Error analysis of filters implemented with floating point arithmetic

Discrete time filters widely used in most control applications are usually implemented on finite precision digital devices which employ either fixed point of floating point arithmetic. Large dynamic range and the inherent normalizing characteristic of floating point representations may, in certain cases, make them preferable over other forms of representations. In a general simulation of system response, the effects due to quantization are not usually considered. This study presents a code for emulation of floating point quantization effects in time domain simulation of the system response