Abstract-This paper describes a new maximum-power-pointtracking (MPPT) method focused on low-power (< 1 W) photovoltaic (PV) panels. The static and dynamic performance is theoretically analyzed, and design criteria are provided. A prototype was implemented with a 500-mW PV panel, a commercial boost converter, and low-power components for the MPPT controller. Laboratory measurements were performed to assess the effectiveness of the proposed method. Tracking efficiency was higher than 99.6%. The overall efficiency was higher than 92% for a PV panel power higher than 100 mW. This is, in part, feasible due to the low power consumption of the MPPT controller, which was kept lower than 350 μW. The time response of the tracking circuit was tested to be around 1 s. Field measurements showed energy gains higher than 10.3% with respect to a direct-coupled solution for an ambient temperature of 26• C. Higher gains are expected for lower temperatures. Index Terms-Energy harvesting, maximum power point tracking (MPPT), solar cells, wireless sensor networks (WSNs).

This paper proposes a closed-loop maximum power point tracker (MPPT) for subwatt photovoltaic (PV) panels used in wireless sensor networks. Both high power efficiency and low circuit complexity are achieved. A microcontroller (µC) driven by a fast clock was used to implement an MPPT algorithm with a low processing time. This leads to a maximum central-processing-unit duty cycle of 6% and frees the µC to be used in the remaining tasks of the autonomous sensor, such as sensing, processing, and transmitting data. In order to reduce power consumption, dynamic power management techniques were applied, which implied the use of predictive algorithms. In addition, the measurement and acquisition of the output current and voltage of the PV panel, which increase circuit complexity, was avoided. Experimental measurements showed power consumptions of the MPPT controller as low as 52 µW for a 2.7-mW PV power and up to 388 µW for a 94.4-mW PV power. Tracking efficiency was higher than 99.4%. The overall efficiency was higher than 90% for a PV panel power higher than 20 mW. Field measurements showed an energy gain 15.7% higher than that of a direct-coupled solution.Index Terms-Dynamic power management (DPM), energy harvesting, maximum power point (MPP) tracking, solar cells, wireless sensor networks.

Abstract-This paper analyzes the energy consumption of direct interface circuits where the data conversion of a resistive sensor is performed by a direct connection to a set of digital ports of a microcontroller (μC). The causes of energy consumption as well as their relation to the measurement specifications in terms of uncertainty are analyzed. This analysis yields a tradeoff between energy consumption and measurement uncertainty, which sets a design procedure focused on achieving the lowest energy consumption for a given uncertainty and a measuring range. Together with this analysis, a novel experimental setup is proposed that allows one to measure the μC's timer quantization uncertainty. An application example is shown where the design procedure is applied. The experimental results fairly fit the theoretical analysis, yielding only 5 μJ to achieve nine effective number of bits (ENOB) in a measuring range from 1 to 1.38 k . With the same ENOB, the energy is reduced to 1.9 μJ when the measurement limits are changed to 100 and 138 k .

No abstract

A new technique to measure a capacitor or a capacitive sensor by means of a direct sensor-to-microcontroller interface circuit that does not need a calibration capacitor is proposed. Basically, the measurement process consists of three consecutive steps of charge, discharge and charge of the capacitor under test. A non-linear equation is obtained and solved that is dependent only on known circuit parameters. Experimental results show that it is possible to measure a wide range of capacitor values with a maximum deviation of 2 % from the reference value, and that temperature changes from 18 °C to 70 °C yield relative errors below 0.1 %. For the lowest measured capacitor range (33 pF to 4.7 nF) the uncertainty holds below 1 pF which enables measurement of commercially available capacitive sensors. The main advantage of the proposed technique is cost and space reduction of the final design.Introduction: Capacitive sensors are used in a growing number of applications. In order to reduce the cost, complexity, space and power consumption of its interface circuits, direct capacitive sensor-tomicrocontroller interfaces have been proposed. There, the sensor is connected directly to the microcontroller (µC) without using either a signal conditioner or an analogue-to-digital converter. These circuits are an evolution of the conditioning circuits found in digital capacitance meters based on measuring the charging/discharging time of an RC circuit that were proposed before µCs became available [1][2] and that needed several digital and analogue discrete components to measure the desired time interval. Fig. 1a shows the simplest direct capacitor-to-microcontroller interface [3]. In this circuit, the unknown capacitor Cm is charged and discharged consecutively by changing the settings of two digital ports, PA and PB. The first stage ensures that Cm is fully charged by setting PA and PB as outputs. PA outputs a high level (VOH, "1") while PB outputs in low level (VOL, "0".) Afterwards, Cm is discharged towards VOL through R by setting PA as an input hence behaving as a Schmitt trigger in high impedance state (HZ) while PB remains the same. Simultaneously, the μC embedded timer starts counting until VPA reaches VTL, the port's lower threshold voltage. At this point, PA is triggered and the timer stops yielding Tm ≈ NmTCLK where TCLK is the clock period and Nm is the timer register count. Fig. 1b shows the waveform at PA that follows the voltage drop across Cm. Then, Cm can be estimated by

Solar energy radiation measurements are essential in precision agriculture and forest monitoring and can be readily performed by attaching commercial pyranometers to autonomous sensor nodes. However this solution significantly increases power consumption up to tens of milliwatts and can cost hundreds of euros. Since many autonomous sensor nodes are supplied from photovoltaic (PV) panels which currents depend on solar irradiance, we propose to double PV panels as solar energy sensors. In this paper, the inherent operation of the low-power solar energy harvester of a sensor node is also used to measure the open circuit voltage and the current at the maximum power point (I MPP), which allows us to determine solar irradiance and compensate for its temperature drift. The power consumption and cost added to the original solar energy harvester are minimal. Experimental results show that the relation between the measured I MPP and solar irradiance is linear for radiation above 50 W/m 2 , and the relative uncertainty limit achieved for the slope is ±2.4% due the light spectra variation. The relative uncertainty limit of daily solar insolation is below ±3.6% and is hardly affected by the so called cosine error, i.e. the error caused by reflection and absorption of light in PV panel surface.

No abstract

Salinity is a key parameter determining water quality. In agriculture, irrigation water with high salinity levels impacts plant growth and yield negatively hence there is a need to monitor it and irrigate with fresh water whenever salinity surpasses the tolerance threshold of a cultivar. Electrical conductivity (EC) of water is usually measured periodically instead of salinity because of its practicality. This work describes a low-cost low-power (and yet inexpensive) autonomous sensor prototype that is able to compute continuously the EC of water from complex electrical impedance measurements based on synchronous sampling, a technique that lowers cost and power consumption. The sensor is easy to assemble and has been verified in the lab for an EC range from 0.35 dS m −1 to 6.18 dS m −1 , showing a maximal deviation of ± 0.03 dS m −1 from the readings of a commercial reference EC meter. The sensor has also been installed in a rice paddy and left unattended for a whole cultivation season (107 days). The maximum deviation observed during this field test is ± 0.14 dS m −1 , which is good enough to detect high salinity levels.