Iet-Pel 2018 6206

Download as pdf or txt
Download as pdf or txt
You are on page 1of 8

IET Power Electronics

Special Issue: Advanced Technologies Utilised in Wireless Power


Transfer Systems

Improved CPT system with less voltage stress ISSN 1755-4535


Received on 11th November 2018
Revised 19th March 2019
and sensitivity using a step-down transformer Accepted on 7th May 2019
E-First on 27th June 2019
on receiving side doi: 10.1049/iet-pel.2018.6206
www.ietdl.org

Tarek M. Mostafa1 , Aam Muharam2,3, Aiguo Patrick Hu4, Reiji Hattori2


1Computer, Electrical and Mathematical Science and Engineering (CEMSE), Ali I. Al-Naimi Petroleum Engineering Research Center (ANPERC),
King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia
2ASEM, Interdisciplinary Graduate School of Engineering Sciences, Kyushu University, Fukuoka, Japan
3Research Centre for Electrical Power and Mechatronics Indonesian Institute of Sciences (LIPI) Bandung, Indonesia
4Department of Electrical and Computer Engineering, The University of Auckland, New Zealand

E-mail: tarek.mostafa@kaust.edu.sa

Abstract: This paper proposes a capacitive power transfer (CPT) system with a step-down transformer on the secondary side
to reduce the circuit quality factor (Q), and thereby reduce the sensitivity to parameters variations, as well as the voltage stress
across the coupling interface. The system operating principle is analyzed mathematically, and the focus is given to understand
the effect of the leakage inductance (Llk) of the non-ideal transformer on the system performance. The analytical and simulation
results show that at a given constant output power, the voltage across the plates is significantly reduced, and the system
becomes less sensitive to the coupling variations by increasing the turns ratio of the step-down transformer. It is found that Llk
can be advantageously utilized as a tuning inductor (L) or part of it. The proposed method is verified by building a prototype
CPT system that delivered 25 W at an operation frequency of 1 MHz, and an efficiency of more than 70%. Simultaneously, the
voltage stress across the single pair of the coupling plates is reduced from 252 V of a conventional CPT system without a step-
down transformer, to 50.4 V using a high-frequency transformer with a turns ratio of 5.

1 Introduction can be achieved by an inductive approach, acoustic approach, or a


capacitive approach.
In 1899, the concept of transferring power wirelessly was Capacitive power transfer (CPT), as one of the near-field WPT
introduced for the first time by Tesla [1]. Since then, researchers technologies, which uses a varying electric field to transfer power
have developed several techniques for moving electricity over wirelessly, is currently gaining significant research attention. CPT
different distances without wires. Some exist only as theories or is distinguished by several advantages such as low-standing losses
prototypes, but others are already in use. These techniques can be and electromagnetic interference, the ability to transfer power
classified based on the distance between the transmitter and through electrically isolated metal barriers, and is small in size and
receiver as shown in Fig. 1. Far-field power transmission light weight [2]. Accordingly, CPT technology is applied to a wide
technologies target long distances, often from multiple kilometres range of applications including integrated circuits [3], biomedical
ranges, to the limits of the earth to space distances. Laser power devices [4], consumer electronics [2, 5–8], and electric vehicles
beaming and microwaves (from purpose-designed antennas) are the [9].
forms of electromagnetic radiation best suited for energy transfer in Fig. 2 shows the structure of the typical CPT system [1, 10–13].
this range. Moderate distance (midrange near field) wireless power In case the capacitive interface has no air gap between the plates, it
transfer (WPT) is referred to the system that power transfers in the has the value typically in the range of hundreds of pico-Farads
range of few metres. The most popular technique used for this while it can reduce to few tens of Pico Farads for air gap existence
range is the magnetic resonance coupling. In the short distance case. This results in high-quality factor (Q) and voltage stress
(near field) WPT, the transmission distance is in centimetres. WPT across the capacitive interface. Consequently, the tuning is

Fig. 1 Wireless power transfer technologies

IET Power Electron., 2019, Vol. 12 Iss. 10, pp. 2634-2641 2634
© The Institution of Engineering and Technology 2019
Fig. 4 Simplified structure for system analysis

Table 1 Parameters used for modelling


Parameter Unit Value
PL Watt 25
frequency MHz 1
Cc1 = Cc2 = 2Cc nF 2
Fig. 2 Typical conventional CPT system
L µH 25.33
RL Ω 10

numbers of turns in primary and secondary sides, Np and Ns,


respectively. Diodes D1 to D4, are constructing a full wave rectifier
to obtain dc voltage and current at the load.
For analysis, the system topology of Fig. 3 is rearranged
assuming the parasitic resistances are neglected for all components
as shown in Fig. 4. The inverter in Fig. 3 produces a square wave
voltage. This voltage is applied to the input terminal of the LC
series resonant circuit. The resonant circuit frequency is tuned to
Fig. 3 Proposed CPT system with a step-down transformer the fundamental component of the inverter square wave, i.e. to the
switching frequency f. In consequence, the resonant circuit current,
challenging to reach practically, and the system is highly sensitive as well as the resonant circuit output voltage, has essentially
to coupling variation and load change. The high voltage stress sinusoidal waveforms at the fundamental frequency f, with
across the coupling interface increases the probability of the negligible harmonics. Accordingly, the sinusoidal approximation is
dielectric breakdown and spark occurrence. The system safety first considered where input voltage vs1 can be obtained from (1).
becomes an issue, the related leakage electric field emissions Equivalent capacitance Cc represents the combination of the two
around the plates may increase than the safe margins. These interface capacitances as expressed in (2). The series compensation
matters cause less efficiency, ability to deliver power becomes inductor L is obtained from (3). The equivalent AC load R [15],
limited, and the difficulty in offering the commercially accepted assuming the transformer and the rectifier are ideal, is described in
system increases. (4).
Different topologies have been reported in the literature to
overcome the aforementioned problems. In [10], a tuning controller 2V D
was used to ensure the circuit operation at the optimum frequency vs1 = sin ωt (1)
π
and maximum efficiency over a wide range of capacitive interface
and load conditions. Although the system showed stability and Cc1 Cc2
80% efficiency, the design is complex and costly. Also, the existing Cc = (2)
Cc1 + Cc2
standard bands, e.g. ISM band, are usually too narrow to
accommodate the frequency tracking; accordingly, the use of such
1
systems become challenging commercially. In [13, 14], inductor– L = (3)
capacitor (LC) and LCLC matching networks were used, Cc ω2
respectively. Both systems are complicated to design and costly
due to the added extra-reactive components. Also, with such 8 2
R = n RL = n2RLoad (4)
topologies, the conduction losses in the cascaded LC resonant π2
parameters may be unavoidable except in one precise operation
point [15]. In [14], Q is very high (>100) which may cause where ω is the angular switching frequency ( = 2πf), and n is the
sensitivity issues especially in low power ranges, i.e. tens of Watts. step-down transformer turns ratio ( = Np/Ns).
This paper proposes a CPT system with a step-down The values in Table 1 are used as a design example to prove the
transformer on the receiving side to reduce the system sensitivity to concept. The value of the compensation inductor has been
parameters changes and lower voltage stress across the interface. calculated according to (3) so that the LC circuit resonates at 1
The rest of the paper is arranged as follows. Section 2 introduces MHz.
the proposed system structure and the main assumptions. Section 3 In the following section, the system's Q and the voltage across
presents a mathematical model and its verification. The the interface Vc are investigated as a function of n. By increasing n,
experimental results of an implemented prototype are shown in the transformer secondary voltage VNs and accordingly the output
Section 4. And finally, the conclusions are drawn in Section 5. voltage VL and the output power PL will reduce. Accordingly, it is
assumed that a sufficient voltage is fed from the source to deliver
2 System structure and modelling the same amount of power to the load at any n, so a comprehensive
The structure of the proposed CPT system is shown in Fig. 3. analysis and fair comparison are achieved. Compensating the
Switches S1 and S2 construct a half-bridge inverter (full bridge can stepped down voltage can be realised in several ways. The simplest
be used for higher power applications) to convert V D into an AC two ways are, first, increasing the input voltage as a function of n
according to
signal. Inductor L is placed in series to compensate the capacitive
interface, where Cc1 and Cc2 are representing the capacitive
V D = nV D_conv (5)
coupling interface. T1 is an ideal step-down transformer with

IET Power Electron., 2019, Vol. 12 Iss. 10, pp. 2634-2641 2635
© The Institution of Engineering and Technology 2019
and

2nV D
vs1 = sin ωt = nvs1_conv, (6)
π

where VD_conv and vs1_conv is the DC input voltage and AC input


voltage of the conventional CPT shown in Fig. 1, respectively.
Fig. 5 shows how the input should change for the design example
with parameters listed in Table 1.
The second possible way is adding a step-up transformer after
inverter as shown in Fig. 6 with turns ration nT 2 given by

Np_T 2 1 Fig. 5 Needed input voltage to deliver 25 W and Q versus n


nT 2 = = (7)
Ns_T 2 n

where Np_T2 and Ns_T2 are the turns number of the primary and the
secondary sides of the step-up transformer T2, respectively.
For analysis, Fig. 6 can also be simplified by using sinusoidal
approximation as shown in Fig. 7, i.e. the input voltage can be
given by

vs1
vin = = nvs1 (8)
nstepup
Fig. 6 CPT system with step-up and step-down transformer
Ideally, both previously mentioned ways will give the same
results in terms of system sensitivity and voltage stress across the
capacitive interface. However, by considering the switching losses
in the inverter and losses due to the parasitic components of the
step-up transformer, the results may differ. A comprehensive
comparison between the two cases has been introduced in [16]. In
this paper, the system with a step-up transformer is chosen for the
analysis and the experimental verification.

3 Mathematical analysis Fig. 7 Simplified structure of CPT system with step-up and step-down
3.1 System sensitivity transformers for system analysis

With reference to (9), adding a step-down transformer to the


typical CPT system will decrease Q by an n2 factor as shown in
Fig. 5. The system with lower Q is less sensitive to parameter
changes such as interface misalignment and load variations. Also,
at a low Q, tuning to the resonant frequency is more attainable
practically. To have a less sensitive system, it is recommended to
use Q≤6 [12] which can be easily achieved by just choosing n≥2
for the chosen values of this design example. The recommended Q
will vary depending on the situation and the system design
parameters.

1 L 1 L Qconv
Q = = = (9)
R Cc n2RLoad Cc n2 Fig. 8 Frequency response of the load power at different turns ratios n

where Qconv is the quality factor of the conventional CPT system at different turns ratio, n. By increasing n, the output will have a
previously presented in Fig. 2. wider response which demonstrates lower Q and less system
Fig. 8 illustrates the frequency response of the output power at sensitivity to parameters changes.
the load that can be obtained from As mentioned before, the system ability to transfer power
becomes impenetrable to the large parametric variations with
(Cc2 Rload V in
2
n2) s2 attaining low Q. The capacitive coupling interface changes may
PL =
1 + 2 Cc Rload n2 s occur due to misalignment or existence of dust or water between
transmitting and receiving plates which change the dielectric value
+ 2 Cc2 L Rload n2 s3 (10) and accordingly the value of the effective capacitances. Fig. 9
+ Cc2 Rload
2
n4 + 2 L Cc s2 shows the output power as a function of the equivalent capacitance
Cc. For the conventional system, a 10% change in Cc of its original
+ (Cc2 L2)s4 value results in degradation of the amount of power delivered to
the load to almost 0 W that resonance frequency will vary by a
and great value due to the high Q. In contrast, by adding the step-down
transformer on the power receiving side, the system would be
n4 ω4 immune to parameters changes and the ability to transfer the power
PL = Cc2 Rload V in
2
2 ,
1 − 2 Cc L ω2 will increase even under extreme conditions. For the designed
(11) example, if step-down transformer with turns ratio n = 10 is used,
2
n4 Rload
2
ω2 the power can be still transferred with almost the same amount for
+Cc
+L2 ω4 any increase in the interface till quadruple the value of Cc. In
addition, decreasing the equivalent capacitance by half of its
2636 IET Power Electron., 2019, Vol. 12 Iss. 10, pp. 2634-2641
© The Institution of Engineering and Technology 2019
load situations, unlike the case of no step-down transformer
addition where Q reaches >100 for some load values. Hence, the
need for optimum load tracking systems to achieve matched load
and low switching losses becomes less.

3.2 Voltage stress across the capacitive interface


As previously mentioned, by having the step-down transformer on
the secondary side in the proposed system, the equivalent AC load
increases by increasing the turns ratio accordingly, as per (4).
Therefore, the input voltage to the LC resonant tank had to be
boosted by adding the step-up transformer as expressed in (8) and
Fig. 9 Tolerance of the output power with the change in the capacitive
the same amount of power could be transferred with smaller
interface value using different turns ratios n
current (I) to the load. The current flowing through the interface (I)
can be found from

Cc V in
I = (12)
Cc Ls2 + Cc Rs + 1

which can be expressed as a function of the turns ratio (n) and the
interface current of the conventional system at the resonance
frequency as

Iconv
I = (13)
n

where Iconv is the current passing through the resonant circuit, i.e.
Fig. 10 Q tolerance with the change of Cc L, Cc1, and Cc2 for the conventional system.
The reduction in the current through the circuit will result in
less voltage stress Vc across the one pair of the coupling plates as
described in

I V in
Vc = = (14)
2 Cc s 2 Cc Ls2 + 2 Cc Rs + 2

Fig. 12 shows how the frequency response of Vc varies with the


turns ratio of the transformer. The voltage stress across one pair of
the coupling plates (Vc) decreases proportionally with n at the
resonant frequency, as given in

V c_conv
Vc = (15)
n
Fig. 11 Q tolerance with the change of RL
where Vc_conv is the voltage across a single pair of the capacitive
coupling interface plates, e.g. Cc1, for the conventional system.
At low frequencies, the terms 2 Cc L ω2 and Cc2ω2 R2 + L2ω2
can be neglected and Vc can be presented as shown in

1
Vc ≃ V , (16)
2 in

which explains the high Vc at the low frequencies. According to


these results, less dielectric losses, and reduction of the likely
occurrence of the sparks and the insulator breakdown can be
fulfilled. The proposed system helps in using CPT technology
safely without the need for shielding or living object detection
Fig. 12 Frequency response of amplitude of voltage Vc across one pair of systems.
the coupling plates at different n
3.3 Non-ideal transformer effect
original value will almost not affect the transferred amount of
power to the load, which still introduces significant improvement As mentioned before by increasing the turns ratio n, the
compared with the conventional system. Thus, by using the step- performance of the system will improve, which raises the question
down transformer as a part of the secondary side, the misalignment as to which limit it can be increased. The answer is mainly in the
issue is less of concern. design of the transformer itself. Practically, the transformer is non-
By adding the step-down transformer, the Q value will always ideal and it has numerous parasitic components unlike what is
be low for any value of the Cc as shown in Fig. 10, which confirms considered so far in this literature. Although the design and
the system robustness. While, for the conventional CPT system, the modelling of the non-ideal transformers are beyond the scope of
Q will change significantly especially at the low values of Cc this paper, in this subsection, a glimpse is given about the effect of
resulting in high Q. the leakage inductance as it has a crucial effect on the proposed
Fig. 11 shows that for CPT system with step-down transformer, CPT system performance. Fig. 13 shows the proposed system
Q is less affected by the alternation of the load, especially in heavy
IET Power Electron., 2019, Vol. 12 Iss. 10, pp. 2634-2641 2637
© The Institution of Engineering and Technology 2019
Fig. 13 CPT system with non-ideal step-up and step-down transformers

sufficiently high, so that it is neglected in the simplified model. By


using high gauge Litz wire, especially on the low turns number
sides of the transformers (i.e. primary side of the step-up
transformer and secondary side of the step-down transformer),
lower winding resistances can be obtained, accordingly, the
winding resistances can be neglected, and the equivalent circuit can
be further simplified as shown in Fig. 14b that

Llk = n2LpT2 + LsT2 + LpT1 + n2 LsT1 (17)

where Llk is the total-leakage inductance.


By considering the leakage inductance, the output power can be
re-given by modifying (10) and (11) to

Fig. 14 Simplified circuit for CPT system with non-ideal step-up and step- (Cc2Rload V in
2
n2) s2
down transformer PL =
1 + 2 Cc Rload n2 s
+ 2 Cc2 L + Llk Rload n2 s3 (18)
2 2 4 2
+ Cc Rloadn + 2 L + Llk Cc s
+(Cc2 L + Llk 2)s4

and

n4 ω4
PL = Cc2 Rload V in
2
2
1 − 2 Cc L + Llk ω2
(19)
n4 Rload
2
ω2
+Cc2
+ L + Llk 2 ω4
Fig. 15 Frequency response of the load power at different leakage
inductance Llk respectively.
It is clearly expectable that a shift in the resonance frequency
presented in Fig. 6 after introducing all the possible parasitic will occur by considering the leakage inductance; this shift will
components for the step-up and the step-down transformers, i.e. increase with the larger values of Llk as shown in Fig. 15 for the
same design example values in Table 1.
• Lp: leakage inductance in the primary side. In most of the systems that use transformers, the leakage
• Ls: leakage inductance in the secondary side inductance is usually considered as a disadvantage and a source of
• Cp: equivalent lumped capacitance at the primary side. concern. Based on the previously achieved results, leakage
inductance can be used for the own benefit of the system.
• Cs: equivalent lumped capacitance at the secondary side.
Combining Llk with the tuning inductance L can result in a
• Rp: DC winding resistance for the primary side winding. reduction of L size; or in some not extreme cases and with the
• Rs: DC winding resistance for the secondary side winding. precise design of the transformers, L can be totally eliminated as
• Cw: lumped winding-to-winding capacitance. shown in Fig. 16 where Llk only is used for full tuning of the
• Rc: equivalent core-loss shunt resistance. capacitive interface at the resonance frequency. Accordingly,
• Lm: magnetising inductance. introducing the transformers to the system will improve the overall
performance in terms of power and efficiency while the cost and
complexity maintenance is low. In the following section, a system
As the leakage inductance effect is the main focus of this
that depends only on the leakage inductance for tuning is
subsection, the complicated Fig. 13 can be simplified to a more
introduced for proving the concept.
accurate than ideal transformers case of Fig. 6 and less complicated
than the exact model of non-ideal transformers [17] as shown in
Fig. 14a. The magnetising inductance Lm can vary significantly 4 Experimental results
from one transformer to another due to material variations, The CPT system in Fig. 16 is implemented for verification as
temperature, and frequency. Lm is required to generate the shown in Fig. 17 with system parameters listed in Table 2. The
magnetic field, and it is desirable to make it as large as practically capacitive coupling interface is constructed by two pairs of 100
possible to reduce the magnetising current, therefore the power mm × 100 mm square copper pads covered by AL-37BLK tape
losses. Accordingly, the core and the number of turns should be from 3M (ɛr=2.4) as a dielectric material. Two identical
designed carefully to reduce its effect. In the proposed CPT system, transformers were made in-house (design guideline is given in
the magnetising inductance Lm of the transformer is made Appendix A) with measured leakage inductance ∼15 μH each and

2638 IET Power Electron., 2019, Vol. 12 Iss. 10, pp. 2634-2641
© The Institution of Engineering and Technology 2019
Fig. 16 CPT system with no tuning inductance L

Fig. 18 Frequency response of the DC output power and efficiency

Fig. 17 Prototype of the proposed system


Fig. 19 Tolerance of the output power and efficiency with the change in
the equivalent capacitive interface
Table 2 Circuit parameters and system specifications
Parameter Unit Value
PL Watt 25
frequency MHz 1
Cc1 nF ∼2.1
Cc2 nF ∼2
n — 5
RL Ω 10
S1 and S2 SiC C2M0080120D
D1–D4 SiC NXPSC06650

Fig. 20 Prototype frequency response of the voltage across the single pair
turns ratio equal to 5. For the system's Q and voltage stress across of the capacitive interface Cc1 (amplitude)
the interface investigation, the frequency was swiped from 0.2–1.5
MHz. Also, misalignment was applied to the interface for the
sensitivity test. common ground effect. Although the same amount of power (25
Fig. 18 shows the frequency response of the DC output power W) is delivered to the load, the voltage stress across the capacitive
and efficiency of the conventional CPT system (a series inductor interface is divided by n by introducing the step-down transformer
with value 25 μH was used for tuning) and the proposed system as derived in (14) from the mathematical analysis section. The
with the step-down transformer. Q of the proposed system is maximum voltage stress across the capacitive interface (Cc1) is
clearly reduced compared to the conventional system as predicted reduced to about 50.4 V at n = 5, corresponding to an output power
by the analysis. In addition, the proposed topology improved the of 25 W, while 252 V is measured for the no step-down
overall system efficiency (from DC to DC) with >30%, as well as, transformer used case. The values of the experimental results of Vc
it is the ability to maintain this efficiency over a wide range of is higher than the calculated ones due to the non-ideal components
operating frequency which confirms the system's high robustness. in the system which increase the losses and decrease the efficiency,
As proved in the analysis section, by reducing Q, the system's so a higher input voltage (VD = 39 V) is needed to deliver 25 W
tolerance to misalignment should improve. Accordingly, for resulting in higher voltage stress across the interface.
verification, the capacitive interface was changed to several values By using 39 V DC input voltage (VD), the system could deliver
higher and lower than the original value, and the equivalent 25 W of power at the n = 5 and operating frequency 1 MHz, while
capacitance was calculated and recorded. As shown in Fig. 19, the the efficiency reached slightly >70%. Fig. 21 shows the
output power and the overall efficiency of the system with step- experimental waveforms of output voltage Vs and output current Is
down transformer are much less affected by the interface of the inverter. The leaky fields from the unshielded capacitive
misalignment compared to the conventional CPT system topology. interface affect the current measurement probe, specifically the
Even with the reduction in the equivalent capacitance value to the inductance of the probe, which results in a distorted current
half, the power and the efficiency could be maintained with almost waveform in the measurements. Fig. 22 shows the measured VCc1
the same numbers of operation at the original design interface and VL.
value. By increasing the number of turns, the range of the operation
with the same output power and efficiency will be wider. However,
parasitic should be considered to adjust to which limit n can be 5 Conclusions
increased, as mentioned before. This paper proposes a CPT system with step-down transformer
Fig. 20 shows the amplitude of voltage stress measured across added on the power receiving side to reduce the system sensitivity
the single pair of the capacitive interface (Cc1). A floating and the voltage stress across the coupling plates. The proposed
oscilloscope was used for the measurement to eliminate the system drastically reduces the system's quality factor; therefore,
IET Power Electron., 2019, Vol. 12 Iss. 10, pp. 2634-2641 2639
© The Institution of Engineering and Technology 2019
Emerging Technologies: Wireless Power Transfer (WoW), Chongqing, 2017,
pp. 1–6
[7] Muharam, A., Mostafa, T.M., Hattori, R.: ‘Design of power receiving side in
wireless charging system for UAV application’. 2017 Int. Conf. on
Sustainable Energy Engineering and Application (ICSEEA), Jakarta, 2017,
pp. 133–139
[8] Hu, A.P., Liu, C., Li, H.L.: ‘A novel contactless battery charging system for
soccer playing robot’. 2008 15th Int. Conf. on Mechatronics and Machine
Vision in Practice, Auckland, 2008, pp. 646–650
[9] Kim, J., Bien, F.: ‘Electric field coupling technique of wireless power transfer
for electric vehicles’. IEEE 2013 Tencon - Spring, Sydney, NSW, 2013, pp.
267–271
[10] Kline, M., Izyumin, I., Boser, B., et al.: ‘Capacitive power transfer for
contactless charging’. 2011 Twenty-Sixth Annual IEEE Applied Power
Electronics Conf. and Exposition (APEC), Fort Worth, TX, 2011, pp. 1398–
1404
[11] Huang, L., Hu, A.P., Swain, A.K., et al.: ‘Z-impedance compensation for
wireless power transfer based on electric field’, IEEE Trans. Power Electron.,
2016, 31, (11), pp. 7556–7563
[12] Mostafa, T.M., Bui, D., Muharam, A., et al.: ‘Capacitive power transfer
system with reduced voltage stress and sensitivity’, Appl. Sci., 2018, 8, (7), p.
1131
[13] Theodoridis, M.P.: ‘Effective capacitive power transfer’, IEEE Trans. Power
Fig. 21 Experimental waveform of the output voltage and current of the Electron., 2012, 27, (12), pp. 4906–4913
[14] Lu, F., Zhang, H., Hofmann, H., et al.: ‘A double-sided LCLC-compensated
inverter capacitive power transfer system for electric vehicle charging’, IEEE Trans.
Power Electron., 2015, 30, (11), pp. 6011–6014
[15] Erickson, R.W., Maksimovic, D.: ‘Resonant conversion’, in ‘Fundamentals of
power electronics’ (Kluwer Academic, USA, 2004, 2nd edn.), pp. 718–719
[16] Muharam, A., Mostafa, T.M., Nugroho, A., et al.: ‘Power loss analysis of
inverter with and without transformer for capacitive power transfer’. 2018
IEEE Wireless Power Transfer Conf. (WPTC), Montreal, QC, Canada, 2018,
pp. 1–4
[17] Louie, H.: Seattle University, Washington, USA. Non-Ideal Single-Phase
Transformers. Presented at Electromechanical Energy Conversion class,
available at: http://www.drhenrylouie.com/Courses.html
[18] ‘Ferrite Cores’, available at: https://www.mag-inc.com/Media/Magnetics/File-
Library/Product%20Literature/Ferrite%20Literature/Magnetics-Ferrite-
Catalog-2017.pdf?ext=.pdf, accessed March 2019
[19] ‘Wire Table’, available at: https://www.mag-inc.com/getattachment/Design/
Design-Guides/Inductor-Design-with-Magnetics-Powder-Cores/
WireTable2017.pdf?lang=en-US, accessed March 2019
[20] ‘Litz Design’, available at: http://www.litzwire.com/litz_design.htm, accessed
March 2019
[21] ‘Keysight Technologies’, available at: https://www.keysight.com/main/
editorial.jspx?
ckey=1356999&id=1356999&nid=-11143.0.00&lc=eng&cc=SA, accessed
March 2019
[22] ‘Power Electronics I’, available at: http://www.engr.colostate.edu/ECE562/
course_info.html, accessed March 2019
Fig. 22 Experimental waveforms of the voltage across the capacitive
interface (Cc1) and the output voltage 7 Appendix A: transformer design guide

easier tuning and less sensitivity to the parameters variation can be


achieved. Moreover, the voltage stress over the capacitive interface The following steps were used to design the transformers used in
was reduced by the turns ratio factor with the same amount of this research, which can be applied for any case, which requires
power delivered to the load. It has been found that the leakage transformer design. First, the amount of power that will be
inductance of the non-ideal transformer can be used to tune the delivered to that load, and the operating frequency should be
circuit partially or fully at the resonance frequency so that the decided. Accordingly, the core material can be chosen. In [18],
tuning inductor of the CPT system can be eliminated or greatly detailed specs of different cores are listed. According to the
reduced. The mathematical analysis has been verified by practical maximum voltage and current that will be applied to the primary
experiments, and the prototype CPT system has demonstrated an and the secondary sides plus the operating frequency, the suitable
output power of 25 W at an operation frequency of 1 MHz at a winding wires type and gauge can be selected. In [19, 20] specs of
power efficiency >70%. The voltage stress across the single pair of different wires can be found. The size of the core should be
the coupling plates is reduced to 50.4 V at a turns ratio of 5, selected to fit the needed number of windings and turns ratio. After
compared to 252 V for the conventional system without using a that, the transformer can be built with the desired turns ratio.
step-down transformer. In the experimental section, the transformers were built from a
toroidal core which has an outside diameter of 41.8 mm, an inner
diameter of 26.2 mm with 12.5 mm thickness (forming 12.5 × 12.5
6 References mm cross-section). The core itself is made of N30 material with µr
[1] Tesla, N.: ‘Electrical transformer or induction device’, US Patent No. = 4300. The windings consist of two twisted wire with #18 AWG
433,702, August 5, 1890
[2] Liu, C., Hu, A.P.: ‘Wireless/contactless power transfer-capacitively coupled
size. One side was wound with 20 turns while four turns were used
solutions’, LAP, 2012 four other sides, so it could give a step-up and step-down
[3] Culurciello, E., Andreou, A.G.: ‘Capacitive inter-chip data and power transfer transformers with 0.2 and 5 turns ratio, respectively.
for 3-D VLSI’, IEEE Trans. Circuits Syst. II, Express Briefs, 2006, 53, (12), The two sides windings are overlapping on one toroidal core as
pp. 1348–1352
[4] Salzman, D., Knight, T., Franzon, P.: ‘Application of capacitive coupling to
shown in Fig. 23.
switch fabrics’. Proc. 1995 IEEE Multi-Chip Module Conf. (MCMC-95), Leakage inductance Llk can be measured by using the
Santa Cruz, CA, USA, 1995, pp. 195–199 impedance analyser as explained in [21]. This method was used in
[5] Wang, K., Sanders, S.: ‘Contactless USB — a capacitive power and this manuscript.
bidirectional data transfer system’. 2014 IEEE Applied Power Electronics
Conf. and Exposition – APEC 2014, Fort Worth, TX, 2014, pp. 1342–1347 Llk can also be estimated for a step-down transformer, for
[6] Mostafa, T.M., Muharam, A., Hattori, R.: ‘Wireless battery charging system example, by using the following equations [22]
for drones via capacitive power transfer’. 2017 IEEE PELS Workshop on

2640 IET Power Electron., 2019, Vol. 12 Iss. 10, pp. 2634-2641
© The Institution of Engineering and Technology 2019
The secondary side value of leakage inductance is
Ls = n2 × Lp = 0.016 μH where x% is the percentage of the
flux, which follows a leakage path.

np2
Lm =

ℜ is the core reluctance which can be obtained by


Fig. 23 Transformer structure
l
Total leakage flux of x%, suggests x%/2 of magnetising ℜ=
Aμo μr
inductance Lm for each leakage coil. Therefore, the leakage
inductance in the primary coil is
where l is the circumference of the core, A is the cross-sectional
area of the core, μr is the permeability of core material, and μo is
Lp = x% × Lm
the permeability of vacuum (4π e−7 H/m).

IET Power Electron., 2019, Vol. 12 Iss. 10, pp. 2634-2641 2641
© The Institution of Engineering and Technology 2019

You might also like