumption, and layout area of Neuron MOS transistor circuits are monitored considering the requirements of modern VLSI design. The Neuron MOS transistor is a recently discovered device principle which has a number of input gates that couple capacitively to a floating gate. The floating gate potential controls the current of a transistor channel. This device can be used in logic circuits. A threshold current through the Neuron MOS transistor can be defined that causes a switching of the output of the logic circuits as soon as the channel current surmounts or falls below the specified value. We designed two different multiplier cells, one based on a Neuron MOS inverter, and the other on a Neuron MOS n-MOSFET which is used as one input device of a comparator circuit. Functionality of both cells is proven for data rates up to 50 MHz which represents the first high-speed measurement of a circuit based on this new design principle. A perspective for the upper speed limit found at more than 500 MHz is given by simulation. The new design principle has a layout area reduced by more than a factor of two compared to usual multiplier cells. Moreover, it is shown, that depending on the design chosen, high speed operation leads to considerable power savings. In view of those advantages it is concluded that the principle of threshold logic qualifies for a major breakthrough for packing density improvement of CMOS-based applications     

Computer-aided design and characterization of digital MOS integrated circuits

A computer-aided circuit-simulation method is developed to enable the design, characterization, and optimization of MOS integrated circuits. The computation of dc and transient characteristics is done in terms of physical device parameters extracted from processing information and incorporated in an analytical device model. It is demonstrated that any MOS circuit configuration (with its associated series resistances and parasitic devices) can be analyzed in terms of an equivalent inverter. Input-output transfer characteristics are obtained by superposition of the load and transistor I-V characteristics, providing the necessary information for dc > `worst-case' design. A simple device model was used to compute circuit transient response. All the computed characteristics are in good agreement with measurements performed on integrated circuits.     

基于SET/MOS混合结构的译码器电路的设计

基于SET的I-V特性以及SET与MOS管互补的特性,以MOS管的逻辑电路为设计思想,首先提出了一个SET/MOS混合结构的反相器,进而推出或非门电路,并最终实现了一个唯一地址译码器.通过SET和MOS管两者的混合构建的电路与纯SET实现的电路相比,电路的带负载能力增强;与纯MOS晶体管实现的电路相比,电路同样仅需要单电源供电,且元器件数目得到了减少,电路的静态功耗大大降低.仿真结果验证了电路设计的正确性.     

基于开关信号理论的三值电流型CMOS电路设计

本文应用开关信号理论对电流型CMOS电路中MOS传输开关管与电流信号之间的相互作用进行了分析,并提出了适用于电流型CMOS电路的传输电流开关理论。应用该理论设计的三值全加器等电路具有简单的电路结构和正确的逻辑功能,从而证明了该理论在指导电流型CMOS电路在开关级逻辑设计中的有效性。     

Methodology for the design of a 4-bit soft-hardware-logic circuit based on neuron MOS transistors

As soft-hardware-logic circuits had been proposed in the literature as an alternative for digital circuits taking advantage the fact that any Boolean function could be implemented with the same cell, just configuring external signals, this work shows a methodology that could be followed particularly for the design of a four bits logic gate, using the so-called neuron MOS transistor (ν-MOS). Simulation results show the feasibility of the design for performing as XNOR, NOR, OR, XOR, AND or NAND logic gates, for instance. In order to extrapolate the design to a higher number of bits, the key issue is to properly consider the weight of the input capacitances in correlation with the number of input bits. A D/A converter can be used as the input stage of the configuration. This design considers the D/A converter-less version, since it helps to increase device integration as the number of transistors used is reduced with no difference in its performance. The design should be based on the theoretical floating potential diagram (FPD) of the desired logic gate.     

A neuron-MOS neural network using self-learning-compatible synapsecircuits

A circuit technology for self-learning neural network hardware has been developed using a high-functionality device called Neuron MOS Transistor (υMOS) as a key circuit element. A υMOS can perform weighted summation of multiple input signals and thresholding all at a single transistor level based on the charge sharing among multiple capacitors. An electronic synapse cell has been constructed with six transistors by merging a floating-gate EEPROM memory cell into a new-concept υMOS differential-source-follower circuitry. The synapse can represent both positive (excitatory) and negative (inhibitory) weights under single V_DD power supply and is free from standby power dissipation. An excellent linearity in the weight updating characteristics of the synapse memory has been also established by employing a simple self-feedback regime in each cell circuitry, thus making it fully compatible to the on-chip self-learning architecture of υMOS neural networks. The basic operation of the synapse cell and a υMOS neural network using the synapse has been experimentally verified using test circuits fabricated by a double-polysilicon CMOS process     

DESIGN OF TERNARY CURRENT-MODE CMOS CIRCUITS BASED ON SWITCH-SIGNAL THEORY

By applying switch-signal theory, the interaction between MOS transmission switch-ing transistor and current signal in current-mode CMOS circuits is analyzed, and the theory oftransmission current-switches which is suitable to current-mode CMOS circuits is proposed. Thecircuits, such as ternary full-adder etc., designed by using this theory have simpler circuit struc-tures and correct logic functions. It is confirmed that this theory is efficient in guiding the logicdesign of current-mode CMOS circuits at switch level.     

In the blink of a silicon eye

In this article, a new device structure called the neuron-bipolar junction transistor (νBJT) was presented. It has been successfully applied to the design of a silicon retina and large-neighborhood cellular neural networks (CNNs). The νBJT-based smoothing array for the silicon retina has a simple and compact structure, which is suitable for the VLSI implementation. It can be integrated with other CMOS retinal signal processing circuits to form smart sensor systems     

Implementation of Multi-Valued Logic Gates Using Full Current-Mode CMOS Circuits

In this paper, a novel multi-valued logic gate set is designed by using only current-mode CMOS circuits. The gate set consists of min, max, inverter, literal, and cyclic operators based on a current-mode, versatile, novel threshold topology. They are shown to exhibit better static and dynamic behavior and consume less area compared to previous MVL design topologies and binary-logic counterparts. The gate circuits are investigated in terms of analog design aspects, such as mismatch and noise. The proposed topology is compared to previous topologies in terms of attainable radix and DC characteristics. A radix-8 multiplex function and a radix-8 full-adder circuit is designed to demonstrate the advantages of new current-mode multi-valued logic circuits.     

Symbolic fault modelling of MOS combinational circuits

This paper presents a new method for fault modelling of MOS combinational circuits at the transistor level. Every transistor is replaced with a conductance controlled by the gate logic value. The specific advantage of the method is use of a symbolic simulator for circuit function extraction. This function is referred as Transistor Logic Conductance Function (TLCF). Starting from a known TLCF, a simple set of rules is used for output state determination. The method is suitable for multiple fault model generation thanks to the fact that only one symbolic analysis of a circuit is sufficient for modelling different stuck-open, stuck-short and stuck-at faults of a logic gate. Moreover, the method can deal also with bridging and cut faults. Finally, the application of the TLCF for test pattern generation is considered.     

Computer analysis of the double-diffused MOS transistor for integrated circuits

The effective length of an MOS transistor can be made narrow by using double diffusion similar to a bipolar transistor. Computations were conducted for an n-channel double-diffused transistor with different surface concentrations, channel lengths, channel gradients, surface-states densities, and substrate concentrations. A shorter channel length and a higher surface-state density, e.g.langle1, 1, 1ranglecrystal, gave a higher drain current and transconductance. The maximum transconductance in many cases occurs at low gate voltages. The computations indicate that a gain-bandwidth product in the gigahertz range can be expected when the graded channel region is less than 1 µm. The difference between an n-type substrate and a p-type substrate is not substantial. The analysis is also useful in predicting the performance of any integrated logic circuit using the diffused enhancement transistor as the active switch and a depletion-mode transistor (without a diffused channel) as the load device. The computation indicates that satisfactory performance can be obtained using a load device with the same geometry and an ON voltage of only a fraction of a volt, This revelation indicates that double-diffused channel MOS transistors not only give higher speed but also smaller chip area for integrated circuits and a lower supply voltage (hence less power dissipation).     

Novel dynamic merged load technology

Two new device concepts for dynamic ratioless inverter logic circuits are presented. Very high circuit density is achieved by replacing the traditional MOS dynamic load transistor with a novel load element which is merged with the switching transistor. Both device types can be implemented with a relatively standard double polysilicon CMOS process and are ideally suited for very low-power digital signal processors, serial memories and correlators, and digital image processors.     

Design optimization of JCMOS structures

JCMOS structures are based on merging an MOS capacitance, a JFET, and a bipolar transistor in an area of a single MOS transistor. The structure performs the basic operations of temporary storage, writing, and sensing of the stored data. It is used in DRAM, serial dynamic memory, and dynamic logic applications. In addition to the advantages of small size and high speed of operation, the use of the JCMOS structure to implement dynamic logic gates overcomes the problem of charge redistribution associated with conventional and domino CMOS logic circuits. In this paper, the JCMOS structure implementation using a retrograde p-well CMOS process is presented. An analytical model relating terminal voltages and currents to device dimensions and doping levels is derived. Simulation results are presented for both reading and writing modes of operation. A test cell was successfully fabricated to verify the principle of operation, and experimental and theoretical results are compared. A simplified lumped component equivalent circuit, to be used in circuit simulators such as SPICE, is presented, and its validity is investigated. The structure design requirements and procedure are presented. The model is used to optimize the design of the structure.     

The Combinational and Sequential Adiabatic Circuit Design and Its Applications

In this paper, a new design of adiabatic circuit, called the quasi-static efficient charge recovery logic (QSECRL) is proposed. To achieve minimum energy consumption, this paper proposes a technique to reduce channel resistance and remove diodes from the signal path. This design method can be implemented in both combination logic and sequential logic. The counter circuit and the 8-bit carry look-ahead (CLA) circuit, a more complex circuit, are selected to evaluate this proposed design. All simulations in this paper have been implemented by SPICE with the 0.8 μm MOSIS technology MOS transistor model under 2-volt (peak-peak) sinusoidal power-clock supply. The results show significantly improved performance of the 8-bit CLA circuit with 20–30 fJ and 70 fJ energy consumption at 10–100 MHz and 500 MHz operating frequency, respectively.     

A unified switching theory with applications to VLSI design

Classical switching theory is shown to have deficiencies when applied to the analysis and design of MOS VLSI circuits. A new logic design methodology called CSA theory is described here which overcomes many of these deficiencies. It is based on three primitive component types: connectors that perform wired-logic operations, switches representing controlled connectors, and attenuators representing resistive load devices. Four basic types of logic values are recognized: Boolean 0 and 1 values, unknown or indeterminate U values, and the high-impedance state Z. The number of logic values can be increased systematically to improve modeling accuracy using a concept of logical strength, which corresponds to current drive capability in analog circuits. It is shown that both the behavior and layout of most types of MOS logic circuits, including contact, gate, and nonclassical mixed circuits, can be treated in a uniform and rigorous manner using CSA network models with either four or seven logic values. The use of a digital charge-storage element called a well to represent sequential behavior is examined. CSA theory is applied to two VLSI design issues, inverter synthesis and fault simulation.     

An integrated wide-band variable-gain amplifier with maximum dynamic range

The design of an integrated wide-band variable-gain amplifier with maximum dynamic range is approached by considering three basic bipolar transistor configurations from which all others can be derived. The analysis of noise and of distortion shows the importance of transistor base resistance in all three circuits. On the basis of these analyses, one configuration is shown to yield maximum dynamic range, and this configuration is then used as the basis for the development of a new circuit called the improved automatic-gain control (agc) amplifier. A unique biasing scheme allows a considerable reduction in distortion and noise, together with a significant increase in bandwidth compared with conventional circuits.     

Multiple-input logic circuit design using BiCMOS-based negative differential resistance circuits

We first propose an inverter circuit design using the negative differential resistance (NDR) circuit composed of the standard Si-based n-channel metal-oxide-semiconductor field-effect-transistor (NMOS) and SiGe-based heterojunction bipolar transistor (HBT). By suitably designing the MOS width/length parameters, we can obtain the ??-type NDR current?Cvoltage (I?CV) characteristic. Expanding the inverter circuit operation, the two-input and four-input NOR logic gates are demonstrated. Especially, the design and fabrication of the logic circuit is based on the standard SiGe BiCMOS process. Compared to the traditional NDR device like resonant tunneling diode (RTD), our MOS?CHBT?CNDR-based applications are much easier to be combined with some Si-based or SiGe-based devices on the same chip.     

RF Power Performance of Asymmetric-LDD MOS Transistor for RF-CMOS SOC Design

This letter presents a new asymmetric-lightly-doped-drain (LDD) metal oxide semiconductor (MOS) transistor that is fully embedded in a CMOS logic without any process modification. The radio frequency (RF) power performance of both conventional and asymmetric MOS transistor is measured and compared. The output power can be improved by 38% at peak power-added efficiency (PAE). The PAE is also improved by 16% at 10-dBm output power and 2.4 GHz. These significant improvements of RF power performance by this new MOS transistor make the RF-CMOS system-on-chip design a step further