Autonomous maximum power point tracking algorithm for ultra-low power energy harvesting

An ultra-low power autonomous MPPT algorithm that maximizes the efficiency of a monolithic 0.98 mm² solar harvester is presented. Using only the pn-junctions of the standard 130 nm single n-well process, the monolithic harvester can serve as supply for wireless sensor grains. Based on the perturbation and observation method, the MPPT algorithm maximizes the output current of the integrated charge pump. The proposed approach inherently optimizes the system efficiency by automatically considering parasitic losses and source characteristics while requiring less than 100 nA supply current. This paper includes a theory section describing the maximum power flow from the illuminated pn-junction to an on-chip charge pump capacitor. Based on these results the analog algorithm is determined and described in detail introducing a resistorless high side current sensor for 100 nA to 1 mA. Concluding with measurement results regarding the tracking efficiency, a maximum deviation from the simulated loaded optimum input voltage of 3% is shown.