Fully Integrated Parametric Filters for Extensive Phase-Noise Reduction in Low-Power RF Front-Ends and Resonant Sensing Platforms

Northeastern University, Boston MA

Investigators

Abstract

Numerous handheld medical devices and other consumer products rely on homodyne or super-heterodyne receivers for wireless connectivity. The maximum data rate achievable by such wireless nodes depends heavily on the frequency stability of the internal oscillators that generate the reference signals for the frequency conversion stages. Furthermore, significant research efforts have been directed towards the development of integrated closed-loop resonant sensing platforms, which rely on reliable oscillators to track the resonance frequency shifts induced in low-power micro and nano mechanical structures upon exposure to targeted physical or chemical signals. While such miniaturized resonant sensors have the potential to achieve unprecedented sensitivities, their detection capability is strongly limited by the stability of the oscillator employed as frequency readout. This research program aims to develop new techniques to achieve an unprecedented level of frequency stability in low-power and high-frequency integrated oscillators, addressing one of the most critical challenges that is currently limiting the performance of RF receivers and resonant sensing platforms. In particular, the proposed research entails the development of a new class of integrated solid-state low-power stabilization circuits, referred to as parametric filters. These integrated circuits exploit the complex nonlinear dynamics of parametric systems to implement the unique functionality of a filter for the phase noise reduction. By increasing the stability of frequency sources, the proposed parametric filter will allow to reduce the power consumption of battery-operated wireless sensor nodes deployed for Internet-of-Things (IoT) applications. In addition, the drastic phase noise reduction in the frequency readout of resonant sensing platforms will allow to surpass the resolution limits of state-of-the-art resonant sensors, leading to unprecedented detection capabilities. A parametric filter consists of a non-autonomous feedback network that includes a passive parametric frequency divider. When a parametric filter is designed to operate in proximity to a point of marginal stability for the parametric frequency divider, it exhibits an increase in relaxation time that renders it immune to the rapid phase fluctuations of its driving signal. As a result, the output signal of parametric filters exhibit orders of magnitude lower phase noise than their input signals. An aim of this project is to develop novel integrated parametric filters that can be connected at the output of gigahertz frequency generators. To demonstrate this concept, a 2.4 GHz frequency generator and the auxiliary circuits will be designed with an overall power consumption less than 0.6 mW. Phase noise improvements exceeding 30 dB at 1 MHz offset from the 2.4 GHz carrier are expected through the use of the new parametric filter architecture. Furthermore, a fundamental goal of this research is to develop a systematic design and simulation approach that allows to manipulate the stability of integrated passive parametric circuits. In particular, the efforts to develop parametric filters for low-power phase-noise reduction will provide means to understand the behavior of parametrically driven circuit components and their capability to manipulate the dynamics of non-autonomous feedback networks. In addition, the project will use commercial circuit simulators to capture such complex dynamics, which will greatly benefit circuit designers in the research community and industry. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.