Transconductance related analysis of EKV MOSFET model for a 0.35 µm CMOS technology node

Transconductance g_m is an important parameter for small-signal analysis of analog circuits. It is paramount to analyse the g_m and transconductance efficiency (g_m/I_D) as it provide useful design guidlines. The sensitivity of g_m and (g_m/I_D) on process related parameters gives deeper insight of circuit performance. Amongst compact MOSFET models, EKV is known for its suitability in low power analog design due to its physical basis. Due to ever increasing complexity of other compact models, EKV is considered ideal as educational an analytical CMOS analog design tool. Notwitstanding ultimate role of BSIM for benchmark validation in the final stage. The paper outlines some useful classroom outcome on EKV experimentation. Firstly, EKV model parameters have been mapped from BSIM model file of 0.35 μm technology node to obtain DL, COX, GAMMA, PHI, KP etc. Next, the I-V curve used as reference data is generated using BSIM3v3, and EKV2.6 model equations are curve-fitted to obtain the best fit value of the EKV2.6 parameters.An optimization program is run for getting global fit value in all operating ranges and conditions, by finding the curve-fitting error. The sensitivity of drain current I_D, g_m, and (g_m/I_D) to process and model parameters is determined for strong and weak inversion conditions. The (g_m/I_D) dependency on inversion coefficient (IC) for different channel length has been shown for EKV2.6 model. The lumped transconductance parameters get resolved into various components providing deeper insight of the relationship of the transconductance parameter with specific physical phenomenon. The method illustrated for long channel can be extended for short channel and the dependence of all component on inversion level can be obtained.