README.md

January 26, 2026 · View on GitHub

Tests PyPI - Downloads DOI

pygmid

A python3 implementation of the gm/ID starter kit
Report bug · Request feature

Table of contents

About

pygmid is a Python 3 version of the gm/ID starter kit by Prof. Boris Murmann of Stanford University. The package also offers some scripts from the Paul Jesper's book.

If you find this package useful for your research, please consider citing it.

The original MATLAB gm/ID starter kit can be found here

Installation

To install pygmid from source, download from Github and run pip:

pip install .

in the root directory.

pygmid can also be installed from PyPI:

pip install pygmid

Usage

Scripting with the Lookup Class

A gm/ID lookup object can be generated with the Lookup class. The lookup object requires lookup data for initialisation. Both .mat files generated using MATLAB or .pkl files generated using pygmid's own characterisation script are supported.

You can create a lookup object as follows:

from pygmid import Lookup as lk

NCH = lk('180nch.mat')

Access MOS Data

The Lookup class allows for pseudo array access of the MOS matrix data. You can access data as follows:

# get VGS data as array from NCH
VGS_array = NCH['VGS']

Data is returned as a deep copy of the array contained in the Lookup object.

Lookup functionality

Lookup of interpolated data occurs as follows:

VDSs = NCH['VDS'] 
VGSs = np.arange(0.4, 0.6, 0.05)
# Plot ID versus VDS
ID = NCH.look_up('ID', vds=VDSs, vgs=VGSs)
# alias function lookup can also be used
ID = NCH.lookup('ID', vds=VDSs, vgs=VGSs)
# check bias
VGS = NCH.look_upVGS(GM_ID = 10, VDS = 0.6, VSB = 0.1, L = 0.18)
print(f'VGS is: {VGS}')

plt.plot(VDSs, np.transpose(ID))

Modes 1 (Simple parameter lookup), mode 2 (arbitrary ratio lookup) and mode 3 (cross lookup of ratios) are implemented. The companion lookupVGS function is also included.

Technology Extraction Functions

The EKV extraction function can be used as follows:

from pygmid import EKV_param_extraction, XTRACT

(VDS, n, VT, JS, d1n, d1VT, d1logJS, d2n, d2VT, d2logJS)\
        = EKV_param_extraction(NCH, 1, L = 0.18, VDS = 0.6, VSB = 0.0)

Sample usage of this and other utility functions can be found in debug_utility.py .

Examples

Usage of lookup scripts are given in debug_lookup.py and debug_lookupVGS.py.

Sample outputs for the 180nm generic pdk used in the Murmann starter pack are given below:

image

image

image

image

image

image

The generic PDK data is hosted here.

Sweeping a Technology

pygmid also features a CLI which can be used to run techweeps to generate transistor data.

Documentation will be added in due course. This functionality is in a state of flux.

Citation

If you find this package useful in your research, please consider citing the following papers:

  1. M. Srivastava*, C. O’Donnell*, B. Griffin, P. Cantillon-Murphy, and D. O’Hare
    (*joint first authors),
    “Efficient Bio-Sensing Amplifier Design: A Python-Based gm/ID Design Methodology,”
    Proceedings of the IEEE Biomedical Circuits and Systems Conference (BioCAS), 2024.
    DOI: 10.1109/BioCAS61083.2024.10798363

  2. T. Cortez, C. O’Donnell, and D. O’Hare,
    “Low Power ADC Buffer: A Python-Based gm/ID Design Methodology,”
    Proceedings of the 32nd IEEE International Conference on Electronics, Circuits and Systems (ICECS),
    Marrakech, Morocco, 2025, pp. 1–4.
    DOI: 10.1109/ICECS66544.2025.11270766

Authors

Contributors

  • José Rui Custódio

A special thanks to Prof. Boris Murmann for giving permission to use his work and release this package under the Apache 2.0 License.