BSIM3v3 to EKV2.6 Model Parameter Extraction

BSIM3v3 to EKV2.6 Model Parameter Extraction and Optimisation

using LM Algorithm on 0.18um Technology node

Kirmender Singh and Piyush Jain

Int. Journal of Electronics and Telecommunications 2018 Vol.64 No.1 pp.5-11

Abstract: The industry standard BSIM3v3 and BSIM4.0 have been replaced by BSIM6.0 compact MOSFET model for deep submicron technology node. The BSIM6.0 is next generation, defacto industry standard model for bulk MOSFET. This model is charge based which is continuous from weak to strong inversion of operation. The core of analytical and physical BSIM6 model[3] is charge, with drain current equation expressed in form of source (qs) and drain charge (qd). This model has all its governing equations continuous and can be used to develop design methodology using IC based approach. But its method of computing qs and qd is complicated which is different from Vittoz traditional charge calculation method. The continuous interpolation equation of drain current as adopted by EKV2.6 although is empirical but its compact expression is preferred by analog designer to get intuitive design guidance. BSIM6 is a combined effort by BSIM and EKV modeling groups based on charge based continuous equations. Although EKV2.6 model is not valid for deep submicron process as it only includes submicron short channel effects like velocity saturation (VS), vertical field mobility reduction (VFMR), Drain induced barrier lowering (DIBL), channel length modulation (CLM) etc. But it still offers some benefits to have first cut design methodology because of its much simplified analytical equations. The inversion coefficient (IC) has found extensive acceptance in designer community as it offers enhanced design elegance in EKV then more complicated BSIM model. This paper discuses first step in analog design process by extracted core EKV2.6 intrinsic model parameters from industry standard BSIM3v3 model on 0.18µ technology node. The 0.18µ technology is chosen as it is still more common technology node in analog circuit design. The model parameters are extracted for different bins and optimisation is done using nonlinear optimisation LM algorithm. The optimised EKV2.6 parameters are validated with currentvoltage(I-V), intrinsic voltage gain (Avi) and Early voltage circuit parameter (VA) with BSIM3v3 model [read more...]

Flow-chart of BSIM to EKV conversion steps
(source:
D. Stefanovic and M. Kayal “Structured Analog CMOS Design" Springer Publications, 2008)