Power Management Techniques for Supercapacitor Based IoT Applications

Abstract

The emerging internet of things (IoT) technology will connect many untethered devices, e.g. sensors, RFIDs and wearable devices, to improve health lifestyle, automotive, smart buildings, etc. This thesis proposes one typical application of IoT: RFID for blood temperature monitoring. Once the blood is donated and sealed in a blood bag, it is required to be stored in a certain temperature range (+2~+6°C for red cell component) before distribution. The proposed RFID tag is intended to be attached to the blood bag and continuously monitor the environmental temperature during transportation and storage. When a reader approaches, the temperature data is read out and the tag is fully recharged wirelessly within 2 minutes. Once the blood is distributed, the tag can be reset and reused again. Such a biomedical application has a strong aversion to toxic chemicals, so a batteryless design is required for the RFID tag. A passive RFID tag, however, cannot meet the longevity requirement for the monitoring system (at least 1 week). The solution of this thesis is using a supercapacitor (supercap) instead of a battery as the power supply, which not only lacks toxic heavy metals, but also has quicker charge time (~1000x over batteries), larger operating temperature range (-40~+65°C), and nearly infinite shelf life. Although nearly perfect for this RFID application, a supercap has its own disadvantages: lower energy density (~30x smaller than batteries) and unstable output voltage. To solve the quick charging and long lasting requirements of the RFID system, and to overcome the intrinsic disadvantages of supercaps, an overall power management solution is proposed in this thesis. A reconfigurable switched-capacitor DC-DC converter is proposed to convert the unstable supercap's voltage (3.5V~0.5V) to a stable 1V output voltage efficiently to power the subsequent circuits. With the help of the 6 conversion ratios (3 step-ups, 3 step-downs), voltage protection techniques, and low power designs, the converter can extract 98% of the stored energy from the supercap, and increase initial energy by 96%. Another switched-inductor buck-boost converter is designed to harvest the ambient RF energy to charge the supercap quickly. Because of the variation of the reader distance and incident wave angle, the input power level also has large fluctuation (5uW~5mW). The harvester handles this large power range by a power estimator enhanced MPPT controller with an adaptive integration capacitor array. Also, the contradiction between low power and high tracking speed is improved by adaptive MPPT frequency.