Skip to main content

β: This English translation is in beta — the Traditional-Chinese original is the authoritative version.

ISF and Oscillator Phase Noise: From Paper to Design Intuition

From paper to design intuition. This site takes the Hajimiri–Lee Impulse Sensitivity Function (ISF) theory from first principles all the way to hands-on design feel for SerDes clock jitter. Every formula is derived step by step, with units, numerical examples, Python code and figures, and explicit paper citations.

Who this site is for​

We assume you are an EE graduate: you know circuit theory, electronics, signals and systems, random processes, the Fourier transform, and basic DSP, but you have not yet truly mastered oscillator phase noise, timing jitter (the deviation of a clock edge from its ideal instant), the ISF, cyclostationary noise, or the LTV (linear time-variant) oscillator model. After finishing this site, you should be able to:

  • read and re-derive on your own the core equations of [P1] Hajimiri–Lee 1998;
  • explain "why an oscillator is LTV — not LTI — with respect to noise";
  • compute phase noise from the ISF and convert phase noise into rms jitter;
  • name the design knobs that lower 1/f² and 1/f³ phase noise;
  • connect all of this to practical SerDes clocking intuition (LC-VCO / ring-VCO / PLL / CDR).

Start from your goal​

The full 9-step sequential path is in the Learning Path; but if you already have a concrete goal in mind, it is faster to jump in through one of the three "entry cards" below.

I want to learn ISF from scratch​

You are not in a hurry and want to understand oscillator phase noise (the random fluctuation of an oscillation signal's phase) from the ground up. First build the physical intuition of "what an oscillator's phase actually is", then follow the planned path.

I have a phase-noise plot and need jitter​

You already have a measured or spec-sheet L(Δf)\mathcal{L}(\Delta f) curve and want to know how to convert it into rms timing jitter, and what that means for a SerDes link.

I am reading a Hajimiri paper​

You are sitting in front of one of the papers and want to know "where on this site is this equation derived step by step".

Quick reference​

You do not want to read a whole page — you just need one symbol, one term, or one formula:

  • Cheat Sheet — signature formulas and numerical feel, packed into one page
  • Notation — site-wide consistent symbols, meanings, units
  • Glossary — intuitive explanations of the English terms
  • Equation Index — formula ↔ paper source ↔ derivation page

Required background​

Linear systems and convolution, Fourier series/transform, random processes and PSD (power spectral density), basic circuits (RLC, capacitor q=Cvq=Cv), and a little Python/NumPy. You do not need any prior knowledge of oscillator noise — that is exactly what this site teaches.

How to run the simulations​

# Install the site dependencies
npm install
# Start the local site (http://localhost:3000)
npm run start
# Re-run all simulations in one shot; regenerates every figure under static/figures/
python scripts/run_all_sims.py

Every figure is traceable to its script and formula in the figure_index.

How to read the equations​

Equations are rendered with KaTeX. Every important formula comes with: its [Px] Eq.(n) page source, its physical meaning, its units, a numerical example, and its conditions of validity and failure. A TODO: marker means that spot still needs manual verification against the original PDF. A "toy model" label means it is a pedagogical simplification, not transistor-level accuracy.

References​

The five source papers and external supplements are in references; per-paper close readings are in the paper deep dives; the paper map is in paper_summary_table.

Honest disclosure: the source folder contains 5 PDFs. 4 of them are Hajimiri-series oscillator phase noise / injection papers, but 1 (Hajimiri_ISCS_98.pdf) is actually a cross-coupled sense amplifier paper, unrelated to the ISF. This site labels it honestly and uses it only as a side note. See the build_report for details.