A Human–AI Co-Design Methodology for Robust and Compact Miller CMOS OTAs Using the Half-Diamond Layout Style for MOSFETs
IEEE Access · 2026 · 14 · 96208-96228
DOI: 10.1109/access.2026.3705653 ↗At a glance
Journal article in IEEE Access, 2026. Written by Egon Henrique Salerno Galembeck and 3 co-authors. 35 references. The full text is freely available (open access, CC BY 4.0). Funded by: São Paulo Research Foundation, Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brazilian Federal Agency for Support and Evaluation of Graduate Education Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for the Article P….
Abstract (original language)
This work presents a systematic two-step design methodology for compact and robust Miller Complementary Metal-Oxide-Semiconductor (CMOS) Operational Transconductance Amplifiers (OTAs) targeting space, nuclear, medical, and military applications. In the first step, a customized Imperialist Competitive Algorithm (ICA), implemented within the iMTGSPICE electronic design automation tool, automates the multi-objective design and optimization of a Miller CMOS OTA using conventional Rectangular MOSFETs (RMs) under a 65 nm Bulk CMOS technology from TSMC, incorporating Corner and Monte Carlo robustness analyses within a single optimization loop. The proposed flow achieves a fully robust solution in approximately 17 minutes on a standard workstation, a result experimentally validated through fabrication and electrical characterization of the optimized RM-based OTA, which confirmed agreement with simulation across key figures of merit including open-loop voltage gain, unity-gain frequency, phase margin, and power consumption. In the second step, a layout-level area reduction strategy is introduced by systematically replacing the optimized RMs with Half-Diamond MOSFETs (HDMs), a second-generation unconventional layout style that exploits the Longitudinal Corner Effect (LCE) and the Parallel Connection of MOSFETs with Different Channel Lengths Effect (PAMDLE) to reduce the effective channel length and, consequently, the gate area, while preserving equivalent electrical performance and robustness under Corner and Monte Carlo analyses. A dedicated SPICE-compatible modeling methodology for HDMs is proposed and applied, enabling accurate circuit-level simulation of the HDM-based OTA. The HDM substitution achieves a 32% reduction in total gate area compared to the RM-based implementation, without requiring re-optimization of the circuit. Although the HDM-based OTA could not be fabricated using the official 65 nm TSMC PDK, the proposed methodology enabled its accurate evaluation through SPICE simulations. Besides, the HDM layout style is structurally expected to enhance tolerance to ionizing radiation through the Deactivation of Parasitic MOSFETs in the Bird’s Beak Regions Effect (DEPAMBBRE), a property supported by previously reported experimental results on HDM devices in the literature, although radiation tests were not conducted in the present study. The proposed methodology is topology-independent and directly extensible to other analog and RF CMOS IC designs. These results establish a practical and efficient path toward rapid development of compact, robust, and reliability-aware analog CMOS ICs for demanding applications.
Funding
- São Paulo Research Foundation
- Conselho Nacional de Desenvolvimento Científico e Tecnológico
- Brazilian Federal Agency for Support and Evaluation of Graduate Education Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for the Article P…
Cite
APA
Galembeck, E. H. S., Moreto, R. A. D. L., Junior, J. L. E., & Gimenez, S. P. (2026). A Human–AI Co-Design Methodology for Robust and Compact Miller CMOS OTAs Using the Half-Diamond Layout Style for MOSFETs. IEEE Access, 14, 96208-96228. https://doi.org/10.1109/access.2026.3705653
Vancouver
Galembeck EHS, Moreto RADL, Junior JLE, Gimenez SP. A Human–AI Co-Design Methodology for Robust and Compact Miller CMOS OTAs Using the Half-Diamond Layout Style for MOSFETs. IEEE Access. 2026;14:96208-96228. doi:10.1109/access.2026.3705653
BibTeX
@article{Galembeck2026,
author = {Galembeck, Egon Henrique Salerno and Moreto, Rodrigo Alves De Lima and Junior, Jair L. Emeri and Gimenez, Salvador Pinillos},
title = {A Human–AI Co-Design Methodology for Robust and Compact Miller CMOS OTAs Using the Half-Diamond Layout Style for MOSFETs},
journal = {IEEE Access},
year = {2026},
volume = {14},
pages = {96208--96228},
doi = {10.1109/access.2026.3705653}
}
EUVIMED