Field Programmable Analog Array

Abstract

A new approach to develop Field Programmable Analog Arrays (FPAAs), which avoids excessive number of programming elements in the signal path to enhance the performance, is presented. The architecture of the FPAA may be considered as a worthy compromise between granularity and performance. High frequency operation, a simple programming methodology, operation over almost three decades of frequency, and use of standard digital CMOS fabrication process are the salient features of the FPAA. The proposed FPAA is based on simple current mode primitives, which may be configured to perform basic functions such as amplification, attenuation, inversion, summation and integration. These elementary operations may then be combined to obtain more complex signal processing, including filtering and data conversion. Emphasis of this work has been on selecting a circuit primitive with low inherent parasitics and on reducing the parasitics of the unavoidable programming elements. An uncompounded scheme has been used for the programming and configuration of the Configurable Analog Blocks (CABs). The programmability of the CAB affords flexibility to the FPAA, to make it a pragmatic solution for building prototypes for implementing analog signal processing functions.