A novel design methodology, enabling extreme design cycle time speed up of DC - DC power converters, is developed. The concept is based on providing high accuracy post-layout RC parasitics aware results, by replacing complicated large RC netlists with small signal approximation models. Scattering parameters analysis is adopted for “on the fly” performance simulation of the power MOSFET switches' routings, which act as large passive linear networks. The RC parasitics aware back end of line (BEOL) S-parameter model is extracted and seamlessly integrated into the schematic testbench, considering the actual circuit as a black box and therefore actively cutting down the design's netlist size to minimum values. Thus, the DC – DC converter performance degradation, that previously could not be simulated, now is accurately predicted and evaluated while the respective simulation time and the number of design iterations needed from layout (physical design) to the schematic and vice versa, are minimized. The proposed methodology is validated using an Integrated Pulse Width Modulation controlled DC - DC converter product vehicle, for light energy harvesting applications, designed, simulated and fabricated in a 0.18 μm CMOS standard process. Experimental results confirm the accuracy and design cycle speed up effectiveness of the proposed novel IC power converter design methodology.

Machine learning (ML) algorithms are utilized for the implementation of a ML python-based model of dynamic behavior of Power DC – DC converters, for energy harvesting applications. This subfield of artificial intelligence, defined as the capability of a machine to imitate intelligent human behavior, is used to perform complex tasks in a way that is similar to how a designer is implementing nonlinear switching circuits, like a power DC-DC converter. The dynamic behavior of this nonlinear vehicle is simulated with ML, and in particular all the related dynamic characteristics, obtained with large signal time domain simulation, such as dynamic voltage drop – electro migration and time domain operation, are now captured rapidly with a ML approach. The related results are benchmarked versus transistor level simulations, depicting superior accuracy in minimum simulation time.