Phase Noise Simulation for Large Time-Constant Oscillators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing phase-noise analysis methods, such as the shooting Pnoise method, fail to accurately simulate phase-noise characteristics for oscillators with large time-constant voltage supplies, as they assume frequency-independent eigen-modes and cannot capture noise contributions from these sources, leading to incorrect results or failure in extracting dominant modes.
Innovation Solution
A modified perturbation projection vector method is introduced, which augments the shooting Newton matrix with a phase-shift factor and a pinning equation to separate the matrix into phase and amplitude modulation equations, allowing for the use of the Arnoldi-vector recycle technique to speed up matrix solving, thereby simulating phase noise contributions from large time-constant sources accurately.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the shooting Pnoise method is used for phase noise analysis, then the method works well for some circuits, but it fails to extract a unique dominant mode and gives wrong phase-noise results for oscillators with large time-constant voltage supplies
Solution Approach 1:
The patent changes the fundamental parameter assumption from frequency-independent eigen-modes to frequency-dependent eigen-modes. This allows the analysis method to adapt to oscillators with large time-constant voltage supplies, where the eigen-modes actually vary with frequency. By modifying this key parameter assumption, the method becomes reliable for both traditional and large time-constant oscillator circuits.
2Device complexity
If the shooting Pnoise method assumes frequency-independent eigen-modes, then the calculation is simplified, but it cannot capture noise contribution from voltage supply and misses the hump at small offset frequency
Solution Approach 1:
The patent introduces dynamics into the eigen-mode representation by making them frequency-dependent rather than static. This dynamic approach captures the varying behavior of the oscillator circuit across different frequencies, enabling accurate capture of noise contributions from the voltage supply and the characteristic hump at small offset frequencies, while still maintaining a systematic calculation framework.
3Measurement precision
If a modified perturbation projection vector method is used with matrix augmentation, then phase and amplitude modulation can be separated, but the matrix size increases requiring Schur-complement approach
Solution Approach 1:
The patent applies segmentation by dividing the augmented matrix into phase modulation and amplitude modulation components using the Schur-complement approach. This segmentation allows each sub-problem to be solved separately and more efficiently, reducing the overall computational complexity despite the initial matrix augmentation. The separation also provides clearer physical interpretation of the results.
Data Source
AI summary
Embodiments may include methods, systems, and computer-readable storage mediums that may be used to simulate phase noise for an oscillator circuit. In some embodiments, a method of simulating phase noise for an oscillator circuit may include providing an oscillator circuit description. A time-domain representation of a small signal phase noise of the oscillator circuit description may be determined. A shooting Newton matrix representation of the time-domain representation of the small signal phase noise may be generated. The shooting Newton matrix representation may be augmented to include a phase-shift factor and a pinning equation. The augmented shooting Newton matrix representation may be solved to determine a signal output of the oscillator circuit.