MOSFET Electrical Simulation Dataset (MESD)

Citation Author(s):
Yuhang
Zhang
Kain Lu
Low
Longfan
Li
Leilai
Shao
Yanan
Sun
Jian
Zhao
Yabin
Sun
Yanling
Shi
Yongfu
Li
Submitted by:
Yuhang Zhang
Last updated:
Tue, 10/15/2024 - 07:35
DOI:
10.21227/wa5a-r991
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

This paper introduces the Metal-Oxide-Semiconductor Field Effect Transistor (MOSFET) Electrical Simulation Dataset, MESD, an extensive collection of I-V and C-V characteristics data simulated across different foundries' Berkeley Short-channel IGFET Models (BSIMs). The MESD dataset covers a range of bias voltages, temperatures, and MOSFET physical dimensions across several technology nodes from 3 to 350 nm. MESD is designed to serve as a standardized benchmark for evaluating and comparing MOSFET models, aiding semiconductor researchers and model developers in assessing the accuracy of their simulations against a robust set of controlled data.

To get more details of the dataset, please refer to the paper: Y. Zhang, K. lu Low, L. Li, L. Shao, Y. Sun, J. Zhao, and Y. Li, Descriptor: MOSFET Electrical Simulation Dataset (MESD), IEEE Data Descriptions.

Instructions: 

Data File Naming Convention

The dataset is in MESD folder. The data is recorded with JSON files. Each JSON file corresponds to a device in a PDK. The files are named as PDK Name-Device Name. MESD includes data from 11 PDKs and 44 types of devices from 350 nm to 3 nm. The PDKs and devices are renamed to avoid leakage of foundry information.

Data Structure

Each file includes metadata such as PDK name, device name, simulator, and model. The Record section is a list, in which each entry includes the Vgs and the corresponding Ids and Cgg under different dimensions (i.e., WL, and Nfin), simulation conditions (i.e., Temp and Corner), and voltage bias (Vds).

Data Simulation Script

The scripts folder provides examples for the simulation of MOSFET devices. netlist.dc.spectre and netlist.ac.spectre show the simulation script for Ids and Cgg collection using Cadence Spectre. Similarly, netlist.dc.sp and netlist.ac.sp are simulation scripts using Synopsys HSPICE.

License

This repository is licensed under the terms of the Creative Commons Attribution 4.0 International License.