No abstract

In this paper, we proposed to use 3D integration technology to create a neuromorphic hardware system that is compatible with current technology, provides high system speed, high density, massively parallel processing, low power consumption, and small footprint. The Through Silicon Vias (TSVs) used in the 3D neuromorphic structure provide high density integration and energy efficient links for transferring information through multiple neuron layers. This work details how a 3D neuromorphic system is benefited from the redundant TSV with substantial design-area reduction. We discussed the yield and reliability issues and explained the impact in neuromorphic 3D system design. A spiking neuron model is developed for the proposed 3D system. Furthermore, a new methodology have been proposed by introducing oxide around the bump that could significantly enhance the TSV capacitance in 3D Neuromorphic Computing (NC) system.

This paper gives a comprehensive study on the modeling and design challenges of Through Silicon Vias (TSVs) in high speed three-dimensional (3D) system integration. To investigate the propagation characteristics incurred by operations within the ultra-broad band frequency range, we propose an equivalent circuit model which accounts for rough sidewall effect and high frequency effect. A closed-form expression for TSV metal oxide semiconductor (MOS) capacitance in both depletion and accumulation regions is proposed. The coupling of TSV arrays and near and far field effect on crosstalk analysis are performed using 3D ElectroMagnetic (EM) field solver. Based on the TSV circuit model, we optimize the TSVs' architecture and manufacturing process parameters and develop effective design guidelines for TSVs, which could be used to resolve the signal integrity issues arising at high frequency data transmission in 3D Integrated Circuits (ICs).

Abstract-An equivalent circuit model of a finite ground plane coplanar waveguide (FGCPW) interconnect in a metal-insulator-semiconductor (MIS) system for an ultra-broadband monolithic IC is proposed and illustrated. An effective substrate considering Maxwell-Wagner Polarization is suggested and demonstrated. The method of modeling the weak skin effect of the conductor is presented. The accuracy of the equivalent circuit model is evaluated. This proposed FGCPW interconnect equivalent circuit model enables a quick and efficient time domain simulation to estimate the time delay and bandwidth of ultra-broadband ICs.

This report outlines National Institute of Standards and Technology (NIST)'s validation tools developed to help address the smart grid's emergent needs and complement research and standards coordination efforts. Rapid advances have enabled an opportunity for the grid to autonomously and simultaneously control generation, balance load, and regulate power quality. The grid will need this coordinated control to regulate power and voltage in realtime, to meet fast-changing power demands, especially with increasing distributed energy resources. To do this, the grid needs measurements from across the distribution network to accurately assess demand and prevent negative impact of variable grid-edge generation. Additionally, such a coordinated control system depends on a new operating paradigm with two-way information flows between domains of the grid. The different operational domains and the information flows between them are outlined in the NIST Framework and Roadmap of Smart Grid Interoperability Standards, Release 4.0 [1].The coordinated control mentioned above requires improved observability of the state of distributed resources and depends on marshalling increasingly decentralized control systems. To implement such a system and meet system level performance objectives, distribution system operators (DSOs) need an understanding of the bandwidth, security and integrity considerations while also ensuring reliability at the grid's edges. Industry is also concerned with how to represent uncertainties like latency and cyber-vulnerabilities in a complex cyber-physical system. Additionally, industry is concerned that standards are focused more on devices than on coordinated control systems.Thus, this report presents an overview of a data acquistion and control system designed by NIST as a validation tool and deployed on the NIST campus in Gaithersburg, MD. This system is comprised of interoperating computing resources, cybersecurity infrastructure, communication links and data interfaces to both take measurements and to transmit control inputs to commonly used distribution system control components. An important proof of concept included in the aforementioned validation system is a data pipeline optimized to stream data from a 5 MW PV inverter and multiple sources of synchronized phasor data. This pipeline utilizes industry standard information modeling and communication standards with the intent of being compatible with industry standard testing equipment. This capability demonstrates how DSOs could test current standards against functionality -such as phasor based control of distribution circuits.