Abstract
This thesis focuses on reducing the latency and energy consumption associated with the start-up of crystal oscillators in (duty-cycled) low-power wireless systems.
A comparison of the state-of-the-art in crystal oscillator start-up shows that energy injection (EI) systems start up faster than negative resistance boosted (NRB) circuits, but do not improve in terms of energy consumption. The main challenge in Energy Injection is the generation of a precise injection signal, which is required to ensure that the injection signal is in-phase with the crystal resonance over the entire injection duration.
This thesis proposes a self-timed injection technique that derives the injection waveform timing from the crystal oscillation itself, eliminating the need for an auxiliary injection oscillator. A prototype demonstrates a start-up time of only 6 µs and a start-up energy of 3.7 nJ for a 50 MHz crystal.
This research also addresses one of the main energy consumption issues in energy injection circuits by introducing step-wise charging, which reduces energy consumption by 2-4 times. The start-up time and energy are further reduced by combining this with further refinements to the self-timed injection technique. A second prototype starts up in only 2.8 µs, consuming only 1.9 nJ in the process when using the same crystal and using identical measurement conditions as for the first prototype.
This work also gives an overview of the design space and limitations in energy injection circuits, aiding in tailoring them to the application and giving insight toward a fair comparison of different techniques. The comparison of various circuits and methods highlights the importance of equal or similar measurement conditions for a fair evaluation.
Overall, this thesis aims to achieve fast start-up times, and minimize energy consumption in crystal oscillators. It introduces self-timed injection, as well the application of step-wise charging to crystal oscillators. Compared to other energy injection methods, the proposed techniques offer competitive start-up times while significantly reducing energy consumption.
A comparison of the state-of-the-art in crystal oscillator start-up shows that energy injection (EI) systems start up faster than negative resistance boosted (NRB) circuits, but do not improve in terms of energy consumption. The main challenge in Energy Injection is the generation of a precise injection signal, which is required to ensure that the injection signal is in-phase with the crystal resonance over the entire injection duration.
This thesis proposes a self-timed injection technique that derives the injection waveform timing from the crystal oscillation itself, eliminating the need for an auxiliary injection oscillator. A prototype demonstrates a start-up time of only 6 µs and a start-up energy of 3.7 nJ for a 50 MHz crystal.
This research also addresses one of the main energy consumption issues in energy injection circuits by introducing step-wise charging, which reduces energy consumption by 2-4 times. The start-up time and energy are further reduced by combining this with further refinements to the self-timed injection technique. A second prototype starts up in only 2.8 µs, consuming only 1.9 nJ in the process when using the same crystal and using identical measurement conditions as for the first prototype.
This work also gives an overview of the design space and limitations in energy injection circuits, aiding in tailoring them to the application and giving insight toward a fair comparison of different techniques. The comparison of various circuits and methods highlights the importance of equal or similar measurement conditions for a fair evaluation.
Overall, this thesis aims to achieve fast start-up times, and minimize energy consumption in crystal oscillators. It introduces self-timed injection, as well the application of step-wise charging to crystal oscillators. Compared to other energy injection methods, the proposed techniques offer competitive start-up times while significantly reducing energy consumption.
| Original language | English |
|---|---|
| Qualification | Doctor of Philosophy |
| Awarding Institution |
|
| Supervisors/Advisors |
|
| Award date | 2 Jun 2023 |
| Place of Publication | Enschede |
| Publisher | |
| Print ISBNs | 978-90-365-5550-0 |
| Electronic ISBNs | 978-90-365-5551-7 |
| DOIs | |
| Publication status | Published - 2 Jun 2023 |