A novel architecture design for VLSI implementation of an FIR decimation filter

A novel architecture design of a one stage FIR filter for decimation is described. It performs the decimation of a 1-bit code at 1024KHz of double integration Sigma Delta modulation output to PCM at 16KHz. This architecture is designed in such a way that it needs only a simple control structure suitable for VLSI implementation. We devised an algorithm for generating the coefficients of the filter with a minimum of required hardware. It does not require storing the coefficients in a ROM and continuously reading it to calculate the convolution. The accumulators needed to perform the direct convolution are arranged in a way that simplifies and minimizes the hardware required for the filter implementation. The filter response is Sinc³(f) which provides sufficient attenuation for modulation generated by means of double integration. The implementation of this filter requires the generation of the coefficients and the performance of the convolution. Three coefficients are needed with every input to obtain the output sequence. The major feature of this architecture is the use of an efficient algorithm to obtain the coefficients thereby reducing the area and power consumption. It is very suitable for VLSI implementation in CMOS technology.