UWA 5G base stations require precise frequency control to handle the rapidly growing data rates and volumes required by today’s users. These high demands drive the need for more precise (low-phase-noise) and tunable oscillators. In 2020, EQUS’ Translational Research Program awarded $55,000 to Chief Investigator Maxim Goryachev and PhD student Graeme Flower to implement EQUS research on precision time-keeping to improve the performance of 5G base stations. The 5G oscillator The team’s goal was to create a widely tunable low-phase-noise oscillator using ferrite-based frequency selection. Such a higher-order magnetic system is capable of narrower linewidths than would be possible with the current technology, while still being highly tunable. It also provides another degree of freedom for improving long-term stability. By equipping the oscillator with an interferometric noise feedback system, pioneered in the UWA laboratory, the team hoped to improve the phase noise even further.

Two resonators were investigated, with the main focus being measuring and minimising sources of noise and loss: a yttrium iron garnet (YIG) resonator and a lithium iron (LiFe) resonator. For the YIG resonator, the power to frequency (phase noise) conversion was shown to be small enough to not be a substantial problem. Temperature stabilisation was achieved by installing a new mount for the tunable magnet, active stabilisation was identified as probably the best option for improving magnetic field noise, and a numerical model was constructed to optimise resonator losses and couplings to a waveguide. For the LiFe resonator, the linewidth and tuning were characterised, with the turning point verified at room temperature, and initial tests of frequency noise (conversion from magnetic field to phase noise) were performed, which suggest the turning point may suppress noise.

Graeme Flower is now a Research Fellow and is continuing to work on this project. For the YIG resonator, the team will implement the active stabilisation (or use a low-noise electromagnet) to reduce magnetic field noise, proceed to oscillator feedback, and make the oscillator tunable while maintaining low phase noise. For the LiFe resonator, the team will retake the measurements with an improved readout set-up to better capture the weak signal, and investigate the strange loss and nonlinear behaviour. They are also aiming to produce a prototype oscillator, and look to working with a commercial partner to refine it for use in industry.

Published in the 2023 EQUS Annual Report