Design of Power Factor Correction Model Using PFC Boost Converter for Control Solid-State Lamps

7 International Journal for Modern Trends in Science and Technology

Design of Power Factor Correction Model Using PFC Boost Converter for Control Solid-State Lamps

Dr Pagidimarri Krishna¹ | M.Aruna² | S. Prem Kumar³

1Professor, Department of EEE, Nalla Malla Reddy Engineering College, Hyderabad, Telangana, India.

2Assistant Professor, Department of EEE, Nalla Malla Reddy Engineering College, Hyderabad, Telangana, India.

3Sr.Assistant Professor, Dept. of EEE, Lakkireddy Balireddy Engineering College, Mylavaram, Andhra Pradesh. India.

To Cite this Article

Dr Pagidimarri Krishna, M.Aruna and S. Prem Kumar, “Design of Power Factor Correction Model Using PFC Boost Converter for Control Solid-State Lamps”, International Journal for Modern Trends in Science and Technology, Vol. 04, Issue 03, March 2018, pp.-07-16.

This paper describes a multiphase pulse width modulation (MPWM) technique used for controlling the luminance of a solid-state lamp (SSL). This letter describes a digital Pulsewidth modulator that enables the generation of a large, adjustable number of Pulsewidth modulated (PWM) outputs of programmable duty cycle and dead time, yet requiring just a small, fixed architecture. In conventional pulse width modulation (PWM) and pulse code modulation techniques, the average HBLED string current is controlled by simultaneously varying the "ON" time duration of each current regulator circuit, resulting in large current transients and pulsating light output. The proposed MPWM technique operates by uniformly time-shifting the individual ON/OFF control signal pulses, thus avoiding large current transients. When compared to conventional PWM dimming, MPWM dimming reduces the risk of visible flicker and lowers the magnitude of audible noise. The reduced magnitude of output current transients results in lower electromagnetic interference and enables optimization of the size of passive components. One promising technology which has the potential for use in specialized energy-efficiency lighting applications is solid-state lamp technology. There are many reasons solid-state lamps may be especially useful for new applications. In this paper, the operation of the MPWM technique is explained in detail. Finally a flyback converter topology with input side power factor correction topology is proposed and simulation results are presented. The simulation results are obtained using MATLAB/SIMULINK software.

Keywords—Digital control, LED dimming, LED lamp driver, multiphase Pulse width modulation, solid-state lamp luminance control.

Copyright © 2018 International Journal for Modern Trends in Science and Technology All rights reserved.

I. INTRODUCTION

The term solid-state lamp refer to light sources based on semi conductor materials III-V compounds such as gallium arsenide (GaAs) and gallium aluminum arsenide (GaAlAS), or II – VI compounds such as zinc selenide (ZnSc). There are several types:

1) Light emitting diodes.

2) Diode lasers

3) Super luminescent diodes 4) Electro luminescent diodes.

1) Light-emitting diodes:

The most common solid- state lamps are light-emitting diodes (LEDs) [9], [10]. They operate at tens of milliamps and a few volts, and have response times of roughly hundreds of nanoseconds. This makes LEDs compatible with integrated circuits and ideal for use as status ABSTRACT

Available online at: http://www.ijmtst.com/vol4issue3.html

International Journal for Modern Trends in Science and Technology

ISSN: 2455-3778 :: Volume: 04, Issue No: 03, March 2018

8 International Journal for Modern Trends in Science and Technology indicators. In illumination-type applications,

however, common LEDs display some drawbacks.

First, while the light is proportional to the current at low values, at higher values the intensity saturates, defining a threshold above which the process is inefficient. Second, an LED radiates brightly over only a narrow solid angle, a few tens of degrees. Third, single LEDs emit light over only a narrow spectral range. Fourth, LEDs based on technology from the 1970s are still widely available; their low-rated performance has led to poor perception of solid-state lamps in general.

However, recent brightness and efficiency gains have led to devices comparing favorably with the performance of some conventional lamp systems.

New manufacturing methods, materials and ultra-large scale processing signal a possibility for lower costs, contributing to a resurgence of interest in LEDs. An example of this is the replacement of conventional bulb-reflector systems in some automobile brake lamps with LEDs or laser-diffuser systems [11].

2) Diode lasers:

Under optical resonance conditions, solid-state light devices may emit coherent light [12], in which the output is extremely mono- chromatic, collimated and in-phase. These devices are called diode lasers, and are used in digital signal storage and retrieval, such as for optical discs, and in telecommunications. While the output of diode lasers is typically higher than in LEDs, the emitted light is coherent, and highly directional. In addition, the bulk of commercial laser diodes radiate in the infrared (IR).

3) Super luminescent diodes:

Recently, a light source with properties intermediate between laser diodes and LEDs [13], [14] has been demonstrated. These devices are called super luminescent diodes (SLDs).

Fabrication methods are similar to those for diode lasers; an added step “spoils” the process leading to laser action, yielding high- output high-efficiency devices with incoherent, narrowband emission.

SLDs were developed for sensor and printing applications, but their high power output and efficiency, and low coherence make them good candidates for use in lighting applications. SLDs, however, are not yet widely available commercially.

4) Electroluminescent devices:

Electroluminescent (EL) lamps consist of a thin-film or powder phosphor sandwiched between two electrodes. Under an applied electric field, the phosphor fluoresces, and light is emitted through the transparent front electrode .Most phosphor

systems involve a ZnS host and an orange activator, Mn. Alternatives include Ce and TbF3 for green, and Pr for white light. EL devices are found in computer display screens; recently, panels have been used as backlighting for exit signs. EL sources are relatively inexpensive, but are relatively inefficient and dim. Their luminance may be increased by raising the operating voltage, but the lifetime then drops drastically.

Solid-state lighting is now far beyond from being just a promising technology [1]–[4].

Phosphor-conversion (PC) LEDs have become the most efficient sources of white light [5] and most manufacturers of high-power LEDs have already surmounted the 100 lm/W efficiency milestone on industrial scale [6]–[8]. Beyond applications in signage, displays, and LCD backlighting, LEDs have become a competitive alternative to conventional sources of light in a vast sphere of general lighting.

The focus of this paper is on the development of an alternative luminance control technique for a SSL implementation, as shown in Fig. 1. The lamp architecture is commonly used in many applications that require a large number of HBLEDs to generate the desired light output, such as liquid crystal display TV backlights, streetlights, or commercial lighting fixtures [6]–[8]. In this architecture, the current through each parallel string consisting of series-connected HBLEDs is controlled by a series-current regulator circuit. For low-power HBLEDs, a linear current sink circuit is preferred because of its small area and low cost [9],[10]. Buck converter-based current regulator circuits are used for high-power HBLEDs to achieve higher system efficiency [11]. The complete array is powered by a single regulated dc–dc converter. For applications that are connected to a low-voltage supply, a boost dc–dc converter is generally used to step-up the input voltage Vg, to a value greater than the HBLED string forward voltage drop Vo.

9 International Journal for Modern Trends in Science and Technology

Fig. 1. Block diagram representation of a SSL consisting of a HBLED array with parallel strings of series-connected devices, series-current regulator circuits, a single dc–dc power converter,

and luminance control block.

The control signals generated by the corresponding modulator circuit for a lamp consisting of four parallel HBLED strings are shown in Fig. 2(a) and (b), respectively. All four control signals {c1, c2, c3, c4}are used to simultaneously turn the LED strings

“ON” and “OFF,” thus resulting in large current transients drawn from the dc– dc power converter, with sharp rise and fall times. The large current transients io(t), result in an increased value of output filter capacitance, increased conducted electromagnetic interference (EMI) and require more complex feedback control of the dc–dc power converter. Further, the range of PWM or PCM dimming frequency that can be implemented is limited by factors affecting the human perception of visible flicker and audible noise.

(a)

(b)

(c)

Fig. 2. Different pulse modulation techniques. (a) PWM. (b) PCM.

(c) MPWM.

II. MPWM DIMMING TECHNIQUE

The study of MPWM technique was motivated by the need to achieve dimming function, while reducing the output current transients and EMI generated by the power circuit. In this section, the operation of MPWM dimming for controlling the luminance of SSL is first explained, and then, implementation details are provided. The performance of SSL is discussed and some of the key advantages of MPWM dimming technique are highlighted.

A. Operation and Implementation of MPWM Dimming Technique:

The MPWM technique achieves lamp dimming by uniformly phase shifting, each PWM control signal {c1 , c2 , c3, . . . , cN }, such that at any given point in time only one string is turning “ON,” while another is turning “OFF,” as shown in Fig. 2(c). The “ON”

time tON of each individual PWM control signal c{i}, is modulated based on the input command Ddim, such that the duty cycle of each control signal dc{i}

is as follows:

(1) Where Tdim is the dimming period, 1/fdim. Even though the current through each individual HBLED string Is{i}, is pulsating, the uniform phase shift among all N string currents results in a net dc current drawn from the power supply, with a peak-to peak ripple limited in magnitude to less than or equal to the rated current flowing through one HBLED string, as shown in Fig. 3. For the case

10 International Journal for Modern Trends in Science and Technology when HBLED array is placed inside light

integrating lamp housing, covered with optically diffusing films, the same averaging effect is observed for the total lamp light output. The measured lamp luminance now has a dc component and the peak-to-peak ripple is limited in value corresponding to that of a single HBLED string output. Further, the magnitude of the peak-to-peak ripple current and luminance is independent of the number of HBLED strings present in the lamp (N>2), and the external-dimming command value, thus making it possible to scale the lamp output to meet the desired specifications. MPWM dimming results in linear relationship between the input-dimming command Ddim, and the lamp luminance Lv, as each control signal is still PWM, in a manner similar to conventional PWM dimming.

Analog and digital hardware implementation techniques for generating MPWM have been described in literature . A digital MPWM circuit based on digital-to-analog conversion (DAC) techniques offers key advantages over the analog counterpart in terms of scalability, power consumption, and the total circuit size and cost.

For digital implementation of MPWM, it is essential to ensure that the lamp luminance can be accurately varied between the maximum and minimum value in discrete number of steps given by the digital-dimming command. In order to achieve the desired resolution in luminance output, a digital modulator capable of accurately generating the control signal duty cycle dc{i} based on the input-dimming command Ddim, is required.

Thus, for a p-bit digital-dimming command Ddim, a modulator capable of generating a pulse with minimum “ON” time given by

(2)

should be used to achieve 2^p discrete luminance values. The area and the power consumption of the modulator circuit now scales with the desired resolution of the luminance output and can be optimized based on the dimming frequency fdim. In this paper, an area efficient digital modulator circuit described in [18] is used to implement a digital MPWM, due to its advantages in terms of size, power consumption, and cost.

Fig. 3. DC–DC power converter output current Io, and the individual HBLED string currents Is{i} when using conventional

PWM dimming (blue) and the proposed MPWM dimming (green) for a lamp consisting of four HBLED string and each drawing a

rated current of 100 mA. The input-dimming command Ddim=0.66 and dimming frequency fdim=1.5 kHz

B. Performance Evaluation of MPWM Dimming Technique:

The visual flicker sensation caused by light modulation at a frequency as high as 75 Hz, is the leading cause of dissatisfaction in home and office environments. Other health problems related to eye stress and headaches have been reported at modulation frequencies as high as 120 Hz. The human flicker sensation is reported to be a function of light modulation frequency and modulation amplitude. In case of PWM and MPWM dimming, the light modulation frequency is based on the choice of dimming frequency fdim, which determines the lamp optical performance and its impact on human visual perception. The visible flicker generated by the SSL can be minimized or eliminated by choosing the highest possible dimming frequency fdim. However, the response time tcr(tcr=trise +tfall) of the linear-current regulator circuits, the optical response time of the HBLEDs tLED, and the resolution of the input-dimming command Ddim, limit the maximum possible dimming frequency fdim_max. In practice, the current regulator response time tcr exceeds the HBLED optical time constant tLED(tcr >> tLED), and hence, limits the dimming frequency to a lower threshold is given by

(3) The lower limit on the dimming frequency fdim min, is based on the critical flicker frequency fcc , threshold which for home and office environment

11 International Journal for Modern Trends in Science and Technology is estimated to be approximately as fcc < 120 Hz.

Therefore, lamp operation above the minimum dimming frequency fdim min = 120 Hz, is essential to prevent visible light flicker. The use of MPWM technique also simplifies the light sensing network used in a feedback control loop designed to regulate the luminance and color temperature of a SSL . For a light sensing circuit, consisting of a photo-pic vision-corrected photodiode and a trans impedance amplifier, a low bandwidth low-pass filter is required to remove the large amplitude of photodiode current variations caused by the pulsating nature of conventional PWM dimming.

C. Lamp Design Procedure When Using MPWM Dimming Technique.

An outline of lamp design procedure to achieve optimal performance when using MPWM dimming is provided next,

Step 1 (Determination of HBLED forward current):

The selection of the forward current Is, flowing through the HBLED depends on the maximum power rating of the HBLED, the rated lumen output of the SSL, the geometry of the lamp, and the thermal solution used to limit the worst-case junction temperature of the device. Based on the power rating of the HBLED, the device forward current is selected to maximize efficient operation of the device.

Step 2 (Determination of HBLED array configuration):

Selection of the number of series connected HBLEDs in a single string and the number of such parallel strings within an HBLED array is based on various design parameters, such as the target system efficiency, the maximum allowable solution size, the HBLED binning process, and the requirements placed by safety standards and institutes. Based on the desired system efficiency, it is possible to calculate the maximum allowable loss, and thus, the maximum allowable standoff voltage across the series connected linear-current regulator circuits. Given a statistical forward voltage variation with a single bin of HBLEDs, the number of series connected devices should be selected, such that the total string-to-string voltage variation is less than the maximum allowable standoff voltage across the current regulation circuit. The number of parallel HBLED strings Nis then calculated based on the total number of LEDs required in a lamp to achieve desired light output.

In practice, safety regulations as specified by agencies, such as Underwriters Laboratory (UL) or by European Norms (EN) restrict the maximum

safe output voltage, which further constrain the HBLED array design.

Step 3 (Selection of dimming frequency):

The MPWM dimming frequency fdim, is selected between the fdim_minand and fdimmax, limit, as given by (3). For a given fdim, the optical and audible response of the lamp is measured and changes are made to achieve optimal performance.

Step 4 (MPWM design):

Analog or digital modulator architecture capable of achieving a minimum “ON” time τmin, calculated based on (2), is selected. For the experimental setup consisting of 64 HBLEDs, with eight parallel strings of eight series connected HBLEDs, a dimming frequency of 1.5 kHz is selected based on the desired 10-bit resolution on the input-dimming command and minimum response time tcr =650 ns of the discrete linear-current regulator circuits.

III. DESIGN OF THE DC–DC BOOST POWER CONVERTER

The impact of MPWM on the design of boost dc–dc converter is studied in this section. As the variation in output current is limited to a value less than one HBLED string current Is, it is now possible to optimize the size of energy storage elements in the boost converter. For the design of closed-loop feedback control, the complete HBLED array can be modeled as an equivalent current sink and converter transfer functions can be derived by making the small-signal approximation. Based on the linear small-signal model, the equations are derived for calculating the value of boost converter inductor and capacitor value. A simple proportional integral compensator Gc based on voltage-mode feedback, is proposed for regulating the output voltage and rejecting external disturbances, such as fluctuations in input power supply Vg, changes in load current based on the dimming command, Ddim, and variations in ambient temperature. An improvement in dynamic response of the boost converter by controlling the output current transient using a dimming filter Gdim, is then demonstrated. The performance of the closed-loop system is evaluated using MATLAB Simulink. In comparison to conventional PWM dimming, it is shown that the MPWM dimming requires a smaller value of boost output capacitor.

It is demonstrated that output voltage regulation of the dc–dc boost converter can be achieved using an outer digital voltage-mode feedback control, while the current through each HBLED string is regulated by series-connected current regulation circuits. This technique can be used as an

12 International Journal for Modern Trends in Science and Technology alternative to the conventional current-mode

control technique for regulating the output voltage of boost dc–dc that is supplying power to the entire HBLED array.

A. Design of Boost Converter Power Stage

The boost converter is designed to meet the specifications based on the allowed variation in inductor current and the output voltage. The inductor current ripple value is generally specified in terms of a parameter k, which denotes the ratio of peak to average inductor current ∆iL, to the maximum inductor current ILmax, and determines that boundary between the continuous conduction mode (CCM) and discontinuous conduction mode (DCM). Based on k, and assuming no losses (ideal operation), the inductor value is calculated using

(4)

Where Vg is the nominal input voltage, Vo is the regulated output voltage, Iomax is the maximum output current (Iomax=NIs), and fs is the switching frequency. The output voltage variations are generally specified in terms of maximum allowable ripple during steady-state operation ∆vrr, and maximum peak overshoot during output current transient ∆vos. The value of output capacitor is first selected to meet the output voltage ripple specifications ∆vrr, and then, modified to meet the maximum overshoot limit based on the closed-loop feedback design. In steady state, the input-dimming command Ddim, is assumed to be a constant and the output current io(t), is modeled as a dc-biased pulse train

(5) Where 𝐼₀ is the magnitude of the dc component and 𝑖_𝑜𝑘 is the magnitude of the kth harmonic component. The phase information of the Fourier coefficients is redundant for the derivation and is neglected. The magnitude of output current variation in steady state is approximated as the fundamental component

(6) Based on the small-signal output impedance in CCM operation, Zo(s) is the amplitude of the

fundamental component of output voltage ripple can then be calculated as follows:

(7)

Fig. 4. Small-signal model of the CCM duty-cycle controlled boost converter derived based on averaged switch modeling technique.

The CCM output impedance Zo(s) and the resonant frequency fo are calculated from the linear average switched mode shown

in Fig. 4 and is as follows:

(8) Where D’ =(1 – D) and D is the steady-state duty cycle and RL is the series resistance of the inductor estimated from the manufacturer’s datasheet or measured using an LCR meter. The derivation neglects the effect of capacitor ESR on the dynamics of the converter by assuming that low ESR MLCCs are used. For a lamp with large number of HBLED strings N, operating at a dimming frequency greater than 120 Hz, the frequency of fundamental component of output current variation can be assumed greater than the resonant frequency (f1 >fo), and the output impedance can be approximated as follows:

(9) Neglecting effect of higher harmonics current variation on the output voltage ripple, we set the value fundamental component of output voltage ripple to be equal to a specified limit ∆vrr1 = ∆vrr.

From (7) and (9), the value of capacitor required to achieve the desired output voltage variation specification is calculated as follows:

(10) The capacitor value calculated using (10) results in a lower estimate as all higher order harmonics in output current variation are neglected for the derivation. Thus, a value capacitor greater than

13 International Journal for Modern Trends in Science and Technology that calculated in (10) is selected. Note that when

MLCCs are used as filter capacitors, a value close to two times larger than that calculated is required to compensate for the loss of capacitance when operating under dc-biased condition. As the boost converter output impedance when operating in DCM is also given by (9), the capacitor value can be calculated based on (10) to meet the desired specifications.

From (10) it is clear that capacitor N times larger would be required to meet the design requirement when conventional PWM dimming is used as compared to MPWM dimming. Thus, by using MPWM dimming instead of PWM dimming, the area of boost converter can be reduced.

The output voltage sensing networkR1andR2in boost dc–dc converter presents a constant resistive load. The sum of resistances (R1 +R2), sets the minimum input power drawn by the converter and limits the range of DCM operation. The selection of resistorsR1 andR2 is done to prevent no-load operation of the converter when all the HBLED strings are turned OFF and to safely discharge the output capacitor.

B. Design of Voltage-Mode Feedback Compensator:

Voltage-mode feedback control can be used along with MPWM dimming to regulate the output voltage and reject the disturbance created by the slow variation in input-dimming command Ddim, and hence, the output current io(t), the input voltage Vg, and HBLED forward voltage drop. This is a key advantage over PWM dimming, where large period steps in output current necessitate the use of current-mode control techniques to achieve wide bandwidth and fast settling time. As sensing of the inductor current is not required, it is now possible to minimize the size of passive components by operating at a higher switching frequency and improve the system efficiency by reducing the conduction loss. Further, the voltage-mode feedback compensator can now be implemented using digital logic, allowing integration of the complete system using either a single microcontroller or a digital application-specified integrated circuit (ASIC).

The design of voltage-mode feedback control is based on the transfer function Gvd(s), from duty cycle D(s) to the output voltage Vo(s), derived using linear small-signal model of the boost converter operating in CCM, as shown in Fig. 4. The expression for boost control to output transfer function is as follows

(11) IV.MATLAB MODELING AND SIMULATION

RESULTS

Here simulation is carried out in different cases, in that

1). Implementation of Proposed Concept using 8 nominal switches. 2). Open Loop Operation of Proposed Boost Converter Concept using single switch. 3). Closed Loop Operation of Proposed Boost Converter Concept using single switch. 4).

Proposed Boost Converter System used in Power Factor Correction Applications.

Case 1: Implementation of Proposed Boost Converter Concept using 8 nominal switches

Fig.5 Matlab/Simulink Model of Proposed Converter using 8 switches

Fig.5 shows the Matlab/Simulink Model of Proposed Boost Converter using 8 switches using Matlab/Simulink Platform.

14 International Journal for Modern Trends in Science and Technology Fig. 6 Switching states

Fig.7 Output Voltage & Output Current Fig.7 shows the Output Voltage & Output Current of Proposed Boost Converter using 8 switches.

Case 2: Open Loop Operation of Proposed Boost Converter Concept using single switch.

Fig.8 Matlab/Simulink Model of Proposed Boost Converter with Sudden Loading Condition Fig.8 shows the Matlab/Simulink Model of Proposed Boost Converter with Sudden Loading Condition using Matlab/Simulink Platform.

Fig.9 Output Voltage & Output Current Fig.9 shows the Output Voltage & Output Current of Proposed Boost Converter with sudden switching.

Case 3: Closed Loop Operation of Proposed Boost Converter Concept using single switch

Fig.10 Matlab/Simulink Model of Proposed Boost Converter with Sudden Loading Condition Fig.10 shows the Matlab/Simulink Model of Proposed Closed Loop Boost Converter with Sudden Loading Condition using Matlab/Simulink Platform.

Fig.11 Output Voltage & Output Current Fig.11 shows the Output Voltage & Output Current of Proposed Closed Loop Boost Converter with sudden switching.

Case 4: Proposed Closed Loop Boost Converter System used in Power Factor Correction Applications

15 International Journal for Modern Trends in Science and Technology

Fig. 12 Matlab/Simulink Model of proposed Boost Converter with PI Controller.

Fig. 12 shows the Matlab/Simulink Model of proposed Boost Converter with closed loop operation conventional PI Controller.

Fig. 13 Output Voltage

Fig. 14 Source Parameters

As above fig. 13 shows the Output Voltage of proposed Converter, Fig. 14 shows the Source Side Parameters of proposed Boost Converter with Conventional PI Controller

Fig. 15 THD Analysis of Source Current with Proposed Boost Converter

Fig. 15 shows the FFT Analysis of Source Current with Proposed Boost Converter, we get 4.81%.

V.CONCLUSION

Here presented one new and interesting DC/DC boost-type converters for PFC applications with closed loop operation using PI controller. Without using any dedicated converter, one converter can be used to eliminate the harmonic current generated by the other non-linear load. A MPWM dimming technique is introduced for controlling output luminance of a SSL consisting of a number of parallel strings of series-connected HBLEDs. A significant reduction in the magnitude of peak-to-peak variation of the total lamp luminance and the output current drawn from the power supply is obtained by uniformly phase shifting the control signals used to pulse the current through each HBLED string. Compared to the conventional PWM dimming technique, an improvement in the quality of light output is observed when MPWM dimming is used, along with a reduced magnitude or elimination of visible flicker perception. With the help of simulation study, it can be concluded that, this configuration removes almost all lower order harmonics, hence with this configuration we can achieve power factor nearer to unity, THD less than 5%, using Fuzzy controller we get fast dynamic response. However, this technique can be limited to application where the non-linear load (pulsating) current is less and fixed. THD Analysis of both controller well within IEEE Standards.

REFERENCES

[1] M. Shur and R. Zukauskas, “Solid-state lighting:

Toward superior illumination,”Proc. IEEE, vol. 93, no.

10, pp. 1691 1703, Oct. 2005.

[2] M. Krames, O. Shchekin, R. Mueller-Mach, G.

Mueller, L. Zhou, G. Harbers, and M. Craford, “Status and future of high power lightemitting diodes for solid-state lighting,”Display Technol. J., vol. 3, no. 2, pp.

160–175, Jun. 2007.

[3] M. Dyble, N. Narendran, A. Bierman, and T. Klein,

“Impact of dimming white LEDs: Chromaticity shifts due to different dimming methods,” Proc. SPIE, vol. 5941, pp.

291–299, Aug. 2005.

[4] S. Muthu, F. Schuurmans, and M. Pashley, “Red, green, and blue LEDs for white light illumination,” IEEE J. Select. Topics Quantum Electron., vol. 8, no. 2, pp.

333–338, Apr. 2002.

[5] A. Zukauskas, M. S. Shur, and R. Gaska, Introduction to Solid-State Lighting. New York:

Wiley-Interscience, 2002.

[6] K. Wang, X. Luo, Z. Liu, B. Zhou, Z. Gan, and S. Liu,

“Optical analysis of an 80-W light-emitting-diode street

16 International Journal for Modern Trends in Science and Technology lamp,” Optical Eng., vol. 47, pp. 013002–0130013, Jan.

2008.

[7] G. Harbers, S. J. Bierhuizen, and M. R. Krames,

“Performance of high power light emitting diodes in display illumination applications,”Display Technol. J., vol. 3, pp. 98–109, Jun. 2007.

[8] J. W. Curran and S. P. Keeney, “Replacement of fluorescent lamps with high-brightness LEDs in a bridge lighting application,” Proc. SPIE, vol. 6337, pp.

633719-1–633719-11, Aug. 2006.

[9] Datasheet, “TLC5940 – 16 Channel LED driver with dot correction and grayscale PWM control,”Texas Instruments Incorporated, Oct. 2007. [Online].

Available:

http://focus.ti.com/lit/ds/symlink/tlc5940.pdf.

[Accessed: Oct. 2009].

[10] Datasheet, “MAX16809 – Integrated 16-channel LED drivers with switch-mode Boost and SEPIC controller,” Maxim Integrated Products, Dallas Semiconductor, March 2007. [Online]. Available:

http://datasheets.maxim-ic.com/en/ds/MAX16809-M AX16810.pdf. [Accessed Oct. 2009].

[11] Datasheet, “LT3595 – 16 channel Buck mode LED driver,” Linear Technology, Oct. 2007. [Online].

Available: Datasheet/3595f.pdf. [Accessed: Oct. 2009].

[12] E. F. Schubert, Light-Emitting Diodes. Cambridge, U.K.: Cambridge University Press, 2006.

[13] I. Ashdown, “Extended parallel pulse code modulation of LEDs,” Proc. SPIE, vol. 6337, pp.

63370W-1–63370W-10, Aug. 2006.

[14] L. Svilainis, “LED brightness control for video display application,” Displays, vol. 29, no. 3, pp.

243–249, Jul. 2008.