Proceedings of the International Conference on Computer and Communication Engineering 2008 May 13-15, 2008 Kuala Lumpur, Malaysia

Microcontroller Based Standalone PV system for Wireless Sensor Node

Md. Anwarul Azim1, 2, S. M. Ariful Huda2, Hafizal Mohammad1, Nowshad Amin3

1
Faculty of Engineering, Multimedia University (MMU), Cyberjaya 63100, Selangor, Malaysia
2
Dept. of CSE, Faculty of MS, International Islamic University Chittagong (IIUC), Bangladesh
3
Dept. of EESE, Faculty of Engineering, National University of Malaysia (UKM),
Bangi 43600, Selangor, Malaysia
Email (azimcse@gmail.com, smarifulhuda@gmail.com, hafizal@mmu.edu.my,
nowshad@eng.ukm.my)

Abstract environment; it should be efficiently integrated into an

embedded system to translate that harvested energy
This study integrates the photovoltaic panel power into increased application performance and system
monitoring, the charge controller as well as switching lifetime [1, 2]. The application space for wireless
options for remote monitoring applications of any sensor networks is dominated by the longevity
wireless sensor network. To increase the life time of a constraint, since the cost of physically deploying the
sensor node, recharge cycle is optimized in such a way sensor nodes often outweighs the cost of the nodes
that the battery is not charged randomly even though themselves. Energy is the limiting factor in achieving
sunlight is available on PV panel, rather it is charged extreme (months to years) system wide lifetime.
while discharge reaches to a certain predefined limit. However, the design of a solar energy harvesting
Thus, it increases the storage battery life time by module involves complex tradeoffs due to the
reducing the recharge cycle. Here, a powerful PIC interaction of several factors, such as the
microcontroller is used for future enhancement of a characteristics of the solar cells, chemistry and
complete sensor node solution for environmental capacity of the batteries used, power supply
monitoring application. However, cost-efficiency can requirements and power management features of the
be obtained by selecting optimum hardware for embedded system. In outdoor settings, the most
particular device of specific applications with the same accessible environmental energy source is solar. While
methodology described in this paper. photovoltaic (PV) power systems are in widespread
use in many settings, the design of a PV system for
Keywords: Solar PV power supply, Charge controller, continuous operation of wireless sensor nodes presents
Sensor node, microcontroller. a number of unique challenges. It should be simple,
robust, and operate with no human intervention for
I. INTRODUCTION many years [3]. It is, therefore, essential to thoroughly
understand and judiciously exploit these factors in
Electrical power in many portable products is
order to maximize the energy efficiency of a solar
supplied by batteries that can be recharged by
harvesting module. Duty cycle and power requirements
electricity from the mains. Due to costs, inconvenient
are low for sensor networks, but the load varies over
use and environmental concern, it would be useful to
huge range–microwatts in standby and mW when
find alternative ways to supply energy to products. The
active. Many applications operate at low duty cycles in
main components of a solar PV power setup (solar
unpredictable environments, so the system should
panel, battery and charge controller) are readily
adapt to the available energy reserve. Finally, the
available, and can be used to supply power for wireless
physical deterioration of the energy storage device is
sensor nodes. However, current designs did not meet
generally the overall limiting factor of lifetime of the
our goals in terms of cost, ease-of-use, and remote
device [5]. For example, rechargeable batteries have
monitoring. Ideally, system design should have the
about 300 to 500 recharge cycles, resulting in at most
least components from an ease-of-use and integration
one to two years of operation if charged daily. The key
in cost standpoint. Solar harvesting technology
design challenge is reducing the strain on storage
provides the ability to extract energy from the
elements while preserving a simple hardware and

978-1-4244-1692-9/08/$25.00 ©2008 IEEE 1136

Authorized licensed use limited to: Universitas Brawijaya. Downloaded on May 20,2020 at 05:38:03 UTC from IEEE Xplore. Restrictions apply.
software architecture. We discuss the design of an be small in size and the voltage needed greater than 6V
intelligent system with lightweight and efficient in order for the microcontroller able to sense the
hardware combined with powerful software that voltage.
actively manages the power subsystem for continuous The charge controller should roughly double the
operation. Since solar energy directly enters the lifetime of the battery, by ensuring proper charging
battery, it experiences recharge cycles daily placing and discharging technique. On the other hand, each
significant stress on the battery. This limits the cell is 1.5V and it is certainly not recommended to
system’s lifetime to no more than two years. Such a discharge an entire battery of cells because when the
lifetime is not dramatically larger than that obtainable battery is discharged below 0.8V per cell, one of the
with batteries alone and far from perpetual operation. cells is inevitably weaker than the others, and goes to
In this paper, we present a simple, low cost and mostly zero first. If the battery is further flattened, this battery
digital design with minimal hardware components. The becomes charged in reverse, which again makes it still
microcontroller is programmed to intelligently control weaker. During the recharging process, once batteries
the operation and easily customizable. The life time, reach their 100% charge, the cell starts to heat and the
therefore, will hopefully be increased due to optimized internal pressure builds up, causing the battery to
recharge cycle proposed and demonstrated in the eventually rupture or vent electrolyte. Thus the battery
system. deteriorates rapidly due to more and more
overcharging. The manufacturer suggests these cells
II. SYSTEM ARCHITECTURE should be good for 500 to 1000 cycles if properly
The system functionality can be categorized into treated. In this study, we demonstrate the solution of
charge controller and system monitoring. The charge this problem in such a way that not to charge the
control aspect involves the solar power conversion, battery randomly when panel voltage are available
sensing of panel and battery voltage, and battery rather the battery starts to charge when its discharge
charging, while system monitoring aspect is concerned reaches to a certain (e.g. user-defined) limit.
with switching to different subsystem to increase the
lifetime of battery as well as sensor node. The system III. SYSTEM DESIGN
has four major blocks, such as solar panel, battery, A microcontroller controls the operation of the
microcontroller, switching and the load. Figure 1 entire sensor node [4]. Here, PIC16F877A
shows the block diagram of the proposed PV module microcontroller is used to control the sensor node. The
for sensor node. Harvesting components, such as solar chips possess CMOS FLASH-based 8-bit
panels, and energy storage elements, such as batteries microcontroller packs Microchip's powerful PIC
or ultra-capacitors, have different voltage-current architecture into a 40 pin package. It has wide
characteristics, which must be matched to each other as operating voltage range: 2.0V to 5.5V, 200
well as the energy requirements of the system to nanosecond instruction execution and 35 single word
maximize harvesting efficiency. Furthermore, battery instructions. The PIC power consumption is very low,
non-idealities such as self-discharge and round trip e.g. below 0.6 mA typical @ 3V, 4 MHz, and 20 μA
efficiency, directly affect the energy usage and storage typical @ 3V, 32 kHz and < 1 μA of typical standby
decisions. current. The built-in ADC can be used to take the
analog input like panel or battery voltage and others
sensors for actual sensor node operation. The
controller takes power from the battery and talks to the
voltage and current sensing circuits to determine the
voltages and currents of the solar panel and the battery.
It will charge the battery from panel through switch
when required and connect the load when panel has
sufficient voltage. When panel does not produces any
voltage (usually at night), the controller connect
battery with the load to perform the operation until the
battery discharge reaches to a certain level.
Figure 1. Block diagram of proposed PV module The goal of this design is to provide continuous
power to a sensor node independent of a power grid.
The solar panel has to supply enough power for the
This is accomplished by using solar power and a
load, consumption of the circuit, and to charge the
rechargeable battery. The primary objectives are
battery if it is empty [4, 6]. The solar panel needed to

1137

Authorized licensed use limited to: Universitas Brawijaya. Downloaded on May 20,2020 at 05:38:03 UTC from IEEE Xplore. Restrictions apply.
compact and mostly digital design, efficient operation,
ability to monitor available battery charge status, and
to provide a regulated 5V output. The major benefit is
the use of the renewable solar energy to provide power
to a sensor or any portable device that require up to
50mW to 100mW of power consumption. In addition,
the battery can be used in lower sunlight conditions or
during nights to supply any loads. The available
battery charge could be monitored and the controller
goes into low power mode when little charge is
available.
A. Program flowchart Figure 4. Microcontroller view.
The program flowchart is shown in figure 2.

Figure 2. Program flowchart.

The program was developed using Eclipse Hi-Tide

C/C++ and generated hex file of the program code. A
snapshot of initialization code of charge process is Figure 5. Circuit diagram.
shown in figure 3. Finally, PIC simulator IDE is used
to download the hex file into the microcontroller to C. Working principle
realize the appropriate operations. Figure 4 shows the
microcontroller view in the simulation process. Panel voltage is directly related to sunlight. So it
can be fluctuating gradually, but PIC microcontroller
needs a constant voltage. Therefore, power is supplied
to PIC from battery. On the other hand if we provide
power to PIC from panel, we can use a voltage
regulator but it would be not efficient because it has an
internal power loss and it must have a certain voltage
difference between input and output.
Voltage monitoring (Panel and Battery): When
Figure 3. A snapshot of initialization code.
panel has available power, it charges capacitor (C1)
through a Scotty diode (SD1). This voltage is
continuously monitored by PIC through RA0/AN0. If
B. Circuit design PIC gets sufficient voltage from VR1, then it compares
The circuit diagram is shown in figure 5. it with predefined reference voltage (Vr). If C1 has

1138

Authorized licensed use limited to: Universitas Brawijaya. Downloaded on May 20,2020 at 05:38:03 UTC from IEEE Xplore. Restrictions apply.
sufficient voltage then it allows the charger circuit to IV. CONCLUSIONS
turn on. The PIC also monitors the battery voltage (Bv) Solar-powered sensor node could be deployed in
via VR2 through RA1. The PIC only charges the outdoor environments where it would be difficult or
battery when the battery reaches its lower limit (should
impossible to change batteries but long-term operation
be set for charging the battery) and C1 (i. e Panel) has
is required. The main design specifications were to
sufficient voltage. It charges the battery until the
achieve the ability to provide 30-50mW continuously
battery reaches its upper limit (set by default).
Load control: Load gets supply from panel or to the sensor node. The system is accomplished
battery. When PIC gets sufficient voltage from C1 then cheaply using off-the-shelf parts and existing
the load connects to panel via P-MOSFET switch electronics are power-efficient enough to built low
(Q1). Otherwise load connects to the battery via P- power consuming devices. Even though efficiency of
MOSFET switch (Q3). When battery voltage reaches the overall design has not been specified, we have
to its predefined lower limit, then PIC automatically verified the efficiency on individual components and
disconnects the load. On the other hand, PIC is always that the internal circuit does not consume much of the
connected with the battery to avoid system down. power from the solar panel. The circuit provides
Switching: Three P-channels MOSFET (Q1, Q3, necessary power output at 5V as well as monitored the
Q5) are used to efficiently switch to different battery charge constantly for the purpose of internal
subsystems of the power supply unit. Q1 and Q2 control. The design of six solar panels in series
provide voltage to load. When panel voltage is provided enough power for the load and internal
available then RD2 output pin of PIC goes high and circuit consumption as well as was able to charge the
Q2 becomes active. In this way Q2 triggers Q1 and battery when it was more or less discharged. The
load gets supply from panel via Q1 and diode (D1). If software to control the charging process as well as
there is no voltage available in solar panel then RD1 other operation is small and simple enough to fit into
output pin of PIC goes high and Q4 and Q3 get an inexpensive microcontroller as shown in this study.
enabled. Thus load gets connected with the battery via
Finally, at a very unlike case, if the battery runs out,
Q3 and D2.
the system proves to be able to go into a stand-by
Charge control: If battery discharges to its lower
mode as an important feature.
limit at C1 then RD0 output pin of PIC goes low. Thus
Q6 becomes disabled, whereas Q5, Q7 become enable.
Then the battery recharges from lower limit to upper REFERENCES
limit through Q5, D3 and Rx. On the other hand, when [1] Gabriel A. Rincón-Mora, Senior Member, IEEE, and Justin P.
Vogt, Student Member, IEEE, “Self-powered wireless sensor
the battery becomes full charge, RD0 pin goes high
nodes: Among other things, a load management feat”, The
and prevent the battery for overcharge. In this way, we Georgia Tech Analog and Power IC Lab.
avoid the battery to charge randomly when panel [2] K. Iniewski, et al, "Ultra-Low Power Circuit and System
voltage is available thus reduce recharge cycle. When Design Trade-Offs for Smart Sensor Network Applications,"
ITI 3rd Int. Conf. on Information and Communications
the system fails due to power failure then Q6 becomes
Technology, 2005, pp.309-321.
inactive because of no bias voltage at the base. After [3] X. Jiang, J. Polastre, and D. Culler, "Perpetual Environmentally
getting sunlight, the system starts charging the battery Powered Sensor Networks," Proc. 4th Int'l Conf. Information
automatically through Q5, D3 and Rx as Q7 is still on Processing in Sensor Networks/Special Track on Platform
while Q6 is off. Tools and Design Methods for Network Embedded Sensors
(IPSN/SPOTS), Los Angeles, CA, Apr. 2005
Finally, a register (R3) is connected with a Zenner [4] Joel Christopher Jordan, "Design and implementation of a
diode (Z1), which provides a constant voltage for stochastic wireless sensor network", Masters Thesis, University
reference (reference voltage may be taken from PIC of Illinois at Urbana-Champaign, 2004.
internal reference voltage register). Here, we use 6V [5] K. Lahiri, A. Raghunathan, and S. Dey, “Battery-driven system
design: a new frontier in low power design”, in Proc. IEEE
voltage regulator IC named LM7806. We need a 5V International Conference on VLSI Design, pp. 261–267, 2002.
supply for load but we use a 6V regulator, as we use a [6] D.L King, J.A. Kratochvil, and W.E. Boyson. “Field experience
diode OR gate to isolate two switches (Q1 and Q3) to with a new performance characterization procedure for
supply to the load. This diode has an internal loss of photovoltaic arrays”. In Proc. 2nd World Conference and
Exhibition on Photovoltaic Solar Energy Conversion, 1998.
0.7V, so it becomes 5.3V which is near to 5V.

1139

Authorized licensed use limited to: Universitas Brawijaya. Downloaded on May 20,2020 at 05:38:03 UTC from IEEE Xplore. Restrictions apply.