CMOS computational circuits with improved performances

摘要

The paper presents an original approach of designing analog signal processing circuits, based on the re-using of the same functional core for implementing two circuit functions: signal gain with theoretical null distortions and signal squaring. The advantages of the increased modularity and controllability and of the reduced design costs represent an immediate consequence of the double function realized by the proposed structures. Because the most important circuit complexity is concentrated for implementing the core of the structure, both circuit area and power consumption per each realized function can be strongly reduced using this method. The overall error of the transconductance amplifier is 0.4% and the approximation error for the squaring circuit is 0.27%, in the condition of a low-voltage low-power operation (a supply voltage of 1.5V and a medium current consumption of 50μA for each implemented circuit function).