PAM4 SerDes TX Linear Fit

A Python framework for extracting channel pulse response from PAM4 SerDes waveforms using linear fitting (least-squares regression).

Live Tutorial & Documentation

What it does

Given a known PRBS pattern and a measured waveform (after passing through a lossy channel), this tool:

1. Builds a convolution matrix from the known symbols
2. Solves y = A * h via least-squares (SVD) to extract the channel pulse response
3. Computes the fitted waveform = conv(symbols, h) for verification
4. Derives the AC frequency response via FFT of the pulse response
5. Reports SNDR, RLM, eye metrics, and confidence intervals per tap

Quick Start

# Install dependencies
pip install numpy scipy matplotlib

# Launch GUI
python gui.py

Click Run Full Analysis to see all results across 8 tabs.

GUI Tabs

Tab	Shows
Waveform	Ideal / measured / fitted signals overlaid (1 sps)
Waveform (OS)	Oversampled waveforms (64 sps) showing transition shapes
Eye Diagram	PAM4 eye with CDR alignment, interpolated, white background
Pulse Response	Extracted vs ground-truth pulse with 95% confidence interval
Step Response	Cumulative sum of pulse (rise time, settling, DC gain)
AC Response	Frequency response: extracted vs true vs pure channel
Linear Fit	Matrix heatmap, SVD, residual, one-row verification
Histogram	PAM4 level distributions, RLM

Mouse Interaction (all plots)

Action	Effect
Scroll wheel	Zoom in/out centered on cursor
Right-click drag	Rectangle select to zoom
Double right-click	Reset to original scale

Project Structure

serdes_tx/          Core library
  prbs.py           PRBS9/PRBSQ9 generation
  pam4.py           PAM4 Gray coding, upsampling
  cdr.py            CDR + alignment
  pulse.py          Pulse extraction (convolution matrix + lstsq)
  metrics.py        SNDR, RLM, eye metrics, insertion loss
  channel.py        Freq-domain channel, AWGN, BW limit, jitter
  visualization.py  Eye diagram, plots
gui.py              Tkinter GUI (8 tabs)
demo.py             CLI demo script
docs/               GitHub Pages tutorial site
  index.html        English tutorial
  tutorial_zh.html  Chinese tutorial
  img/              Screenshots

Documentation

• English: docs/index.html
• Chinese: docs/tutorial_zh.html

The tutorial covers the complete mathematical derivation, explains every GUI panel, and includes FAQ, channel model comparison, and PRBS pattern guide.