Abstract:In this paper, we address the problem of compact model parameter extraction to simultaneously extract tens of parameters via derivative-free optimization. Traditionally, parameter extraction is performed manually by dividing the complete set of parameters into smaller subsets, each targeting different operational regions of the device, a process that can take several days or even weeks. Our approach streamlines this process by employing derivative-free optimization to identify a good parameter set that best fits the compact model without performing an exhaustive number of simulations. We further enhance the optimization process to address critical issues in device modeling by carefully choosing a loss function that evaluates model performance consistently across varying magnitudes by focusing on relative errors (as opposed to absolute errors), prioritizing accuracy in key operational regions of the device above a certain threshold, and reducing sensitivity to outliers. Furthermore, we utilize the concept of train-test split to assess the model fit and avoid overfitting. This is done by fitting 80% of the data and testing the model efficacy with the remaining 20%. We demonstrate the effectiveness of our methodology by successfully modeling two semiconductor devices: a diamond Schottky diode and a GaN-on-SiC HEMT, with the latter involving the ASM-HEMT DC model, which requires simultaneously extracting 35 model parameters to fit the model to the measured data. These examples demonstrate the effectiveness of our approach and showcase the practical benefits of derivative-free optimization in device modeling.

Scalable Compact Modeling for On-Chip Passive Elements with Correlated Parameter Extraction and Adaptive Boundary Compression

A Numerical Framework to Simplify CAD Models for Reliable Estimates of Physical Quantities

A Scalable Model of On-chip Inductor Including Tunable Dummy Metal Density Factor

A Derived Physically Expressive Circuit Model For Integrated Passives In Rfic

A Compact Circuit Model of On-Chip Multi-layer Spiral Inductors on Silicon Substrate

Quasi-Static Derived Physically Expressive Circuit Model for Lossy Integrated RF Passives

Passive-Assured Rational Function Approach For Compact Modeling Of On-Chip Passive Components

Modeling and Characterization of On‐Chip Interconnects

A Physics-Aware Methodology for Equivalent Circuit Model Extraction of TSV-inductors

An automatic parameter extraction and scalable modeling method for transformers in RF circuit

Scalable Compact Circuit Model and Synthesis for RF CMOS Spiral Inductors

Modeling of On-Chip Rf Passive Components

Efficient Surrogate Modeling and Design Optimization of Compact Integrated On-Chip Inductors Based on Multi-Fidelity EM Simulation Models

An Efficient Optimization Method of RF Passive Components Using RBF Model

A Layout Optimization Method to Reduce Commutation Inductance of Multi-Chip Power Module Based on Genetic Algorithm

Compact Model Parameter Extraction via Derivative-Free Optimization

Scalable Modeling of Layout Parameters in Cmos Integrated Stacked Millimeter Wave Transformer

An Accurate Compact Model for On-Chip Vertically Coiled Transformers

Scalable Compact Circuit Model for Differential Spiral Transformers in CMOS RFICs

A New Highly-Scalable Equivalent Circuit Model for On-Chip Symmetrical Transformer with Accurate Substrate Modeling