Indian Journal of Science and Technology

Abstract

Background:Distributed Arithmetic (DA)technique is a bit serial operation used to perform vector to vector multiplication in convolution, which is an essential operation in digital filters. This paper presents realization of Finite Impulse Response (FIR)filtersusingsixdifferentDAtechniquesnamelyconventionalDA,ModifiedDA(MDA),OffsetBinaryCodingDA(OBC-DA), Offset Binary Coding Modified DA (OBC-MDA) and LUT-Less DA. Methods: A novel DA method is also introduced which is a combination of multiplexers and Look-Up-Tables (LUT). In DA based FIR filter architecture the partial products of the FIR filter coefficients are pre-calculated and stored in LUTs. The filtering operation is done by shift-accumulation unit. Findings: The designs are simulated using Xilinx ISE and synthesized in Cadence RC using 180-nm technology library. From synthesis reports, it is found that for 4-tap FIR filter, Modified DA consumes 30% less power and 40% less area, whereas for 8-tap FIR filter the saving in power consumption is 30-80% and 35-80% in area. Conclusion: Compared to all the DA techniques mentioned, MDA requires least area and least power consumption because of its less memory requirement. All architectures are designed for 4-tap and 8-tap FIR filters.
Keywords: Distributed Arithmetic, FIR Filter, LUT-Less DA, Modified DA, MUX-LUT DA, OBC-MDA, Offset-Binary Coding DA