Optimized RF CMOS Low Noise Amplifier Design Via Geometric Programming

We present an innovative geometric programming (GP)-based approach for optimizing the design of a CMOS low noise amplifier (LNA). An LNA with common source inductive degeneration is formed in the GP framework for optimization. Globally optimal design parameters satisfying design constraints and specifications are obtained with high computational efficiency. Tradeoff curves for analyzing various design metrics, such as noise figure and quality factor, are rapidly computed through GP. Numerical examples show that the GP-optimized values are in excellent match with the known optimal values. Moreover, the GP framework allows fast computation and easy incorporation of additional design constraints that are generally difficult, if not impossible, in the traditional approach