Design of a Contactless Energy-Transfer System for Desktop Peripherals

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 47, NO. 4, JULY/AUGUST 2011

1643

Design of a Contactless Energy-Transfer System for Desktop Peripherals Pascal Meyer, Paolo Germano, Miroslav Markovic, and Yves Perriard, Senior Member, IEEE

Abstract—This paper presents an approach to design a contactless energy-transfer system to supply desktop computer peripherals. Energy is transferred using coreless transformers from an array of primary coils (placed on a table) to one or several secondary coils (placed under the peripheral). Knowing the geometrical structure of the coils of a coreless transformer, accurate calculations allow predicting the power that can be transferred to a load. By applying these calculations to one basic energy-transfer cell, it is then possible to design a whole table that is able to supply many consumers at any location. The main problem with such systems is that the energy transfer is not constant along the surface and may not be even enough to supply an electronic device. This paper proposes several solutions to solve this problem. Index Terms—Contactless, transfer.

coreless

transformer,

energy

I. I NTRODUCTION

M

ANY INDUSTRIAL applications are based on contactless energy transfer through inductive coupling [1], [2]. Such applications often require solving complicated problems such as the position detection and the activation of the corresponding primary coils. A simple solution to transfer energy to a few loads is developed in [3], where the magnetic-field intensity is almost uniform due to the use of two primary coils with different sizes. In [4], four rectangular printed circuit board (PCB) primary coils allow supplying one or a few loads with a detection system composed of smaller integrated coils. An array of square primary coils is a very common structure to generate a magnetic flux. It can be used to supply a moving load [5], [6], where the secondary coil has the size of about four primary coils and the primary side is composed of nine coils. To ensure the desired amount of power transfer, the voltage supply is continuously adjusted during movement. An optimization of a similar system with magnets integrated in its structure is proposed in [7] in order to supply a contactless planar actuator. For this application, the coils are composed of Litz wires, Manuscript received September 30, 2010; revised January 10, 2011; accepted February 20, 2011. Date of publication May 12, 2011; date of current version July 20, 2011. Paper 2010-EMC-361.R1, presented at the 2010 IEEE Energy Conversion Congress and Exposition, Atlanta, GA, September 12–16, and approved for publication in the IEEE T RANSACTIONS ON I NDUSTRY A PPLICATIONS by the Electric Machines Committee of the IEEE Industry Applications Society. The authors are with the Integrated Actuators Laboratory (LAI), Ecole Polytechnique Federale de Lausanne, 1015 Lausanne, Switzerland (e-mail: [email protected]; [email protected]; [email protected]; [email protected]). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TIA.2011.2153812

and their number is drastically increased in both primary and secondary sides. In order to generate an uniform flux density over the whole surface, a three-layer array of hexagonal PCB windings is simulated and measured in [8] and [9]. In [10], a study on different array structures is shown. Two shapes (square and hexagonal) are taken into account, and multilayer arrangements are tested. Simulations show interesting results, but experimental measurements remain quite far from simulations. A contactless energy transfer that uses an array of hexagonal coils is proposed in [11]. Clusters of three primary coils are excited by a current with the same amplitude but with a phase shift of 120◦ . That allows reducing the stray magnetic field. It is also shown that secondary coil topologies with multiple small coils are better suited than a single large coil. However, for this application, the maximum power efficiency of 13% is really low. In [12], the efficiency is improved, but no numerical value is provided. In [13], a desk with a cord-free power supply is presented. The primary is composed of an array of small circular coils. The energy is transferred from multiple primary coils to a larger secondary one. The tested system provides high energy-transfer efficiency of 85% ± 5%, depending on the location of the secondary. In [14], a single-layer array structure of primary coils is presented. A study on the best arrangement of shapes for primary and secondary coils is performed in order to ensure a good efficiency, whatever the position of the device. As a result, this paper states that the best solution is a hexagonal packing of circular primary windings coiled around a ferrite core and a larger secondary circular spiral coil designed so that it covers at least one primary coil wherever it is placed on the surface. The system is able to detect the device and localize the charging flux within the covered area of the secondary coil. Practical measurements show an efficiency in the range of 86%–89%. Such a system is very convenient because it needs only one primary layer. However, it requires a threedimensional structure that cannot be realized with PCB coils. The original contributions of this paper is to present an approach to design a coil array for desktop computer peripherals. The first step consists of a study that allows determining the required shape and topology of the primary and secondary coils. Then, a basic cell is designed to fulfill the required specifications in the most operating cases. An analytical model has been implemented in order to compute inductances and electric magnitudes for a given geometry. Finally, the primary coil of the basic cell is replicated to form a full array.

0093-9994/$26.00 © 2011 IEEE

1644

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 47, NO. 4, JULY/AUGUST 2011

TABLE I E LECTRIC S PECIFICATIONS OF THE M OUSE , K EYBOARD , AND L OUDSPEAKER

Fig. 2.

Mutual inductance between two loops.

Fig. 3.

General circuit of a coreless transformer.

Fig. 1. Parameters of the primary coil.

II. S YSTEM S PECIFICATIONS The required energy-transfer system consists of a table with arrays of PCB primary coils integrated into it and desktop peripherals with PCB secondary coils. It must be able to supply three computer devices, i.e., a mouse, a keyboard, and a loudspeaker. The energy must be continuously transferred to the three peripherals, whatever their position. Their electric specifications have been measured and are reported in Table I. For the application presented in this paper, the geometrical design is constrained by the size of the mouse. Furthermore, the whole induction table should not be constantly supplied but only the primary coils where a device is placed. That is why the system has to be able to detect the presence of a device and to activate only the coil(s) situated under the device. This paper presents an approach that is able to provide the design that ensures the sufficient amount of energy for any peripheral position.

III. A NALYTICAL M ODEL A. Generalities According to [10], square and hexagonal shapes are best suited for coil array structures. In this paper, the study is focused on the square ones because the amount of coils needed for a given surface is smaller and, therefore, the system requires less electronic components. Furthermore, it is shown in [10] that three primary layers are needed for hexagonal coils and only two for square ones to generate an uniform flux density. As a result, the square form is chosen for primary and secondary coils. The geometric parameters of a coil, shown in Fig. 1, are the number of turns n1,2 , the numbers of layers N1,2 , the track width W1,2 and the copper thickness T1,2 , the distance between two turns D1,2 , and the external length of the coil a1,2 . The vertical distance H between primary and secondary coils is fixed at 5 mm. In this paper, the insulator of the PCB is in epoxy (FR-4) with a theoretical thickness of 0.2 mm. A maximal

thickness of 0.45 mm is measured, where there are two layers of copper. B. Inductances Calculation For coreless transformers that are composed only of one primary coil and one secondary coil, an analytical model is generated to accurately calculate the self-inductance and the mutual inductance. It is based on a discretization of the Neumann formula. For convenience, the square spiral coils are assumed to be closed loops with the same number of turns. For calculating the mutual inductance between two closed loops (see Fig. 2), the Neumann formula is given by M=

μ0   Δl2 · Δl1 . 4π R

(1)

C2 C1

The formula for mutual inductance between two coils of more than one turn is obtained by summing all the mutual inductances between pairs, i.e., L12 =

n1  n2 

Mij

(2)

i=1 j=1

where L12 is the mutual inductance between the primary and secondary coils and n1 and n2 are the number of primary and secondary turns. For the self-inductance (L11 and L22 ) computation, the model follows a similar procedure. C. Resistance Calculation The energy transfer described in this paper is planned to work around 200 kHz. For this frequency, the influence of the skin effect is insignificant when using PCB windings [15], [16].

MEYER et al.: DESIGN OF A CONTACTLESS ENERGY-TRANSFER SYSTEM FOR DESKTOP PERIPHERALS

Fig. 4.

1645

Flow diagram of the design methodology.

Therefore, the resistance of the coils are calculated with the well-known formula for the resistance, i.e., R=

ρl S

(3)

where ρ is the resistivity of the copper (ρ = 17 nΩm), l is the length of the conductor, and S is the cross-sectional area of the conductor. D. Electrical Circuit The electrical parameters of the coreless transformer are given in Fig. 3. For the application developed in this paper, the connection of the resonance capacitors is chosen in series (C2 ) for the secondary coil and in parallel (C1 ) for the primary coil [2], [17]. To compute the power transferred to the load, the following voltage equations in the primary and secondary coils must be considered: U 1 = R1 I L1 + jωL11 I L1 − jωL12 I 2 jωL12 I L1 = (R2 + RL )I 2 + jωL2 I 2 +

1 I . jωC2 2

(4) (5)

Here, IL1 is the current in the primary coil and I1 (see Fig. 3) is the current delivered by the voltage supply. By resolving the electrical circuit as proposed in [5], it is possible to compute the secondary current, i.e., I 2 = I L1 ·

jωL12 . R2 + RL

(6)

From this point, it is not difficult to determine the voltage and the power transferred to the load. IV. C ORELESS T RANSFORMER D ESIGN A. Iterative Design Process For this paper, the parameters described in Section III are defined through an iterative process so that they fulfill the requirements of the mouse in terms of geometry and of the loudspeaker in terms of the power transferred. The approach to design the system is as follows (see Fig. 4). 1) The study of the specifications allows determining some parameters. The size of the mouse limits the secondary

coil to a surface of 40 mm× 40 mm. Furthermore, the primary and secondary coils must have similar dimensions in order to increase their magnetic coupling. The thickness of copper tracks T is fixed to 105 μm. 2) The numbers of turns n1 and n2 are chosen (randomly in the first iteration). The track widths W1,2 and the track spacings D1,2 largely depend on this choice but are also limited by the heat dissipation due to Joule losses. 3) The analytical model is applied to the set of geometric parameters in order to calculate the self-inductances L11 and L22 and the mutual inductance L12 . 4) By fixing the primary current or voltage, it is possible to compute the secondary voltage and current for a given load, due to the method presented in Section III. 5) At this point, two conditions must be checked, i.e., the heat dissipation and the power transferred to the load. The evaluation criterion for the first one is a threshold of the heat generation power. For the second one, the criterion is a threshold on the minimum power transferred to the load. If one of this conditions is not fulfilled, a return to point 2 is necessary in order to generate a new set of parameters. For choosing the maximal Joule losses Pj for the coils, a simple thermal model is implemented, taking into account only the convection. The increase in temperature of the coils is evaluated by ΔTemp =

Pj RI 2 = . αAα αAα

(7)

Here, α is the convection coefficient of surface Aα of the copper. We choose α = 9 W/m2 K according to [18]. Experimental tests showed that a coil temperature of more than 80 ◦ C caused no damage to the PCB and the surrounding, but for the user’s convenience, the maximal allowed temperature is fixed at 60 ◦ C. This choice leads to maximal Joule losses of 0.5 W for each coil. Concerning the condition on the output voltage URL , it must be chosen according to the application. In this case, URL,min was fixed at 6 V [root mean square (RMS)], which ensures the working of the loudspeaker with a good margin. With the two conditions well defined, the iterative process can be applied for the application presented in this paper. It is not a classical optimization since the process is stopped when the specifications are met. The values obtained at the end of this process are given in Table II.

1646

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 47, NO. 4, JULY/AUGUST 2011

TABLE II F INAL VALUES OF THE C OIL PARAMETERS

Fig. 6. Voltage induced on a load of 5 Ω as a function of the number of primary and secondary turns.

Fig. 5. Computed mutual inductance as a function of the number of primary and secondary turns.

B. Parametric Analysis Using the analytical model, it is possible to perform a parametric analysis of the coreless transformers. The benefit of such a study is double; it gives a good understanding of the impact of the parameters on the coreless transformer, and it provides the tools to modify the geometrical design toward the desired solution during the iterative design process. For this application, the most relevant parameters to be analyzed in this paper are the number of primary turns n1 and secondary turns n2 . The final design obtained for the primary and secondary coils described in this paper are used as the reference for this parameters analysis. The final design has 18 primary turns and 20 secondary turns, both distributed on two layers of copper. Fig. 5 shows the mutual inductance L12 when the number of primary turns is varied from 2 to 18 (n1 = 2, 4, . . . , 18) and the number of secondary turns from 2 to 20 (n2 = 2, 4, . . . , 20). The geometrical configuration of the coils is such that the first turns are placed on the external dimensions of the coils. The turns are then added toward the center of the coils by pairs of two superimposed turns. As expected, when the number of primary and secondary turns is increased, the mutual inductance is also increased. Fig. 6 shows the calculated voltage induced on a 5-Ω load as a function of n1 and n2 . The input voltage is U1 = 12 [alternating-current (ac) RMS] at 200 kHz, the secondary capacitor C2 is chosen so that it cancels the secondary inductance L22 , and the primary capacitor C1 is chosen so that the primary voltage U1 is in phase with the primary current I1 , which corresponds to the case where there are no reactive losses. The result shows that, when n2 is increased, the output voltage URL is also increased. On the contrary, increasing n1 leads to an output voltage drop. The reason is that the impedance of the

Fig. 7. Efficiency of the coreless transformer as a function of the number of primary and secondary turns.

primary coil, given by Z1 = R1 + jωL11 , becomes higher with the number of turns and therefore limits current IL1 flowing through it. From (6), it is clear that it leads to a decrease in the secondary current I2 . This effect dominates the increase in the mutual inductance. Fig. 7 shows the efficiency of the coreless transformer. It is given by the ratio of the primary power and the power of the load as follows: η=

URL I2 U1 I1 cos(φ1 )

(8)

where φ1 is the angle between the primary voltage U1 and the primary current I1 . With the correct value of C1 , there is no phase shift between U1 and I1 so that cos(φ1 ) = 1. One can note here that only the efficiency of the coreless transformer itself is considered. The losses in the rectifier are not taken into account. As a result, the efficiency increases if n1 and

MEYER et al.: DESIGN OF A CONTACTLESS ENERGY-TRANSFER SYSTEM FOR DESKTOP PERIPHERALS

Fig. 8.

Photograph of the prototype.

n2 increase and values over 90% are obtained for six turns on primary and secondary coils. As a conclusion for parameters n1 and n2 , a few general statements may be formulated. First, the number of secondary turns n2 has an essential influence on the induced voltage. Increasing n2 leads to an increase in the induced voltage. However, the number of primary turns n1 has a major impact on the current flowing in the primary coil IL1 , so that increasing n1 leads to a decrease in the induced voltage even if the mutual inductance is in the same time increased. Finally, a high number of primary and secondary turns becomes more interesting in terms of efficiency. These few general statements have been also observed for other coreless transformers, which are different in size and topology. They are largely used in the iterative process presented in the previous subsection and show how the design may be changed between each step.

1647

Fig. 9. Electronic circuit of the coreless transformer. TABLE III C OMPUTATIONS OF I NDUCTANCES AND R ESISTANCES C OMPARED W ITH E XPERIMENTAL M EASUREMENTS

V. P ROTOTYPE Two small prototypes have been built for the measurements presented in Section VI. The first one consists of a small array of nine primary coils distributed on one layer. The design of the coils is issued from the result obtained by the iterative process presented at the end of Section IV-A. The second prototype consists of an array of 18 primary coils distributed on two shifted layers, again with the same design for each coil. A photograph of the setup that was used is shown in Fig. 8, with the two-layer prototype and a full bridge converter (see Fig. 9) that was developed for this project. For supplying only small consumers such as a mouse, a keyboard or a speaker, a halfbridge or a one-transistor bridge would be enough. However, for the prototyping, the full bridge generates clearer electronic signals, which is more convenient to provide accurate measurements. The full bridge is composed of four metal–oxide– semiconductor field-effect IRFZ44 transistors. The value of the primary and secondary capacitors are C1 = 55 nF and C2 = 47 nF, respectively. The frequency is adjusted to 220 kHz so that the secondary inductance is better compensated. The load resistance RL shown in Fig. 3 represents the electronic device to be supplied. The output voltage URL has the same frequency as the current in the primary coil. It is clear that, for the mouse, the keyboard, and the loudspeaker, the secondary side must

Fig. 10. Output voltage in millivolts (ac RMS).

integrate a rectifier with a voltage regulator. The rectifiers are simply realized with four diodes and a smoothing capacitor. The voltage regulators are realized with LM2940-5 for the loudspeaker and two of LM317 for the mouse and the keyboard. In conclusion, RL must be regarded as an equivalent load that includes the rectifier, the regulator, and the device itself. From many theoretical and measured examples, we deliberately choose RL = 5 Ω because it represents a bigger consumer than the three devices presented in Table I. Therefore, if it is possible to show that the system works for this value of the load, it is obligatory the case for bigger values of resistance. VI. E XPERIMENTAL M EASUREMENTS A. Inductance and Resistance First of all, computed values of self-inductance and mutual inductance are compared with experimental measurements in Table III. The self-inductances and resistances are measured with a HP Precision LCR Meter 4284A (20Hz–1MHz) at 200 kHz. The connecting wires are as short as possible.

1648

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 47, NO. 4, JULY/AUGUST 2011

Fig. 11. (a) Primary coil of 18 turns. (b) Output voltage in millivolts (ac RMS).

Fig. 12. (a) Array of nine coils. (b) Output voltage in millivolts (ac RMS) for one layer of primary coils. The nine coils are supplied, and the measurements are performed on a portion of the platform due to its geometrical periodicity.

For the mutual inductances, two measurements are needed. First, the two coils of the coreless transformer are connected in series, which gives the value of inductance LA = L11 + L22 + 2L12 . Second, the connection pins of one coil are inverted, which gives the measurement of inductance LB = L11 + L22 − 2L12 . The mutual inductance can be then deduced with L12 = (1/4)(LA − LB ). B. Output Voltage Knowing the magnetic parameters of a coreless transformer, it is possible to compute the voltage of a given load as a function of its position. Fig. 10 shows a comparison between the predicted induced voltage and measurements, and Fig. 11 shows the voltage measured in the secondary coil when it is loaded by a resistance of 5 Ω displaced beside the primary coil. The position of the secondary coil is given by its center and is moved along x- and y-axes with measured points every 5 mm. A current of 1 A (ac RMS) is generated in the primary coil by a full bridge converter (see Fig. 9) working at about 220 kHz. The computations for multiple primary and secondary coils are too long if good accuracy is desired. As a first set of measurements, the primary coil presented in Table II is replicated in an array of nine coils [see Fig. 12(a)]. The experimental conditions are the same as previously mentioned, i.e., a 1-A (ac RMS)

current generated in each primary coil. Because of the repetitive coil pattern, the measurements are realized only on four coils, while the nine coils are supplied. An interesting result is that the behavior of multiple primary coils is very similar to the case of only one coil but with magnitudes of the measured voltage a bit smaller, due to the magnetic interactions between the primary coils. It can be also seen that the output voltage is critically low (330 mV) when the secondary coil center is located above the corners of primary coils, and that prevents the electronic consumers from being supplied at these positions. It is possible to solve this problem by superimposing a second layer of coils so that the corners of four adjacent coils are placed above the center of a coil of the other layer. Fig. 13 shows the measured voltage at the secondary coil for such a two-layer structure. As a result, the minimal voltage is assured at every location. Using two layers of primary coils implies that one layer is more distant from the secondary coil, so that it should have less effect on the induced voltage. However, we did not measure a significant difference between the two primary layers because the thickness of the two PCB layers is less than 0.9 mm, which is relatively small in comparison with the distance of the secondary coil (H = 5 mm). Finally, the negative values obtained in Figs. 10–13 must be commented. Indeed, it is absurd to speak about negative values for ac RMS voltages. However, we decided to artificially add

MEYER et al.: DESIGN OF A CONTACTLESS ENERGY-TRANSFER SYSTEM FOR DESKTOP PERIPHERALS

1649

Fig. 13. (a) Array of 18 coils. (b) Output voltage in millivolts (ac RMS) for two layers of primary coils. The 18 coils are supplied, and the measurements are performed on a portion of the platform due to its geometrical periodicity.

the negative sign because this makes the figures more readable. This is done for positions where the magnetic-field lines are closed, which gives rise to an inversion of the phase in the ac voltage. It is not possible to avoid this effect, whatever the number of supplied coils. C. Efficiency and Temperature For the measurements of efficiency, we decided to supply only one primary coil. The reason is that we selectively plan and supply the primary coils where the device is placed. Therefore, it would make no sense to measure the efficiency of the system when many coils are simultaneously supplied. Unfortunately, we were not able to measure the efficiency directly on the devices; thus, we decided to use a variable electronic load (HP 6060A) connected to the rectifier. The power of the load is evaluated by measuring its dc voltage and current, and the input power is evaluated by measuring the dc voltage and the current one generated by the voltage supply. Using this approach, the losses in the transistors and in the rectifier are taken into account. Two sets of measurements are proposed here. First, in Fig. 14, the efficiency is plotted as a function of the resistance, when the primary and secondary coils are perfectly aligned. The best efficiency (67%) is obtained with load values between 10 and 15 Ω, which corresponds to the loudspeaker (see Table I). This result was expected because the value of C1 = 55 nF has been chosen for this specific load since the loudspeaker is the most consuming device. The efficiencies obtained for the mouse and the keyboard are 62% and 44%, respectively. However, these values are not the total ones because no voltage regulator was used. Finally, we estimate the efficiency of about 30% for the mouse and the keyboard with the voltage regulator. Fig. 15 shows the efficiency measured when the secondary coil is moved along x-axis with a load of 12 Ω, which corresponds to the loudspeaker. The displacement is zero when the two coils are perfectly aligned. The efficiency becomes lower than 30% when the secondary coil is displaced at 20 mm from the aligned position. This corresponds to the transient position where the adjacent primary coil should have the same contribu-

Fig. 14. Measurement of the efficiency as a function of the load.

tion. Even if this value is relatively low, the necessary voltage to supply the loudspeaker remains sufficient as two primary coils situated under the device are activated, as shown in Section VI-B. When using relatively small consumers such as the three devices proposed in this paper, it is not possible to obtain very good efficiencies because the losses induced by the electronic circuits have a significant impact. For small consumers, a contactless energy transfer providing an efficiency of 67% is satisfying, and we estimate that efficiencies between 20% and 50% are acceptable since the system is properly working and the involved power is really low. For example, the power needed by the mouse is about 0.06 W so that having an efficiency of less than 30% is not dramatic. Finally, the temperature is not a critical matter for the application presented in this paper. We measured the temperature with a probe directly in contact with the copper tracks of the coils. For a typical use that involves currents in the order of 1 A (ac RMS) in primary and secondary coils, the measured temperatures are situated between 35 ◦ C and 40 ◦ C.

1650

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 47, NO. 4, JULY/AUGUST 2011

Fig. 15. Measurement of the efficiency as a function of lateral displacement.

VII. C ONCLUSION Due to the approach presented in this paper, we have been able to supply the mouse, the keyboard, and the loudspeaker at every positions above a prototype composed of two layers of nine primary coils. Another possibility to ensure the necessary level of the output voltage in the secondary coils consists in using multiple secondary coils with only one layer of primary coils. With two secondary coils shifted in a similar manner, the results are equivalent when the primary and secondary coils remain parallel. However, because of the eventual rotation of the secondary coils, a third one is required to transfer enough energy for every possible configurations. R EFERENCES [1] S. Tang, S. Hui, and H.-H. Chung, “Characterization of coreless printed circuit board (PCB) transformers,” IEEE Trans. Power Electron., vol. 15, no. 6, pp. 1275–1282, Nov. 2000. [2] C.-S. Wang, O. Stielau, and G. Covic, “Design considerations for a contactless electric vehicle battery charger,” IEEE Trans. Ind. Electron., vol. 52, no. 5, pp. 1308–1314, Oct. 2005. [3] X. Liu and S. Hui, “Optimal design of a hybrid winding structure for planar contactless battery charging platform,” IEEE Trans. Power Electron., vol. 23, no. 1, pp. 455–463, Jan. 2008. [4] F. Sato, J. Murakami, H. Matsuki, S. Kikuchi, K. Harakawa, and T. Satoh, “Stable energy transmission to moving loads utilizing new CLPS,” IEEE Trans. Magn., vol. 32, no. 5, pp. 5034–5036, Sep. 1996. [5] J. de Boeij, E. Lomonova, and A. Vandenput, “Contactless energy transfer to a moving load part I: Topology synthesis and fem simulation,” in Proc. IEEE Int. Symp. Ind. Electron., Jul. 2006, vol. 2, pp. 739–744. [6] J. de Boeij, E. Lomonova, J. Duarte, and A. Vandenput, “Contactless energy transfer to a moving load part II: Simulation of electrical and mechanical transient,” in Proc. IEEE Int. Symp. Ind. Electron., Jul. 2006, vol. 2, pp. 745–750. [7] J. de Boeij, E. A. Lomonova, and A. J. A. Vandenput, “Optimization of contactless planar actuator with manipulator,” IEEE Trans. Magn., vol. 44, no. 6, pp. 1118–1121, Jun. 2008. [8] S. Hui and W. Ho, “A new generation of universal contactless battery charging platform for portable consumer electronic equipment,” IEEE Trans. Power Electron., vol. 20, no. 3, pp. 620–627, May 2005. [9] X. Liu and S. Hui, “Simulation study and experimental verification of a universal contactless battery charging platform with localized charging features,” IEEE Trans. Power Electron., vol. 22, no. 6, pp. 2202–2210, Nov. 2007.

[10] J. Achterberg, E. A. Lomonova, and J. de Boeij, “Coil array structures compared for contactless battery charging platform,” IEEE Trans. Magn., vol. 44, no. 5, pp. 617–622, May 2008. [11] C. Sonntag, E. Lomonova, J. Duarte, and A. Vandenput, “Specialized receivers for three-phase contactless energy transfer desktop applications,” in Proc. Eur. Conf. Power Electron. Appl., Sep. 2–5, 2007, pp. 1–11. [12] C. Sonntag, E. Lomonova, and J. Duarte, “Variable-phase contactless energy transfer desktop part I: Design,” in Proc. Int. Conf. Elect. Mach. Syst., Oct. 2008, pp. 4460–4465. [13] K. Hatanaka, F. Sato, H. Matsuki, S. Kikuchi, J. Murakami, M. Kawase, and T. Satoh, “Power transmission of a desk with a cord-free power supply,” IEEE Trans. Magn., vol. 38, no. 5, pp. 3329–3331, Sep. 2002. [14] W. Zhong, X. Liu, and S. Hui, “Analysis on a single-layer winding array structure for contactless battery charging systems with freepositioning and localized charging features,” in Proc. IEEE ECCE, 2010, pp. 658–665. [15] C. Fernandez, R. Prieto, O. Garcia, P. Herranz, J. Cobos, and J. Uceda, “Modelling core-less high frequency transformers using finite element analysis,” in Proc. IEEE 33rd Annu. PESC, 2002, vol. 3, pp. 1260–1265. [16] C. Fernandez, O. Garcia, R. Prieto, J. Cobos, S. Gabriels, and G. Van Der Borght, “Design issues of a core-less transformer for a contact-less application,” in Proc. 17th Annu. IEEE APEC, 2002, vol. 1, pp. 339–345. [17] O. Stielau and G. Covic, “Design of loosely coupled inductive power transfer systems,” in Proc. Int. Conf. PowerCon, 2000, vol. 1, pp. 85–90. [18] M. Markovic, L. Saunders, and Y. Perriard, “Determination of the thermal convection coefficient for a small electric motor,” in Proc. 41st IEEE IAS Annu. Meeting, 2006, vol. 1, pp. 58–61.

Pascal Meyer was born in Switzerland, in 1984. He received the M.S. degree in microtechnology engineering in 2008 from the Ecole Polytechnique Federale de Lausanne, Lausanne, Switzerland, where he is currently working toward the Ph.D. degree, with the thesis on contactless energy-transfer systems for different types of peripherals.

Paolo Germano was born in Lausanne, Switzerland, in 1966, and is a native of Italy and Switzerland. He received the M.Sc. degree in microengineering from the Ecole Polytechnique Federale de Lausanne (EPFL), Lausanne, Switzerland, in 1990. Since 1990, he has been with the Laboratory of Electromechanics and Electrical Machines (LEME), EPFL, where he is currently a Senior Researcher with the Laboratory of Integrated Actuators (LAI). His activities as the Project Leader are mainly focused on test benches dedicated to the measurement and the analysis of stepping motors, brushless direct-current motors, linear actuators, and other devices (tactile watch). In the domain of inductive energy transmission, he took part in projects in the biomedical field (implanted device supply and transcutaneous stimulation), in the field of electrical vehicle (direct supply, battery recharge stations, and dynamic recharge) and in various projects related to machine tools and wireless computer peripheral supplies.

MEYER et al.: DESIGN OF A CONTACTLESS ENERGY-TRANSFER SYSTEM FOR DESKTOP PERIPHERALS

Miroslav Markovic received the M.Sc. degree from the University of Belgrade, Serbia, in 1996 and the Ph.D. degree from the Ecole Polytechnique Federale de Lausanne (EPFL), Lausanne, Switzerland, in 2004. From 2000 to 2004, he was a Research/Teaching Assistant with the Integrated Actuators Laboratory (LAI), EPFL, where he has been a Project Leader since 2004. His main research area is the design optimization of electric motors and electromechanical actuators.

1651

Yves Perriard (SM’10) was born in Lausanne, Switzerland, in 1965. He received the M.Sc. and Ph.D. degrees in microengineering from the Ecole Polytechnique Federale de Lausanne (EPFL), Lausanne, Switzerland, in 1989 and 1992. He is a Cofounder and was the Chief Executive Officer of Micro-Beam SA, Yverdon, Switzerland, which is a company involved in high-precision electric drive. He was a Senior Lecturer in 1998 and has been a Professor since 2003 with the EPFL, where he is currently the Director of the Integrated Actuator Laboratory (LAI) and the Vice Director of the Microengineering Institute. His research interests are in the field of new actuator design and associated electronic devices. He has authored or coauthored more than 100 publications and patents.