Energy harvesting is a technology that can be applied to IoT systems to eliminate the need for batteries. Many types of energy sources are available for energy harvesting, such as light, thermal, vibration, and electromagnetic energy. Indoors, where most IoT devices are located, artificial light, such as from LED lamps, can be used for energy harvesting in circuits with very ultra-low power consumption. Integrated switch-capacitor DC-DC converters are required for this type of system to convert the harvested energy into a constant output voltage suitable for powering an electronic circuit. The idea of this work is to use a hysteretic feedback control consisting of comparators and a logic system to adjust the switching frequency and the voltage conversion ratio (VCR) of the converter. With this, the equivalent output resistance is tuned to a value that results in a constant output voltage. A new method for modeling the equivalent output resistance based on charge flow analysis is proposed, which also considers the effects of source resistance. An integrated energy-harvesting system consisting of a switched-capacitor DC-DC converter is implemented to obtain an output voltage of 400 mV using a small photovoltaic cell for energy harvesting from indoor light. The proposed system can power an ultra-low-power device between 20 $\mathsf{\mu }$W and 40 $\mathsf{\mu }$W with a minimum input voltage of 230 mV. Electrical simulation results show that the implemented converter can achieve a peak efficiency of 81.24% at an input voltage of 260 mV for a 20 $\mathsf{\mu }$W load.

Energy Harvesting is the act of collecting energy from the environment and transforming it into electrical energy. Energy Harvesters in integrated-form are composed of many components, including a DC-DC converter. If this converter is implemented with switched-capacitors, an oscillator is necessary to provide the switching frequency. There is a great demand for ultra-low power and ultra-low voltage oscillators in Energy Harvesting applications, since switching frequency of the DC-DC converter can be adjusted to optimize conversion efficiency. This work proposes a 0.7 V open-loop digitally-controlled oscillator (DCO) in a frequency range of 300 kHz to 5 MHz and maximum power consumption of $5\ \mu \mathrm{W}$ in the TSMC 180 nm technology. It is based on a ring-oscillator with variable capacitive loading for a fine adjustment in frequency and current-starving transistors for frequency calibration. The frequency setting is performed with a 4-bit digital input. The implemented circuit layout occupies an area of 0.0348 mm ² . Post-layout simulations have shown monotonic and linear frequency variation with a peak power consumption of $4.9\ \mu \mathrm{W}$ at the maximum frequency of 5 MHz. The calibration capability compensates frequency degradation in post-layout simulations due to significant parasitic capacitances.

Energy-harvesting circuits are promising candidates to enable ultra low power Internet-of-Things devices. In the case of solar-powered circuits, they suffer from output voltage variations due to unstable incident light intensity. Those circuits require power management circuitry to supply a stable voltage to the main circuit despite the light-source variations. Self-starting the power management circuit from an ultra low-voltage (ULV) source is a challenging task, and this work introduces the design of a cold-start integrated system based on a ring-oscillator, clock-boosting techniques and a charge pump for use in an energy harvesting DC-DC converter for ULV applications. The cold-start circuit must provide a 650 mV, 10 µW signal to drive the main converter. It is designed in 180-nm CMOS technology and requires a minimum start-up voltage of 195.79 mV, providing a start-up time of 30.7 µs.

Electrical models are used to compare the properties of PV cells and solar panels and to predict the characteristics of their I-V curves. Even though there is much research on the evaluation of these models for outdoor conditions, their indoor performance is still not fully investigated, despite the increasing interest for IoT applications. This work describes a method for extracting the parameters of a $20\times 40$ mm PV cell for the 1D2R electrical model at indoor conditions. An analytical procedure is presented, in which the 5 model parameters are calculated from 6 PV cell characteristic points. Translation equations are estimated for the characteristic points in function of illuminance level, reducing the amount of measurements. Results demonstrated good accuracy of the electrical model concerning measurement data for both procedures, with R ² greater than 0.97. This modeling is fundamental for emulating the electrical behavior of PV cells under indoor conditions and can be used as input for characterization of energy harvesting circuits.