A High-Efficiency RF Harvester with Maximum Power Point Tracking ^†

Abstract

This paper presents the implementation of a high-efficiency radiofrequency (RF) harvester, which consists of a rectenna and a maximum power point tracker (MPPT). The rectenna was characterized from −30 dBm to −10 dBm at 808 MHz, achieving an efficiency higher than 60% at −10 dBm. Experimental results also show that the rectenna can be well modelled as a Thévenin equivalent circuit, which allows the use of a simple ensuing MPPT. The complete RF harvester was tested, achieving an overall efficiency near 50% at −10 dBm. Further tests were performed powering a sensor node from a nearby antenna.

1. Introduction

RF energy harvesting has been extensively proposed to power tiny devices such as RFID or IoT (Internet of Things) nodes [1,2,3,4]. RF energy can be harvested either from dedicated sources or from the RF energy already present in the ambient and coming from unintentional sources. Figure 1 shows the generic block diagrams of: (a) an RF harvester powering a sensor node, and (b) a sensor node. The rectenna transforms the RF signal to a dc voltage and the MPPT provides the optimum load to the rectenna in order to transfer maximum power to the sensor node.

This paper continues and complements the work presented in [5] in two ways. First, the rectenna is fully characterized and modelled as a Thévenin equivalent circuit. Secondly, an MPPT is added to automatically operate at the maximum power point (MPP) of the rectenna. At an input power of −10 dBm, the efficiencies of both the rectenna and the MPPT are high, achieving an overall efficiency around 50%. As an application example, the RF harvester is used to power a sensor node that forms part of a smart gas meter [6].

2. RF Harvester

Figure 2a shows the schematic circuit of the rectenna, which includes an antenna, a high-pass L-matching network (composed of a capacitor Cm and an inductor Lm), a half-wave rectifier and an output filtering capacitor (Co). The antenna is modelled as a sinusoidal voltage source va in series with a radiation resistance Ra and provides a power Pav at matching conditions. The parameters Vo, Io, and Po are the DC voltage, current, and power at the rectenna output, respectively. An equivalent resistance Ro is defined as Vo/Io.

The rectenna output can be modelled as a Thévenin equivalent (a dc voltage source Voc in series with a resistance RT), such as in Figure 2b. The calculus of the Thévenin parameters is simple assuming no losses in the components and the diode [7] but otherwise is rather complex [2]. On the other hand, the parameters can be inferred by experimental characterization [3], which will be the strategy followed here. According to this model and to the maximum power transfer theorem, maximum power will be achieved when the ensuing stage presents an equivalent input resistance Ro = RT, which implies Vo = 0.5Voc.

In general, a sensor node directly connected to the output of the rectenna will not accomplish that condition and an intermediate stage, such as an MPPT, has to be used. An MPPT mainly consists of a DC/DC converter and a tracking algorithm. In particular, the fractional open circuit voltage (FOCV) technique leads to simple power-efficient implementations. In this technique, the open circuit voltage, V_oc in Figure 2b, of the energy transducer is first measured and a fraction k of V_oc is used in order to fix V_o to the MPP voltage, V_MPP. Assuming the model of Figure 2b, k = 0.5 is the proper choice and thus V_MPP = 0.5V_oc. Figure 3 illustrates the block diagram for an implementation of the FOCV-MPPT technique, where P_load is the power transferred to the sensor node. The operation is the following. First, S1 is closed and S2 is open (sampling period). For high values of R_oc1 and R_oc2, the output of the rectenna can be considered as open and thus V_o = V_oc. The voltage divider formed by R_oc1 and R_oc2 fixes V_MPP = kV_oc. The input capacitor C_L momentarily stores the incoming harvested energy. Secondly, S₁ is open and S₂ is closed (regulation period). Thus, V_MPP holds constant thanks to C_REF, and the DC/DC converter settles V_o around V_MPP and transfers the harvested energy by the rectenna to the node. In order to increase the efficiency at light loads, the DC/DC converter uses special control regulation techniques such as pulse-frequency modulation (PFM) or burst-mode [8].

3. Materials and Methods

The rectenna presented in [5] was used here. It consists of a printed circuit board with Rogers substrate and the following components: Cm = 0.5 pF (AVX), Co = 1 nF, Lm = 27 nH (Coilcraft), and a Schottky HSMS-2850 diode (Avago Technologies). The circuit of Figure 2a was used for its characterization, where an RF generator (Agilent E4433B) is connected at the input instead of the antenna and a Source-Measure Unit (SMU, Agilent B2901A) configured as a voltage sink (quadrant IV) at the output. The generator was set at a tuned optimal frequency of 808 MHz and at Pav of −30 dBm, −20 dBm, and −10 dBm. For each value of Pav, the SMU was set at different values of Vo while measuring Po. Then, ηrect was obtained as Po/Pav.

As for the FOCV-MPPT, a BQ25504 chip (Texas Instruments) was used, and in particular an evaluation board from the manufacturer. The chip contains a boost converter with PFM control and the board includes CL = 4.8 μF and Cload = 104.8 μF. The default values of ROC1 and ROC2 were modified to 10 MΩ in order to fix k = 0.5 (by default 0.78). The sampling and regulation periods are prefixed to 256 ms and 16 s, respectively. Then, the efficiency of the whole RF harvester (rectenna plus MPPT) was characterized by using the RF generator at the input of the rectenna and the SMU set at 3 V at the output of the MPPT. The overall efficiency, ηT, was calculated as Pload/Pav. The efficiency of the MPPT, ηMPPT, was not directly measured but can be inferred from

η_T = η_rect,maxη_MPPT,

(1)

where ηrect,max is ηrect at the MPP.

A setup was also used to power a sensor node with the RF harvester. The selected sensor node is used to upgrade a mechanical gas meter to a smart device [6]. For the tests, the node was programmed to stay in a standby mode, consuming about 1.4 μA. The input power (Pav) was set to keep the voltage supply of the sensor node around 3 V, thus Pload ≈ 4.2 μW. The required value of Pav can be estimated from

P_load = η_TP_av.

(2)

As for the RF harvester input, two configurations were used: (1) an RF generator and (2) a receiving monopole antenna. In the second case, another identical monopole antenna was connected to a nearby RF generator, jointly acting as a wireless energy transmitter. The antennas showed an insertion loss higher than 10 dB at 808 MHz.

4. Experimental Results and Discussion

Figure 4a shows the experimental values of ηrect (in dots) versus Vo for the three values of Pav together with least-squares fits (continuous lines) corresponding to a Thévenin equivalent model such as that of Figure 2b. The continuous lines show good agreement with the experimental data, thus validating the Thévenin model.

Table 1. Inferred and measured parameters from the data of Figure 4a.

Pav/dBm	RT/kΩ	Voc/mV	ηrectmax/%	VMPPexp/mV	Vocexp/mV
−10	3.80	958	60.3	480	960
−20	4.79	275	39.3	130	280
−30	6.29	58.7	13.6	27	60

shows for each value of Pav, the inferred Thévenin parameters RT and Voc, the achieved ηrect,max and its corresponding voltage (VMPP,exp), and the experimental open circuit voltage (Voc,exp) of the rectenna. In Figure 4a, ηrect,max, VMPP,exp, and Voc,exp are also marked for Pav = −10 dBm. Achieved efficiencies (ηrect,max) are among the highest published in the literature for similar designs [5]. On the other hand, Voc (the Thévenin voltage) nearly matches Voc,exp. Finally, VMPP,exp equates or nearly matches 0.5Voc,exp, the regulated voltage at the input of the MPPT. Thus, the MPPT will be able to extract the maximum power (or nearly) from the rectenna.

Figure 4b shows the experimental values of ηT versus Pav. At −20 dBm, ηrect,max = 39.3% but ηT = 6.5%, resulting, from (1), in ηMPPT = 16.5%. This low value of ηMPPT is due to both a low input voltage value (140 mV = 0.5Voc,exp) and a low value of Po (≈3.9 μW = ηrect,maxPav). Contrariwise, at −10 dBm, ηrect,max = 60.3% and ηT = 48.6%, resulting in ηMPPT = 80.6%. At higher values of Pav (−5 dBm) ηT reached a value of 55.6%. Compared to [4], where a similar chip for the MPPT is used, ηT is quite higher.

When powering the sensor node, the required value of Pav was −17.6 dBm. This value fits well with (2), considering the corresponding efficiency in Figure 4b (≈24%). This performance was also tested with the antennas at a distance of 0.5 m and 1 m. The power output of the remote RF generator was tuned at appropriate values in order to operate the node, resulting in 8.0 dBm and 13.2 dBm, respectively. These values accounted for the respective link budgets.