Design and optimization of variable gain LNA for IoT applications using meta-heuristics search algorithms

Abstract

In this paper, Variable Gain LNA (VG-LNA) parameters are optimized using the Particle Swarm Optimization (PSO), Firefly Algorithm (FA), and Genetic Algorithm (GA). A comparison of the three optimization techniques has been done and FA is depicting better results over GA and PSO. VG-LNA is composed of a Complementary Common Gate (CCG) and Variable Gain Amplifier (VGA). Gm-boost topology helps in increasing the gain while the current reuse technique provides less power consumption. Optimization algorithms simulated on MATLAB and the result shows minimum Noise Fig. (NF) is 2.62 dB, maximum gain is 17.8 dB, S11 i.e. input reflection coefficient is −13.5 dB and S22 i.e. output reflection coefficient is −14.7 dB at 50 Ω impedance matching while Figure of Merit1 (FoM1) is 36.14 dB using FA. The FA optimized parameters when simulated on Cadence Virtuoso software using GPDK 45 nm CMOS technology for the frequency range of 26–32 GHz then results show a minimum NF of 2.6 dB at 30.9 GHz, maximum gain of 16.9 dB at 30.5 GHz, S11 is −17.7 dB at 30.5 GHz, S22 is −21.2 dB at 29 GHz and FoM1 of 34.19 dB. The layout of the realized circuit has an area of 231.695 μm*164.48 μm i.e. 0.03811mm2.