See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/331730526

Ambient light adaptive LED light dimmer

Article in TELKOMNIKA (Telecommunication Computing Electronics and Control) · February 2019

DOI: 10.12928/TELKOMNIKA.v17i1.10114

CITATIONS READS
2 4,597

4 authors, including:

Taufik Taufik Saidah Saidah

California Polytechnic State University, San Luis Obispo University of Surabaya Bhayangkara
228 PUBLICATIONS 1,504 CITATIONS 12 PUBLICATIONS 33 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The DC House Project View project

Control Straategy for PWM Voltage Source Converter Using Fuzzy Logic for Adjustable Speed DC Motor View project

All content following this page was uploaded by Taufik Taufik on 26 April 2019.

The user has requested enhancement of the downloaded file.

TELKOMNIKA, Vol.13, No.2, June 2015, pp. 125~132
ISSN: 1693-6930, accredited Aby DIKTI, Decree No: 58/DIKTI/Kep/2013
DOI:10.12928/TELKOMNIKA.v13i2.xxxx 281

Ambient Light Adaptive LED Light Dimmer

Taufik1, William Xiong1, Jonathan Sato1, Saidah2

1Department of Electrical Engineering, California Polytechnic State University
1 Grand Avenue, San Luis Obispo, USA, +1-805-7562781
e-mail: [email protected]
2Department of Electrical Engineering, Bhayangkara University

Jl. Ahmad Yani 114, Surabaya, Jawa Timur 60231, Indonesia, 031-8285602
e-mail: [email protected]

Abstract
This paper presents the design of an Adaptive Light Dimmer based on the method of
sensing ambient light content to adjust lamp's light intensity accordingly, and thus regulating the
room's light content. The device is designed to work with renewable energy sources such as wind
and solar energy. This would be useful in less developed countries where AC electricity is not
well spread and renewable DC sources, such as solar, can be better utilized. It functions by using
the TSL2561 light sensor, ATTiny85 microcontroller to output PWM to the LED driver, LT3795
LED driver to output current to an LED and LT3014 LDO to lower the input voltage and power the
microcontroller and sensor. The dimmer is designed to work with a 48V input voltage and operate
from an input light range of 20 to 100 Lux. Above 100 Lux the light is off and below 20 Lux the
light is fully on.

Keywords: light dimmer, LED driver, adaptive dimmer

1. Introduction
With regard to efficiently using sustainable energy, there is a significant amount of data
supporting the energy saving capabilities of light dimming technology. Some state that smart
lighting, using sensors and controllers to control lighting, saves between 50% and 70% of energy
compared to an uncontrolled lighting system [1-5]. These energy savings have huge impacts,
considering that lighting accounts for about 19% of the electrical energy generated worldwide. In
commercial buildings, lighting accounts for even more, 30-40% [6-11]. This being said, smart
lighting seems to be the next logical step in saving electricity and tackling the problems of world-
wide electrification in today’s world.
In order to improve smart lighting technology, it makes sense to utilize electrical devices,
such as microcontrollers, to increase efficiency and usability of lighting systems. Microcontrollers
would be a reasonable choice for controlling light dimmer due to their decreasing costs, versatility,
and ease of use [12-14]. In addition to microcontrollers, DC-DC converters are also essential to
improving smart lighting systems [15-17]. These converters are the most efficient ways of
converting DC power and consequently are used in many stages of power conversion. Without
them, electronics, such as lighting systems, would be much more inefficient. Any improvements
to these DC-DC converters directly improve the efficiency and performance of smart lighting
systems.
There are existing solutions for AC lighting in the form of smart lighting and socket to bulb
interfaces [18-22]. Smart lighting uses “smart systems” to control lighting with applications in
computers, tablet devices, or smart phones. On a larger scale, they can also be used to control
the lighting system of a room, house, or building. They can wirelessly turn on and off lights, control
their brightness, set the lights on a timer, and even integrate sensors. Existing socket to bulb
interfaces use a sensor in an attempt to dim the bulb appropriately but they do not function very
well. An example of an AC powered smart lighting device, called a home light control module
(HLCM) and designed by Ying-Wen Bai and Yi-Te Ku, uses passive infrared (PIR) sensors, light
sensors, a microprocessor, and an RF module to control light intensity in all the rooms of a house
[23]. The system is illustrated in Figure 1. A single HLCM controls one set of luminaires. As a
result, multiple microprocessors determine lighting levels, rather than a single central controller.

Received February 23, 2014; Revised May 29, 2014; Accepted June12, 2014
282 ISSN: 1693-6930

This device uses the PIR sensors to determine the presence of any people in a room, turning off
the lights if no one occupies the room. The light sensors determine the room's brightness levels.
If outside sources, like daylight, provide enough light, then the luminaires are turned off.
Otherwise, the system activates the appropriate amount of luminaires to achieve the desired
brightness levels. The device's RF module allows communication between different HLCMs. In
the case that brightness levels are insufficient even when all luminaires are on, communication
between HLCMs allows an adjacent HLCM to increase the number of lights to activate. This then
affects the light intensity of the first device's room. While these components help this device
operate with high efficiency, the design presented in this paper needs not contain any
unnecessary modules if their inclusion greatly increases purchase costs. For example, there won’t
be any need to use the RF module if its use produces costs that outweigh its energy benefits.
Additionally, the HLCM does not use light dimming technology while the proposed solution is
based on light dimming.

Figure 1. The configuration of an HLCM system for a 2 story house [20]

However, there are currently no existing DC dimmers that are inexpensive and entirely
autonomous. The objective of the proposed solution is therefore to design and construct an
adaptive DC light dimmer that autonomously dims a set of LEDs depending on the ambient light
sensed in the room, providing the appropriate amount of light to the environment it is in. The end
product of this project will save electricity usage; thus, reducing electrical costs and increasing
the viability of renewable energy.

2. Design Methodology and Requirements

For the proposed design, the adaptive light dimmer must be able to produce the same or
more light than a normal 60W incandescent light bulb at maximum brightness. This was
determined to be roughly 800 lumens. If the adaptive light dimmer is incapable of producing the
appropriate amount of light, a house may not be bright enough during darker hours of the day.
Another important aspect is the efficiency of the device. Having better efficiency leads to less
costs in terms of both energy and money. While both are definitely important, reduction in energy
costs will likely have a greater impact on those rural areas depending on renewable energy
sources, since these energy resources are much more limited. The adaptive light dimmer should
have at least an efficiency of 70% at full load.
In addition, the device should consume a relatively low amount of power. Considering
that the LED being used is a 5 W rated LED, the maximum output power must be 5 W. Derived
from the efficiency requirement, the maximum input power must be 7.14 W. The device must be
able to operate from a 48 V input, for example to match the DC bus used by the DC house project
[24]. It should ideally be able to operate at other DC voltages as well for other applications. The
device must also be quick to respond without causing abrupt changes. So, the device should be
able to adjust its output at least once every two seconds. Each adjustment may need to be gradual
for any sudden changes in the surrounding brightness levels.
The sensor should be sensitive enough that it can differentiate changes in daylight
throughout the day. The sensor measures light intensity in lux, so the placement of the sensor
can have a significant effect on measurement values. The best place to put the sensor is near the
floor, as this minimizes the vertical movement of the sensor. Near the floor, the expected lux

TELKOMNIKA Vol. 13, No. 2, June 2015 : 125–132

TELKOMNIKA ISSN: 1693-6930  283

measurements range from 35 to 150 lux.The device will be connected to a lamp, so the entire
system will be fairly large. Approximately, the height must be 4 to 5 feet, and the base should be
about a foot in diameter. The height will also help to spread the light to the entire room.
Figure 2 depicts the block diagram of the design, consisting of specialized blocks each
with their individual inputs and outputs (I/O). It specifies the types and connections of these I/Os
and how it achieves the desired output to the level 0 block diagram. The system first takes the 48
V DC and sends it through a Low-Dropout Regulator (LDO) to lower the voltage to 5 V so that it
may be used to power the microcontroller and photosensitive sensor. The photosensitive sensor
communicates via i2c to the microcontroller to give a light intensity reading. The microcontroller
then maps this reading to a corresponding pulse width modulated (PWM) control voltage that it
outputs to the LED driver. Based on the PWM voltage it receives from the microcontroller, the
LED driver outputs appropriate PWM current using the 48 V DC from the bus. The LED is driven
by the current output and produces the correct amount of light out.

Figure 2. Block diagram of the adaptive light dimmer

The software flow diagram as illustrated in Figure 3 describes the process performed by
the microcontroller. The microcontroller initializes the photosensitive sensor and then goes into
an endless loop consisting of three stages. First it requests a reading from the sensor. Once it
has received and stored this reading, it maps the reading to an appropriate PWM output value
through a lookup table. The PWM is then written out to the appropriate pin.

Figure 3. Software flow diagram of the adaptive light dimmer

Title of Paper..Title of Paper.. (First Author)

284 ISSN: 1693-6930

3. Design and Simulation Results

The main purpose of the Adaptive Light Dimmer is to both minimize costs by using
cheaper materials while maximizing efficiency and to operate from a DC voltage source, allowing
renewables to be a direct source of power. Simplifying the design can also allow the device to
operate without worry of installation. The Adaptive Light Dimmer has three main parts to design.
First is the LED driver, which has to drive a 5 V LED at a maximum current of 800 mA. The second
is the LDO linear regulator, which had to step down the 48 V input to 5 V in order to power the
microcontroller and the light sensor. The last portion of the design is the combination of the
microcontroller and the light sensor, both of which have to operate at 5 V.
The LED driver used is Linear Technology’s LT3795 since it can operate with an input
voltage of 48 V. The circuit is a buck mode LED driver meant to operate from 24 V to 80 V and
output a controlled current to the LEDs. This current can be controlled directly by changing the
control voltage or by sending a PWM signal into the PWM pin of the device.The components that
has to be changed are the input capacitor CIN, the inductor L, the input current sense resistor
RINSENSE, the LED current resistor RLED, the PMOS, the NMOS, and the PNP BJT. These
components directly affect the operation of the circuit, so they are changed in order to produce
the correct maximum output current and to keep the circuit from damaging any components in
the design. Because the input voltage was 48 V, during transient operation, the voltage could
reach upwards of 63 V; thus, requiring careful consideration when designing the circuit. The
equation used to find the appropriate value of CIN in microfarads is:

(V − V ) 10μF
C =I ∙V ∙ ∙T ∙
V A ∙ μs
(1)
(48V − 5V) 1 10μF
C = 0.8A ∙ 5V ∙ ∙ ∙ = 1.87μF
(48V) 400000Hz A ∙ μs

The equation for sizing the inductor:

1
(T ∙R ∙V ∙ [V − V ]) ∙ 0.015Ω ∙ 5V ∙ [48V − 5V]
L= = 400000Hz (2)
V ∙ 0.02V 48V ∙ 0.02V
= 8.4μH

The input current sense resistor is determined more loosely. Because the configuration is a buck
LED driver, the input current should not be very high, so RINSNS is chosen as 100 mΩ. By the
equation below, the maximum input current can be determined.

0.06V 0.06V
I ( ) = = = 600 mA (3)
R 0.1Ω

The LED current resistor is chosen such that the maximum LED current is 800mA. Because the
control pins are tied high, the equation used to determine the LED current resistor is simplified to:

0.25V 0.25V
R = = = 312.5 mΩ (4)
I ( ) 0.8A

The PMOS and PNP BJT transistors are chosen differently because the ones from the datasheet
are not readily available for both simulation and hardware tests. The same is true for the NMOS
transistor, but in addition, the NMOS transistor is chosen to minimize the gate charge. Higher gate
charge values increases the current in the INTVCC pin and could disrupt the operation of the
circuit by dropping the INTVCC pin below the threshold.

I 0.02A
Q ( ) = = = 50nC (5)
f 400000Hz

The timing resistor, which determines the switching frequency of the LED driver, is determined to
yield 400kHz. This frequency is neither too low, which would increase component sizing, nor too

TELKOMNIKA Vol. 13, No. 2, June 2015 : 125–132

TELKOMNIKA ISSN: 1693-6930  285

high, which would increase losses. The feedback resistors, which set the maximum output
voltage, are adequate because the LEDs operated at 5 V. The sense resistor RSNS is used to
measure the actual current through the LEDs. Its value must be less than 0.07V/ILED, which for
800mA, is calculated to be 87.5mΩ. The final design of the LED driver is shown in Figure 4.

Figure 4. Final schematic of LED driver for the adaptive light dimmer

Simulations of the design were done using LTSpice. Results are shown in Figures 5 and 6 for the
output voltage and the input current at 100% duty cycle, respectively.

Figure 5. Simulation result for output voltage at 100% duty cycle

Title of Paper..Title of Paper.. (First Author)

286 ISSN: 1693-6930

Figure 6. Simulation result for input current at 100% duty cycle

The second portion of the dimmer is the low drop-out regulator (LDO) circuit which lowers
the 48 V input to a usable 5 V for the microcontroller and sensor. The LT3014 was chosen
because it fits well into the design; it has the capability of dropping 48V to 5V, and it would incur
little power loss since the microcontroller/sensor circuit would not pull a significant amount of
current. Figure 7 shows the configuration of the LDO to be implemented taken from the LT3014
datasheet. This configuration is setup for “5V Supply with Shutdown” according to the datasheet
and fits the needs of the light dimmer.The value of the capacitor on the output was specified to
be at least 0.47 µF to prevent oscillations.

Figure 7. The LDO section of the adaptive light dimmer

The final portion of the dimmer consists of the sensor and microcontroller circuits. The
VEML6070 is chosen due to its ability to communicate via I2C and to detect UV light, and because
its operating voltage range that is the same as the ATTiny85. The ATTiny85 is chosen for its low
cost, I2C capability and ability to produce a PWM output. In addition, it also can be programmed
with the Arduino UNO using Sketch’s built in “ArduinoISP” library. Since the VEML6070
communicates via I2C, the SDA (data line) and SCL (clock line) have 4.7 kΩ pull up resistors to
ensure proper logic levels. These are connected to the 5 V from the LDO, which supplies voltage
to the microcontroller and UV light sensor. The output PWM on pin six on the microcontroller
connects to the PWM input of the LED Driver. The ACK pin of the VEML6070 sends a signal if
the UV light reading drops below a built in threshold and was not connected since the feature is
not necessary for the function of the device. The ground is common to the LED driver and LDO
circuits as well.The design of the circuit is shown in Figure 8.

TELKOMNIKA Vol. 13, No. 2, June 2015 : 125–132

TELKOMNIKA ISSN: 1693-6930  287

Figure 8. Sensor and Microcontroller Circuit Diagram of the adaptive light dimmer

3. Hardware Construction and Results

The hardware of the light dimmer was built and tested in separate stages. The LED driver,
LDO, microcontroller, and sensor were tested individually and then combined to construct the
entire system. After constructing the circuit on the PCB as shown in Figure 9, testing the individual
components, and checking all possible connections, power was applied and the entire system to
evaluate its functionality. Results from the test demonstrate that when dark below 10 Lux, the
LED driver received 100% duty cycle and the lamp went to full brightness, and when bright above
100 Lux, the LED driver receives 0% duty cycle and the lamp outputs no light. Different duty
cycles were assigned to different ambient light levels in between 0 and 100 Lux. This approach
reduces the complexity in the operation of LED drive using the pulse density modulation [25-26].
The LED dimmed and grew brighter according to how much light reached the sensor, and
the brightness changed smoothly without visible leaps in light output. The adaptive light dimmer
worked as expected as ambient light levels increased and increased in brightness during low light.
When dark, below 10 Lux, the LED driver received 100% duty cycle and the lamp went to full
brightness and when bright, above 100 Lux, the LED driver receiver 0% duty cycle and the lamp
outputs no light. Different duty cycles were assigned to different ambient light levels in between 0
and 100 Lux. The LED dimmed and grew brighter according to how much light reached the
sensor, and the brightness changed smoothly without visible leaps in light output. Furthermore,
the proposed design was quick to respond to changes in the environment, as the LED was able
to change more quickly than once every two seconds and each change was gradual. This is
important to prevent the LED’s dimming from bothering users. Overall, the proposed adaptivelight
dimmer is an improvement in functionality compared to the results presented in [2].
However, this achievement comes at the cost of efficiency because the microcontroller
and sensor draw about 30 mA, increasing the input power needed while the output power of the
LED stays constant. This current is relatively significant compared to 115 mA input current needed
for the LED driver. Thus, the microcontroller and sensor increased the input current to yield input
power by 26%. The high input voltage of 48 V meant that input power increased by about 1.44
W, about 1.29 W of which was dissipated through the LDO. Figures 10 through 12 detail the
hardware test results of the adaptive light dimmer prototype. These results indicate the linear
operation of the light dimmer as duty cycle is varied. The efficiency after connecting the
microcontroller and LDO was 48.7%, which is well below that presented in [27]. Based on these
results, the main function of sensing ambient light levels and dimming appropriately has been
achieved by the proposed solution.
There are several ideas to improve the performance of the proposed adaptive light
dimmer. One example involves the addition of a 0.47 µF LDO output capacitor will be included on
the PCB design to ensure that the microcontroller and sensor circuit were receiving a more stable
voltage. This capacitor needs to be connected close to the supply of the microcontroller and
sensor. The easiest way to do this would be to connect a 0.47 µF through-hole ceramic capacitor
between the input voltage to the sensor on the PCB and the ground to the sensor on the PCB.

Title of Paper..Title of Paper.. (First Author)

288 ISSN: 1693-6930

Figure 9. Adaptive light dimmer hardware prototype

Input and Output Currents vs. Duty Cycle

350

300
Input and Output Currents [mA]

250

200

150

100

50

0
5 15 25 35 45 55 65 75 85 95
Duty Cycle [%]

Input Current [mA] Output Current [mA]

Figure 10. Hardware test results on input and output currents with varying duty cycle

Output Voltage vs. Duty Cycle

11.0

10.5
OUTPUT VOLTAGE [V]

10.0

9.5

9.0

8.5

8.0

7.5
5 15 25 35 45 55 65 75 85 95
DUTY CYCLE [%]

Figure 11. Hardware test results on output voltage with varying duty cycle

TELKOMNIKA Vol. 13, No. 2, June 2015 : 125–132

TELKOMNIKA ISSN: 1693-6930  289

Lux 4 ft Radius vs. Duty Cycle

7000

6000

5000

LUX 4 FT RADIUS
4000

3000

2000

1000

0
5 15 25 35 45 55 65 75 85 95
DUTY CYCLE [%]

Figure 12. Hardware test results on Lux 4 ft radius with varying duty cycle

4. Conclusion
An adaptive LED light dimmer which operates automatically based on ambient light
condition was presented in this paper. The design has been shown through computer simulation
and hardware prototype tests to successfully perform the dimming function following the changing
condition of the ambient light. The proposed solution reduces the complexity of previous
approaches and it can be adapted to any DC system operating at different DC voltage levels.
Further improvements of the proposed solution involves increasing the overall system efficiency.
This could be done by utilizing a snubber to the NMOS switch to reduce switching losses at 400
kHz. The PCB design could also be optimized for better signal integrity since many components
are farther apart on the current PCB design. Additionally, an alternative to the LDO utilized in the
proposed design must be sought to further improve the overall efficiency.

References
[1] C. Basu et al. Sensor-Based Predictive Modeling for Smart Lighting in Grid-Integrated
Buildings. IEEE Sensors Journal. 2014; Vol. 14, no. 12, pp. 4216-4229.
[2] J. Byun, I. Hong, L. Byongjoo, S. Park. Intelligent household LED lighting system considering
energy efficiency and user satisfaction. IEEE Transactions on Consumer Electronics. 2013;
Vol. 59, no. 1, pp.70-76.
[3] C. Huang, Chi-Han, B. Wang, C. Kang, J. Chen, C. Jen-Yao and I. Chou. Energy-saving
LED Control Module for Agent-based Micro-Grid Systems. Proceedings of IEEE
International Conference on Consumer Electronics. 2014; pp. 127-128.
[4] D. Edwards. Energy Saving LED Screw-in-base Bulbs: A Residential Demonstration.
Proceedings of World Energy Engineering Congress. 2014, pp. 1372-1380.
[5] J. Kuo. On LED Luminous Efficacy and Energy Saving Technique in Lighting. Advanced
Materials Research. 2012; Vol. 512-515, pp. 2693-2696.
[6] I. Galkin, L. Bisenieks and A. Suzdalenko. Impact of pulse modulation method of LED
dimmer for street lighting on its efficiency. Proceeding of 4th European Education and
Research Conference. 2010; pp. 160-164.
[7] K. Jain, K. Kumar, P. Thangappan, K. Manikandan, P. Magesh, L. Ramesh, and L. Sujatha.
Lighting electrical energy audit and management in a commercial building. Proceedings of
2nd International Conference on Intelligent Computing and Applications, 2015. Pp. 463-474.
[8] S. Bannamas and P. Jiraping. An intelligent lighting energy management system for
commercial and residential buildings. Proceedings of IEEE Innovative Smart Grid
Technologies – Asia. 2015.
[9] G. Lowry. Energy saving claims for lighting controls in commercial buildings. Energy and
Buildings. 2016; Vol. 133, pp. 489-497.

Title of Paper..Title of Paper.. (First Author)

View publication stats

290 ISSN: 1693-6930

[10] A. Kumar, A. Kajale, P. Kar, A. Shareef and S. Panda. Location-aware smart lighting system
for individual visual comfort in buildings. IEEE 6th Global Conference on Consumer
Electronics. 2017; pp. 301-306.
[11] N. Xavier, A. Kumar, S. Panda. Design fabrication and testing of smart lighting system.
Proceedings of the Future Technologies Conference. 2016; pp. 763-768.
[12] R. Goyal, L. Arya. Home automation and intelligent light control system using
microcontroller. Proceedings of International Conference on Computing, Communication
and Automation. 2017; pp. 997-1000.
[13] Z. Shu. Design of White Light LED Lighting Control System. Proceedings of International
Conference on Intelligent Transportation, Big Data & Smart City. 2018; pp. 561-563.
[14] L. M. Thet, A. Kumar, N. Xavier, S. K. Panda. A smart lighting system using wireless sensor
actuator network. Proceedings of Intelligent Systems Conference. 2017; pp. 217-220.
[15] J. Zhou, H. Guo, B. Bai, M. Duan, C. Lin. Power-white LED dimming detection system based
on NC PWM. Proceedings of the 11th IEEE International Conference on Anti-counterfeiting,
Security, and Identification. 2017; pp. 187-190.
[16] M. L. G. Salmento, R. Brier, J. Amaral, L. Diniz, V. M. Albuquerque, G. M. Soares, H. A. C.
Braga. Application of a flyback converter and variable pulse position modulation for visible
light communication. Proceedings of Brazilian Power Electronics Conference. 2017; pp. 1-
5.
[17] S. Mahadeokar, M. Sardeshmukh. Energy efficient PWM Dimmable Smart Digital LED
driver. International Conference on Energy Systems and Applications. 2015; pp. 3016-311.
[18] L. Martirano. A Sample Case of an Advanced Lighting System in a Educational Building.
Proceedings of IEEE Environment and Electrical Engineering Conference. 2014; pp. 46-51.
[19] J. Xiao, C. Liu and Y. Mo. Lighting Control and Its Power Management in Railway Passenger
Station. Proceedings of Asia Pacific Power and Energy Engineering Conference. 2009; pp.1-
4.
[20] S. Y. Chun and C. S. Lee. Applications of Human Motion Tracking: Smart Lighting Control.
Proceedings of IEEE Computer Vision and Pattern Recognition Workshops Conference.
2013; pp. 387-392.
[21] F. Wahl, M. Milenkovic and O. Amft. A distributed PIR-based approach for estimating people
count in office environments. Proceedings of IEEE Computational Science and Engineering
Conference. 2012; pp. 640-647.
[22] M. Miki, R. Okunishi, K. Matsutani, H. Ikegami and S. Fujimoto. An Intelligent Lighting
System with a Seat Management Mechanism. Proceedings of IEEE Systems, Man, and
Cybernetics Conference. 2013; pp. 4571-4576.
[23] Y. W. Bai and Y. T. Ku. Automatic room light intensity detection and control using a
microprocessor and light sensors. IEEE Transactions on Consumer Electronics. 2008; Vol.
54, no. 3, pp. 1173-1176.
[24] T. Taufik. The DC House project: An alternate solution for rural electrification. IEEE Global
Humanitarian Technology Conference. 2014; pp. 174-179.
[25] S. Rohit, R. Gunabalanand M. P. Kumar. Pulse Density Modulation Flyback Converter for
LED Automotive Lighting. Indonesian Journal of Electrical Engineering and Computer
Science. 2017; Vol. 8, No. 1, pp. 85-91.
[26] Y. Liu, S. Wang, and Y. Luo. Digital dimming control of CCFL drive system using pulse
density modulation technique. Proceedings of IEEE Regon 10 Annual International
Conference TENCON. 2007.
[27] R. N. Hasanah, Y. D. Handari, S. Soeprapto, and T. Taufik. LED-Lamp Design for
Renewable Energy-Based DC House Application. International Journal of Power Electronics
and Drive Systems. 2018; Vol 9, No 3.

TELKOMNIKA Vol. 13, No. 2, June 2015 : 125–132