IEEE PEDS 2019, Toulouse, France

9 – 12 July 2019

Design of an Adaptive Impedance Matching Converter in

Microwave Power Transfer System
Xirui Zhu, Ke Jin, Qi Hui
Nanjing University of Aeronautics and Astronautics, Jiang Su, China
[email protected]

Microwave Microwave
transmitter receiver
Abstract—Microwave power transfer (MPT) is one of the most
promising systems in the long distance wireless power transfer
system. In MPT system, conventional diode rectifying circuit
produces harmonic energy which has great impact on the overall Microwave Microwave Microwave
power generator radiator Rectifier Rload
efficiency as well as the power reflection. In order to gain adaptive
impedance matching and high efficiency, this paper applies power
electronic devices and topology to construct a two-stage 100MHz Microwave Microwave Receiving Rectifying
converter. 100MHz belongs to the radio frequency (RF) and generating unit radiation unit Free space antenna circuit

belongs to the conception of microwave frequency. The two-stage

adaptive impedance matching circuit and its control method are
Fig. 1. Microwave power transfer system
evaluated by simulations. The design of microstrip antenna is also
presented in this paper, experimental results are shown to verify
the effectiveness of the proposed RF-DC converter and the Rectifier Circuit
efficiency is up to nearly 65% as well.
Load
Matching DC Pass
Antenna Resistor
I. INTRODUCTION Network Filter
(RL)
As a novel technology of power transfer, wireless power
transfer can transmit power without any electrical wire or
physical connections. Microwave power transfer (MPT) has Fig. 2. Conventional rectenna schematic
become a popular technology both on academic and industry
area. It takes a lot of advantages to the people who live far parameters of the converter are designed critically and the
away from the city and are uneasy to get electricity. Also, when simulation results show that the converter realizes the stable
transmitting power in microwave, it produces less power loss input impedance. Microstrip antenna design is also presented
in the air than the laser. What’s more, it provides the new idea in this paper, its characteristics could meet the requirements of
for the further power supply system. Therefore, it is receiving antenna in MPT system. Finally, the simulation and
meaningful to raise the efficiency of whole MPT system. experimental
Fig. resultsasare
1. Magnetization shownoftoapplied
a function verifyfield.
the effectiveness
Note that “Fig.” of
is
abbreviated.
the proposed There is a period RF-DC
two-stage after the figure number, followed by two spaces.
converter.
The structure of the microwave power transfer system is It is good practice to explain the significance of the figure in the caption.
shown in Fig.1. Microwave transmitter and receiver constitute II. DESIGN OF RF-DC CONVERTER
the enormous system. The receiving terminal which composes
A. Parameter design of Class E rectifier
of receiving antenna and rectifying circuit is called rectenna.
The circuit of two-stage RF-DC converter is shown in Fig.
Conventional rectenna schematic is shown in Fig. 2 [1].
3. It contains two parts, including a current-driven resonant
Rectifying circuit as an important part of the receiver has
Class E rectifier and a simple Buck DC-DC converter. The
great influence on the overall RF-DC efficiency. When
input source of the converter which is a sinusoidal current
rectified only by a single Schottky diode, there should be some
source chosen for appropriate analysis purposes can be
harmonic component generated which seriously affects the
substituted by a sinusoidal voltage source and a LC resonant
overall efficiency. In addition, power reflection often occurs in
tank connected in series [3]. In the proposed design, the
the occasions of high frequency which the source impedance
resonant tank of the first-stage rectifier is tuned on resonance
does not match the load impedance or the transmission line. So,
at 100MHz in order to make the input current substantially
this mismatch will reduce the efficiency, either [2]. Moreover,
sinusoidal. Different from other traditional rectifiers [4], the
in MPT system, the efficiency of the rectifying circuit changes
proposed Class E rectifier considers the inductor Lr and
when the input RF power or load impedance changes. This
capacitor Cr which are consisted of the diode capacitance plus
phenomenon often happens due to the distance changing
extra capacitance in the resonant condition. The current
between the transmitter and the rectenna.
through Lr is not dc any more. There are some ac current
Therefore, in this paper, in order to raise the RF-DC
ripples through it. The second-stage converter is designed
efficiency regardless of input or output conditions, a strategy
traditionally and works at 200kHz. Because of the frequency
of designing two-stage RF-DC converter is proposed in this
gap between the first-stage rectifier and the second-stage DC-
paper. This strategy could keep input impedance constant. The

978-1-5386-6499-5/19/$31.00 ©2019 IEEE

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 09:30:33 UTC from IEEE Xplore. Restrictions apply.
   Zr 
L s Cs Lr Q L  I in Lr sin(t ) sin( ) −  L sin(r t ) sin( ) 
vm + vo
-
  2  r

vin D1 Cr Cm D2 (
  ) − 1  + cos(r t ) cos( ) − cos(t ) cos( ) 
Zin Zo vD (t )=  r (5)
Zbuck −Vm cos(r t ) − I in Z r sin(r t ) sin( ) + Vm
Fig. 3. Architecture of two-stage RF-DC converter  T0  t  T1

DC converter, it is reasonable to analyze the circuit operation 0 T1  t  T2
process separately. Some assumptions are made to discuss the
former parameter analytically:    Lr 
1) The diode is ideal without any on-resistance except for  I in  cos(r t ) sin( ) + Z sin(r t ) cos( ) 
  2  r

the diode capacitance (Reference [5] mentions that the  (  ) − 1  − cos(t ) sin( ) − sin(t ) cos( ) 
influence of junction capacitance of switching devices on class  r
E rectifier); iLr (t )=  V (6)
− sin(r t ) + I in cos(r t ) sin( )
m

2) The forward voltage drop of the diode is neglected;  Zr

3) The diode capacitance does not change due to the voltage  T0  t  T1
above the diode;

0 T1  t  T2
4) The output ripple voltage is so small that the output In the equations (5) and (6), 𝜔ｒ is the resonant frequency of
voltage is constant. Lr-Cr and Zr is characteristic impedance of the network. The
specific values are shown in equations (7)-(8).
During a single switching period, there are two switching
modes in the rectifier. The details of waveforms can be found 1
r = (7)
in the Fig. 4 which divided into on and off mode of the diode. Lr Cr
a) The diode turns off at t=T0 when the current through it Lr
reaches zero. At the same time, the diode capacitor which is Zr = (8)
resonant with inductor starts to charge with zero initial current. Cr

Suppose the input current is: In addition, the rectifier shown in Fig. 2 should satisfy three
constraints to make it stable. The first constraint is diode
iin = Iin sin(t +  ) (1) voltage should be zero when the diode turns on. The second
In the diode-off mode, according to KVL and KCL constraint is average value of diode voltage should be equal to
equation, it can be concluded: output voltage Vm. The third constraint is average value of
current through Lr should be equal to output power Po divided
diLr (t ) by output voltage Vm.
vD (t ) = Lr + Vm (2)
dt vD (T1 ) = 0 (9)
dvD (t )
iin = Cr + iLr (t ) (3) 1

T

dt vD (t ) dt = Vm (10)
T 0
b) The diode turns on at t=T1 when the current through
capacitor converts to the diode itself immediately. Then on 1 Po

T
iL (t ) dt = (11)
account of the constant voltage drop on the inductor, the current T 0
Vm
through it decrease linearly. Not until the current through diode
According to the three constraints, three independent
reaches zero does it turn off again at t=T2 and the rectifier fulfill
equations (9) to (11) can be derived in terms of 8 independent
its circulation. According to the voltage drop, it can be
parameters, including duty ratio Drec, switching frequency fs,
concluded:
output power Po, output voltage Vm, resonant inductor Lr,
diLr (t ) capacitor Cr, input current magnitude Iin and phase φ.
Lr + Vm = 0 (4) Since the Ls-Cs is tuned on resonance, the input current is
dt
sinusoidal. There is only fundamental component in the input
Use equations (2), (3) and (4), the voltage on the diode current without any harmonics. So, it is logical to determine
capacitor and the current on the resonant inductor can be the input impedance of the current-driven rectifier at the
derived as: operating frequency. As shown in Fig. 4, the input impedance
is made up by a series combination of the input resistance Rin
and the input reactance Xin, both of which are determined at the

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 09:30:33 UTC from IEEE Xplore. Restrictions apply.
 TABLE I
vXin PARAMETERS OF RECTIFIER

fs 100MHz Cs 10pF
jXin
vrec Rin vRin Vin 31.65V Cr 69.9pF
vin
φ -50.37deg Lr 27nH

Zin Ls 250nH Vm 22.4V

Fig. 4. Circuit model about RF-DC converter
could be calculated and the parameters are listed in TABLE I.
operating frequency [5]. The Zin can be calculated as equation The simulation results are shown in section IV.
(12).
Z in = Rin + jX in (12) B. Conventional impedance matching strategy
At low frequency, where the electrical wavelengths of the
The fundamental component of the input voltage can be
signals are much longer than the physical dimensions of the
defined as
wires and lumped components, the phase of the voltage and
vrec = vR + vX current waveforms do not change significantly along the length
(13)
in in

= Vn , R sin(t +  ) + Vn , X cos(t +  ) of wires or components. Therefore, power is transmitted from

the source to the load with little loss. At RF or microwave
As mentioned above, the Ls-Cs resonant tank is tuned on frequency, where the signal waveforms are shorter than the
resonance at the operating frequency, therefore, the physical dimension of the network components. The amplitude
fundamental component of the input voltage is equal to the and phase of voltage and current waveforms change
diode voltage. Thus, input impedance can be calculated by the significantly as they travel from source to load. In this case,
expression of diode voltage vD using trigonometric Fourier power is reflected at discontinuities and the maximum power
series for real-valued signals shown as equations (14)-(15). will not reach the load. To achieve the maximum power
1 2
Vn , X =   vD (t )  cos(t +  ) dt  (14) transfer, the conventional strategy is adding the impedance
  0  matching network after the antenna, when the receiving power
1 changes, the characteristics of impedance could keep constant.
Vn , R =   2 vD (t )  sin(t +  ) dt  (15) Fig. 5 shows the diagram of antenna and impedance
  0 
matching network. In conventional method, impedance
Therefore, Rin and Xin can be expressed: matching network usually cascade after receiving antenna,
Vn , X
X in = =f ( Drec , f s , Po , Vm , Lr , Cr , I in ,  ) (16)
I in Rs RIN
+jXs +jX IN
Vn , R
Rin = =g ( Drec , f s , Po , Vm , Lr , Cr , I in ,  ) (17) Impedance
I in Vs Antenna
Network
The expression details about Rin and Xin are too complex to
list, but in fact, they can be expressed by the 8 independent
parameters mentioned above.
Besides, the two-stage RF-DC converter works as a receiver Fig. 5. Conventional strategy of impedance matching network
of MPT system, so the magnitude of the input impedance when the receiving power changes, the characteristics
should be kept at 50Ω and the phase difference between impedance will also change. To maintain the characteristics
voltage and current should be kept at 0. Because of the theory impedance, a passive impedance matching network is
of impedance matching in microstrip line. connected [6]-[7]. Usually use L-network for matching, in
Based on the equations (9), (10), and (11), 3 independent order to expend the range of impedance, two L-networks could
equations and 8 independent parameters can be concluded. also be used to achieve broadband impedance matching.
These equations can be solved numerically by code in However, this conventional strategy is passive, couldn’t reach
MATLAB using fsolve function. But actually, the switching the target of adaptive impedance matching.
frequency fs is determined by the applications and also fixed
on 100MHz. The output power Po is also determined by C. Proposed adaptive impedance matching strategy
applications as well. The only changeable parameter is the In this paper, a novel strategy of adaptive impedance
output voltage Vm. matching is presented. This strategy uses a DC-DC Buck
On account of these constraints and limitations, Vm that can converter which cascade after the rectifier and control the duty
make a tradeoff on most of the situations. The parameters ratio of MOSFET, could keep input impedance constant.
Owing to the difficulty in sampling the voltage and the current

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 09:30:33 UTC from IEEE Xplore. Restrictions apply.
in the very high frequency circuit, it is hard to know the 𝑊
(𝜀𝑒 + 0.3)(
+ 0.264)
accurate value of power even though using high speed digital ∆L = 0.412h ℎ (26)
sampling chip. For this problem, achieving maximum power 𝑊
(𝜀𝑒 − 0.258)( + 0.8)
ℎ
from the input is equivalent to set the input impedance constant
on 50Ω. On the other hand, as long as the input impedance is Microstrip Patch
Antenna
kept constant on 50Ω, maximum power from the input can be
achieved as well. So, derive the relation with duty ratio of DC-

1
W

W
DC Buck converter could achieve constant input impedance.

L0
L Feeding Point

Equations (18) to (21) show the derivation process clearly. L1

Dielectric
Vo = DBuckVm (18) Substrate
h

Pbuck = Po Ground
Plane

2 2 Fig. 6. Schematic of microstrip antenna

Vm Vo (19)
= Fig. 6 shows the schematic of microstrip antenna, since the
Rbuck Ro operation frequency is 100MHz, the size of antenna is quite big.
1 In this paper, the derivation of RF-DC lays a foundation of
Rbuck = 2
Ro (20) future development and the design procedure of microstrip
D
2
antenna can still be used in high frequency which up to GHz.
Vm
Rbuck = (21) TABLE II.
Po PARAMETERS OF ANTENNA
Parameter Value
III. DESIGN OF MICROSTRIP ANTENNA
W 0.7m
In MPT system, antenna is also an essential component. In L 0.9m
L0 0.18m
order to achieve the high transfer efficiency, antenna should h 1.6mm
have less power reflect. Some of the bandwidth extension εr 4.4
techniques include use of frequency selective surfaces, low W1 1.4m
L1 3.0m
dielectric substrate, multiple resonators, thicker substrate,
employing stacked configuration and slot antenna geometry IV. SIMULATION AND EXPERIMENTAL RESULTS
[8]-[10]. The parameters are marked on the Fig. 6. L, W, L0 are
the length, width, and the feeding point length respectively. Fig. 7 shows the waveform of input voltage and current
In this paper, general parameters are calculated. Limited to under input power 5W and 10W in simulation, which confirms
paper’s length, the specific derivation will not be presented in that the RF-DC converter can realize the input impedance
detail. equal to nearly 50Ω. Because of the step size of the simulation,
Single microstrip antenna has some disadvantages, such as the waveform is a little bit distorted. Fig. 8 shows the
low radiation gain, poor directional performance. So, the
receiving side of MPT system usually has an antenna array. In
this paper, a microstrip antenna is described in detail, and the
impedance of antenna is set to 50Ω.
Using equations (22) to (26) can calculate the general
parameters, C is the speed of light, f is the operation frequency,
𝜀𝑒 and ∆L are effective dielectric constant and effective
radiation gap length respectively. However, these parameters
still need to be validated in simulation, in this paper, the
software HFSS is used to simulate the parameters. The Fig. 7 The input voltage and current simulation
parameters after simulation is shown in table II.
60 Pin=10W
c 𝜀𝑟 + 1 −1 iL
W= ( )2 (22) 50 vD
2f 2
𝑐 40
λ𝑒 = (23)
𝑓√𝜀𝑒 30

𝑐 20
Time:[2ns/div]
L= − 2∆𝐿 (24)
2𝑓√𝜀𝑒 10

𝜀𝑟 + 1 𝜀𝑟 − 1 ℎ 0

+ (1 + 12 ) (25)
2 2 𝑊 Fig. 8 The diode voltage and resonant inductor current simulation of the
Class E rectifier

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 09:30:33 UTC from IEEE Xplore. Restrictions apply.
waveform of the diode voltage and the resonant inductor Lr C1:VD [5V/div]
current, which confirms that the parameter chosen and rectifier
are in the normal operating condition. In this paper,
experimental results also used to verify the analysis above. C1 C2
The front and back of the two-stage converter are shown in
Time: [5ns/div] C2:IL [400mA/div]
Fig. 9 and Fig. 10.
In Fig. 11 shows the experimental waveforms of the diode Fig. 11 Schematic of microstrip antenna
voltage and inductor Lr current. The waveforms are not good C1:VD [20V/div] C2:Iin [700mA/div]
as the simulation results because of every diode has its own
turn-on threshold, forward voltage and on-resistance. Besides,
C1 C2
the value of the diode voltage below zero should be noticed. At
the same moment, the diode is still on the “diode off” mode C3:Vin [40V/div]
C3
while the capacitor Cr is resonating with the inductor Lr. And
Time: [5ns/div]
this explains why the duty ratio looks a little bit bigger in the
Fig. 12 Schematic of microstrip antenna
experiment.
Fig. 12 to Fig. 15 show the key waveforms of the two-stage C4:Dbuck [10V/div]
rectifier, including diode voltage vD, input current iin, input
voltage vin, duty ratio of the Buck converter Dbuck, output C4

voltage Vo and output current Io. From Fig. 12 and Fig. 14, the C5:Vo [10V/div]

input voltage and input current still have a little phase shift C5 C6
Time: [5µs/div] C6:Io [1A/div]
because of the detecting error and the parasitic inductor and
capacitor in the circuit. When Buck converter works on the Fig. 13 Schematic of microstrip antenna
different duty ratio, there is hardly any phase difference C1:VD [20V/div] C2:Iin [700mA/div]
between the vD and iin. It means adaptive impedance matching
network is working and when the output changes, it maintains
C1 C2
the input impedance and keep it to 50Ω.
Fig. 16 shows the efficiency of the two-stage converter when C3:Vin [40V/div]
C3
operates on Dbuck=0.5 and Dbuck=0.8. As shown in Fig. 16,
Time: [5ns/div]
compared with the lower duty ratio, the higher duty ratio has
Fig. 14 Schematic of microstrip antenna
higher efficiency because Buck converter works more
efficiently on the higher duty ratio. Besides, the overall C4:Dbuck [10V/div]
rectifier efficiency is not so high because the working C4
frequency is too high that some power may be radiated to the
outside space. Even though there is some power loss, the C5:Vo [5V/div] C6:Io [200mA/div]
converter still achieves expectations. The experimental results C5
Time: [5µs/div]
C6

and theoretical analysis are well matched. Fig. 15 Schematic of microstrip antenna
80
70
60
Efficiency (%)

50
40
30
20 D=0.5
10 D=0.8
0
0 2 4 6 8 10 12 14
Output Power (W)
Fig. 9. Front of the proposed converter
Fig. 15 Schematic of microstrip antenna

V. SIMULATION AND EXPERIMENTAL RESULTS

A methodology of designing two-stage RF-DC converter is
proposed in this paper. The characteristics impedance is
maintained at 50Ω, so that the two-stage RF-DC converter
could transmit at maximum power. The parameters about first
stage Class E rectifier and microstrip antenna are well designed.
Fig. 10. Back of the proposed converter
Moreover, an impedance-controlled strategy is also presented.
The simulation and experimental results are shown to verify

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 09:30:33 UTC from IEEE Xplore. Restrictions apply.
the theorical analysis above. To control the input impedance
when the load varies is an effective way to achieve adaptive
impedance matching. 100MHz belongs to RF frequency band
and also belongs to the category of microwave. Although the
frequency of 100MHz still has a gap with 2.4GHz or 5.8GHz,
the theoretic analysis which presented in this paper could lays
a foundation of future research. In the future, MPT could
increase the power level and transmission efficiency, thus
expending the scope of application.
ACKNOWLEDGMENT
This work has been supported by the 2018 Delta Young
Scholar Award Program, and by the Natural Science
Foundation for Distinguished Young Scholars of Jiangsu
Province (BK20181293).

REFERENCES
[1] B. Strassner, and K. Chang, “Microwave power transmission: historical
milestones and system components,” Proceedings of the IEEE, vol. 101,
no. 3, pp. 1379-1396, Jun. 2013.
[2] Y. Huang, N. Shinohara, and T. Mitani. “Impedance matching in wireless
power transfer,” IEEE Trans. Microwave The. and Tech., vol. 65, no. 2,
pp.582-590, 2013.
[3] J. A. Santiago-Gonz’alez, K.M. Elbaggari, K. K. Afridi and D. J.
Perreault. “Design of class E resonant rectifiers and diode evaluation for
VHF power conversion,” IEEE Trans. Power Electron., vol. 30, no. 9,
pp. 4960-4972, Sep. 2015.
[4] M. K. Kazimierczuk. “Analysis of class E zero-voltage-switching
rectifier,” IEEE Trans. Circuit and Systems, vol. 37, no. 6, pp. 747-755,
Jun. 1990.
[5] K. Sakuma, and H. Koizumi. “Influence of junction capacitance of
switching devices on class E rectifier,” 2009 IEEE Circuits and Systems,
pp. 1965-1968, Feb, 2017.
[6] C. S. Hemsi and C. M. Panazio, "Adaptive baseband fre-equalization for
RF impedance matching correction," 2017 IEEE International
Symposium on Circuits and Systems (ISCAS), Baltimore, MD, 2017, pp.
1-4.
[7] P. R. Chin, A. K. M. Wong, K. I. Wong and N. Barsoum, "Modelling of
LCRC adaptive impedance matching circuit in narrowband power line
communication," 2015 IEEE 11th International Conference on Power
Electronics and Drive Systems, Sydney, NSW, 2015, pp. 132-135.
[8] G. R. DeJean, T. T. Thai, S. Nikolaou and M. M. Tentzeris, "Design and
Analysis of Microstrip Bi-Yagi and Quad-Yagi Antenna Arrays for
WLAN Applications," IEEE Antennas and Wireless Propagation Letters,
vol. 6, pp. 244-248, 2007.
[9] H. Wang, D. G. Fang and M. Li, "A Single-Channel Microstrip
Electronic Tracking Antenna Array With Time Sequence Phase
Weighting on Sub-Array," IEEE Transactions on Microwave Theory and
Techniques, vol. 58, no. 2, pp. 253-258, Feb. 2010.
[10] W. Wang, J. Wang, A. Liu and Y. Tian, "A Novel Broadband and High-
Isolation Dual-Polarized Microstrip Antenna Array Based on Quasi-
Substrate Integrated Waveguide Technology," IEEE Transactions on
Antennas and Propagation, vol. 66, no. 2, pp. 951-956, Feb. 2018.

Authorized licensed use limited to: Auckland University of Technology. Downloaded on June 08,2020 at 09:30:33 UTC from IEEE Xplore. Restrictions apply.