Hi, I am interested in sub-vth analog/rf design and I am wondering how to relate my hand calculations using EKV model to simulation results using BSIM4 models. Mainly I am interested in how much stock I should put into the 'region' parameter. I want to design for weak inversion, which Vittoz says is the same as the more common term, subthreshold. When I do a DC sweep of Vgs I find the region for which the device is in subthreshold (region 3), but when I apply a ramp input (Vpulse with long rise time) and plot the region parameter from transient result, the value of Vg for which the device is in region 3 is not the same as the result obtained in DC. Is this correct? Should I not consider this parameter and only focus on Inversion Coefficient? Also the spectre model guide states that in subthreshold the only device capacitance is Cgb, however I am seeing nonzero capacitances for all capacitances whose values are greater than Cgb by several orders of magnitude in some cases (self-capacitances on the order of pF). What is the reason for this?

let me recommend following readings:

• D. Stefanovic and M. Kayal 'Structured Analog CMOS Design'; Series: Analog Circuits and Signal Processing, 2009, ISBN: 978-1-4020-8572-7
• Paul G.A. Jespers; 'The gm/ID Design Methodology for CMOS Analog Low Power Integrated Circuits'; ISBN: 978-0-387-47100-6
• D. M. Binkley; 'Tradeoffs and Optimization in Analog CMOS Design', ISBN: 978-0-470-03136-0