Research On The Key Technology Of Low Phase Noise Quartz Crystal Oscillator

Crystal oscillator is a common frequency source in electronic products at present,and it is widely used in various fields including communication equipment,automatic driving equipment,navigation system and so on.With the advent of the digital age,the demand for various electronic devices continues to increase,and the requirements for the accuracy and stability of clock signals are also getting higher and higher.Crystal oscillator with low phase noise is one of the current research hotspots.This paper studies the characteristics of crystal oscillators.The principle of the quartz crystal plate with piezoelectric effect in the crystal oscillator as a resonator is analyzed,and the temperature-frequency characteristics of different cut crystal resonators are studied.In this paper,the oscillation principle of the crystal oscillator is analyzed through the feedback model analysis method and the negative resistance analysis method,and the start-up and stable conditions of the crystal oscillator are obtained.After comparing three different crystal oscillators,this paper chooses the Pierce oscillator as the oscillation form of the high-frequency crystal oscillator circuit.Starting from the concept of phase noise,this paper deduces the application of the Leeson phase noise model in crystal oscillators,calculates the load factor of the oscillator,obtains the relationship between the load factor Q_L and the load capacitance,and introduces the oscillator temperature compensation technology and the role of this technique in optimizing phase noise.This paper proposes a crystal oscillator system using BP neural network temperature compensation,including a control signal module,a temperature compensation voltage generation module and a crystal oscillator module,which control the learning and training state and working state of the neural network through the control signal module.And the artificial neural network model was established based on MATLAB,and its compensation ability was simulated and verified.The verification results show that the neural network temperature compensation has high frequency stability,and the frequency deviation is less than 0.24ppm.Finally,in order to verify the feasibility of the low phase noise design method,a483.2MHz high-frequency crystal oscillator is designed in this project,and the simulation is carried out using Cadence Virtuoso software.The simulation is carried out under the0.18μm BCD process,and the results show the validity of the low phase noise design method.The design was taped and tested under the process.The actual chip works at a power supply voltage of 3.3V,the operating temperature range is-40℃～85℃,the power consumption is 18m A,and the phase noise performance:128.18dBc/Hz@10kHz,143.24dBc/Hz@100kHz,better than the frequency source of the same frequency.