SPICE simulation support for me to make gate and drain sweeps between my different PDKs.

This repository contains device abstraction layers and characterization scripts for nmos/pmos devices. The repo has examples for running ngspice simulations with the SkyWater 130nm PDK, but I also use HSPICE and commerical processes.

• ngspice
• Python 3 with numpy and matplotlib
• SkyWater 130nm PDK models (can be installed with the installskywater.sh script)

Set the SPICE_LIB environment variable to point to your PDK models:

export SPICE_LIB=/path/to/skywater-pdk/libraries/sky130_fd_pr/latest/models

The sky130/devices.cir file provides abstract device definitions that allow swapping between different PDKs without modifying your circuit files:

Usage in your SPICE netlist:

.include devices.cir
Xn1 drain gate source body abnmos
Xp1 drain gate source body abpmos

Sweeps gate voltage from 0 to 1.8V for NMOS and PMOS at two drain voltages (saturation and linear regions).

ngspice -b gatesweep.cir
python plotgatesweep.py

Sweeps drain voltage from 0 to 1.8V at multiple gate voltages for both NMOS and PMOS.

ngspice -b drainsweep.cir
python plotdrainsweep.py

spicesupport/
├── README.md
├── devices.cir          # Top-level abstraction (gitignored, copied from sky130/)
├── gatesweep.cir        # Gate sweep simulation
├── plotgatesweep.py     # Gate sweep plotting script
├── gatesweep.png        # Gate sweep output
├── drainsweep.cir       # Drain sweep simulation
├── plotdrainsweep.py    # Drain sweep plotting script
├── drainsweep.png       # Drain sweep output
└── sky130/
    └── devices.cir      # SKY130 device definitions

1. Create a new directory under spicesupport/ (e.g., gf180/)
2. Add a devices.cir file mapping abnmos and abpmos to the appropriate devices
3. Copy the desired devices.cir to the top level before running simulations

MIT