Design of a Neural Network-Based VCO With High Linearity and Wide Tuning Range

Abstract

In this paper, a 2 GHz LC-VCO with neural network (Multilayer Perceptron) has been designed in a 0.13 ţm CMOS technology. With the integrated neural network, the linearity and tuning range of the LC-VCO has been substantially improved. Compared to a conventional VCO design, the proposed technique can improve the linearity by selecting optimized bias voltages obtained from the output of the neuron network. The result shows that the tuning nonlinearity of the proposed VCO is further optimized from 0.335% to 0.254%.

Analog and Mixed-signal Integrated Circuits of 5Gand 6G Data Communication

“…BVCEO describes the maximum voltage that a transistor can withstand across its collector-emitter junctions before breakdown occurs, a function of its doping concentrationtherefore an important parameter to describe its power handling capabilities. fT and fmax (typically, fmax is higher than fT ) can be verified in harmonic distortion analyses and power and noise figure measurements [34]. Silicon has dominated microelectronic circuits since 1965 and as a result, has also been extensively researched ever since.…”

“…SiGe active components, particularly the HBT, are a highly feasible candidate for high frequency (mm-wave and THz included), high bandwidth, and high data rate wireless communications. The alloy, formed by adding germanium to silicon during the material processing phase, increases the carrier mobility, decreases noise during operation, and increases the cutoff frequency of the transistor, as has been presented in [7], [16], [34], [62], [63], [64], and [65]. The bandgap and electron mobility of Si1−xGex depends on ranging between pure silicon (x = 0) and pure germanium (x the material composition and doping level of the material, = 1).…”

“…Learning how conversion gain, noise figure, linearity, and isolation are impacted in topologies and under operating conditions, while constantly learning from and improving the datasets can lead to improved performance, as reported by [16], [34], and [35]. As with feedback and control systems, AI-assisted mixers also require large and accurate datasets that may be difficult and time-consuming to generate.…”

Design of a constant loop bandwidth phase-locked loop based on artificial neural network

“…According to Eq. (1), a suitable set of biasing voltages can improve the linearity of varactor.Ref [23]. applied a neural network to optimize the biasing voltages.…”

“…applied a neural network to optimize the biasing voltages. Compared with result in[23], more parameters like VCO gain, divider ratio and phase scaling should be considered to optimize loop bandwidth, according to Eqs.…”