Solar Power Source for Sensors - DOC by fta11207


									  Solar Power Source for Sensors


        Steven Portscheller
          Pavlina Akritas
          Gunjan Tejani


          SPRING 2006

      TA: Dwayne Hagerman

           May 1, 2006

          Project No. 10

This paper discusses the implementation of a solar panel and rechargeable battery system to provide
power to a remotely located sensor node. The design implements a peak power tracking technique to
obtain maximum power from the solar panel and supply it to the load and the internal circuit. The peak
power tracking is adaptable under various light and load conditions. The solar power is not just used to
supply power to the load, but also to charge a battery. The battery is constantly monitored for its
capacity or the available charge. Finally, it provides a regulated constant output for a sensor.

Furthermore, the topics include different testing procedures and verifications. Testing provides a high
efficiency and thus high performance design. The efficiency tests are executed by changing the input
voltage for a particular resistance at the output and determining input and output power. Technical data
shows the reliable and low cost design meeting the performance specification. Schematics as well as the
results of component testing are provided in this paper.

Finally, the topics discussed in this paper also consider accomplishments, future work and alternatives,
as well as the ethical considerations.

                                                         TABLE OF CONTENTS

1.   INTRODUCTION .................................................................................................................. iii
     1.1 Purpose ...............................................................................................................................1
     1.2 Specifications ......................................................................................................................1
     1.3 Subprojects .........................................................................................................................1

2.   DESIGN PROCEDURE AND DETAILS................................................................................3
     2.1 Design Decisions and Details .............................................................................................3

3.   DESIGN VERIFICATION .......................................................................................................8
     3.1 Testing ................................................................................................................................8
     3.2 Conclusions.......................................................................................................................10

4.   COST ......................................................................................................................................12
     4.1 Parts ..................................................................................................................................12
     4.2 Labor .................................................................................................................................12

5.   CONCLUSIONS ....................................................................................................................13

     APPENDIX – TITLE .............................................................................................................14

     REFERENCES .......................................................................................................................21

                                         1. INTRODUCTION

The goal of our design was to provide continuous power to a sensor node independent of a power grid.
This is accomplished by using solar power and a rechargeable battery. We implemented a peak power
tracking method in order to achieve efficient energy collection and maximize power extracted from
given solar panels. We used a gas gauge IC to monitor the battery charge status. The use of a switching
regulator provided higher efficiency and a 3.3V output.

1.1 Purpose

The primary objectives were compact design, efficient energy collection, ability to monitor available
battery charge status, and to provide a regulated 3.3V output. The major benefit was the use of the
renewable solar energy to provide power to a sensor or any portable device that required up to 50mW of
power consumption. In addition, the battery can be used in lower sunlight conditions or during nights to
supply any load requiring power to operate. The available battery charge could be monitored and the
circuit can go into low power mode when little charge was available. Moreover, the peak power
tracking method helped to extract maximum power from the solar panel at any given condition.

1.2 Specifications

The main design specifications were to achieve the ability to provide 30-50mW continuously with the
ability to provide up 3.3W short term. The compact size and a constant output voltage of 3.3V were also
specified. Even though efficiency of the overall design was not specified, we verified efficiency on
individual components and made sure that the internal circuit does not consume much of the power from
the solar panel. We verified these specifications through testing and made sure that components
functioned together efficiently.

1.3 Subprojects

The design involves individual component testing and subprojects to determine key parameters and
thresholds as well as to employ different techniques to meet the specifications.

1.3.1 The solar panel

The solar panel had to supply enough power for the load, consumption of our circuit, and to charge the
battery if it is empty. The solar panel needed to be small in size and the voltage needed to be varied
from 0 to 3.3V in order for the microcontroller to be able to sense the voltage.

1.3.2 Power tracking circuit

The power tracking circuit consisted of a dc/dc boost converter, a current sense amplifier, and a
microcontroller. The microcontroller monitored voltage and current of the solar panel and calculated
power. It enabled or disabled the dc/dc converter based on the change in power over time in order to
maintain maximum power conditions for the solar panel.

1.3.3 Battery

The 3.6V NiMH battery was rated at 700mAH, which based on the circuit consumption and the
maximum load conditions can provide 21 hours of power without charging. The output of the dc/dc
converter was an important variable. We needed to find a threshold for the converter output voltage
somewhere between where we provided enough voltage to charge the battery and also protected it from

1.3.4 Battery monitoring system

The battery monitoring system consisted of the battery, the gas gauge IC and the microcontroller. The
gas gauge IC was used to keep track of the available charge of the battery. The microcontroller utilized
serial communication to communicate with the gas gauge IC. It transmits an 8-bit address to the gas
gauge IC to read the NAC (nominal available charge) register and receives 8-bits back from the gas
gauge IC. For this to occur, we needed to find a compatible bit rate at which the gas gauge IC can
operate therefore, bit rate was an important variable to control. The system can be updated based on the
received information through the microcontroller.

1.3.5 Switching regulator

The switching regulator provided the final regulated output of 3.3V through buck-boost operation. The
input voltage was an important variable since the regulator operated between 2V and 5.5V. However,
2V across the battery terminal was sufficiently low to provide the load and the internal circuit.
Therefore, we shut-off the system before it reaches that low of a voltage. The maximum output current
of the switching regulator was 1300mA, which was well suitable for our maximum load condition.

The general block diagram of our design and all of the interconnections between major components are
shown in Figure (1.1).

                                     2. DESIGN PROCEDURE

2.1 Design Decisions

   2.1.1 Photovoltaic

Choosing a photovoltaic (PV) array was very important given that it was the source of the input power.
Crystalline silicon (c-Si) is the leading commercial material for PV cells with an efficiency of
approximately 20%, therefore we searched for PVs that used this type of semiconductor. The initial
concern was the low voltage and small size required by our design. Given this, the Power Film 3.6 V,
50mA flexible solar panel by Sundance Solar was initially chosen. Its size was measured to be 2.9” x 3”,
which was the exact size we were looking for.

However, the small rated short circuit current, ISC, of 50mA resulted in the power being too low
compared to the 30-50mW that was initially specified. A new search for PVs with higher ISC was put in
effect. The low size and voltage requirements made the search difficult given that PVs in such small size
and voltage were not readily available. We decided to use a PV from Edmund Scientific, item
#3039811, rated at an open circuit voltage (VOC) of 0.5 V, and ISC of 800mA. Given the low VOC, six
panels were put in series to increase the input voltage as shown in Figure 2.1.

   2.1.2 Power Tracking Design and Details

The IV characteristics of solar panels were such that there was a voltage for which maximum power was
extracted from the solar panel under a given set of operating conditions (insulation, temperature, etc.)
The goal of Maximum Peak Power Tracking (MPPT) was to actively drive the solar panel to this voltage
by changing the load the solar panel experiences. This was accomplished with a dc/dc converter and a
way to monitor whether or not the solar panel was operating at peak power voltage. After speaking with
Jonathan Kimball, a research engineer in power at University of Illinois, the initial MPPT method
pursued was constant voltage fraction. In this method it was assumed that the voltage at which
maximum power output occurs Vpp was a constant fraction of the panel‟s open circuit voltage Voc , which
varied depending on insulation. An open circuit test was performed periodically to determine Voc , and
then the converter was controlled to maintain the calculated Vpp , which was found as follows:
                                            Vpp  kVoc Vpp                                        (2.1)
This necessitated experimentation to determine the fraction constant k , which varied from solar panel to
solar panel. This design allowed MPPT to be implemented by monitoring solar panel voltage only, with
no need to monitor current. However, this was a purely model based design and was not adaptable to
changing temperatures or internal changes to the solar panel over time. Furthermore, the accuracy of this
method was heavily dependant on testing of the solar panel to be used and this testing consumed much

In light of the disadvantages of the constant voltage fraction implementation, an implementation that
used voltage and current readings to calculate the change in power and detect the peak of the power
curve was implemented. This offered an adaptable design that would work with different solar panels
without extensive testing.

   2.1.3 Current Sense Amplifier Design and Details

In order to implement a MPPT design that measured the change in power, a suitable current sense
amplifier and sense resistor was needed. The Maxim MAX4173 amplifier was chosen for its compact
design, low power consumption, and ability to work in 3.3 Volt circuits. The MAX4173 also came in
three different gains (20 V/V, 50 V/V, and 100 V/V), which would allow testing of different gain/sense
resistor combinations with minimal changes to the overall design. Assuming the maximum current
output of the solar panels would be 500mA (based on previous experience with the manufacturer‟s short
circuit current rating) a sense resistance was calculated using the following equation:
                                             Gain  I max (50)(0.5)
                                    Rsense                         0.13                             (2.2)
                                              Vout max       3.3
where Vout max was assumed to be 3.3 V since that would be the maximum voltage the microcontroller
could measure. Although the 50 V/V gain was chosen initially as good compromise between measurable
current range and power losses in the sense resistor, all three available gain levels were ordered along
with a range of sense resistors to allow testing of different combinations.

   2.1.4 Dc/dc Converter Design and Details

The dc/dc converter stepped up the voltage of the solar panel to a level sufficient to charge the battery as
well as acted as the actuator to control the operating voltage of the solar panel. Initially, we planned on
designing and fabricating our own converter and controlling the voltage by pulse width modulation
(PWM) by the microcontroller. This was going to be challenging to accomplish in an efficient and
compact design. Thus, we investigated ways to control the solar panel voltage with an off the shelf
converter. A suitable design was found in [1].

In this design a 1 F capacitor was placed in parallel with the solar panel and the positive side was
connected to the input of the dc/dc converter. The voltage of the solar panel was then controlled by
either enabling or disabling the converter. When the converter was disabled, the current being drawn
through the converter dropped to nearly zero and the solar panel charged the capacitor, raising the
voltage across the solar panel. When the converter was enabled, current was drawn through the
converter and consumed by the load. As long as the power consumed by the load was greater than the
power produced by the solar panel, the capacitor would begin to discharge and the voltage across the
solar panel would drop. The large capacitance was used to provide a damping effect on the changes in
voltage. The design found in [1] used a MAX1675 boost converter, and it was decided to utilize this as it
offered a compact design and high efficiency (up to 94% claimed). Testing of the battery indicated an
output voltage of 4.3 V would be appropriate. The appropriate resistance for the voltage divider on the
feedback pin can be found using Equation (2.3).
                                           V                  4.3
                                  R5  R6[ out  1]  100 K [      1]  230 K                         (2.3)
                                            Vin               1.3
Here, the input voltage was taken to be the lower limit that the converter can handle and still output the
appropriate voltage.

   2.1.5 Microcontroller Design and Details

The target circuit for this power system contained a MSP430 microcontroller, which was to be utilized
to implement the MPPT control. However, a Microchip PIC16F876A was used in the design and
development of the power system due to availability and support provided in the class (examples, etc.).
All code was written in C programming language to allow adaptation to other microcontrollers. The PIC
was able to operate at 3.3 V supply, but only with a slightly reduced clock speed. A FOX F1100E
oscillator rated at 4 MHz was utilized due to availability and it allowed the fastest clock cycle at which
the PIC could operate at 3.3 V. The algorithm implemented in the software was similar to that presented
in [2]. The algorithm in [2] involved taking three power and voltage measurements and comparing the
three readings and making decisions based on this. The algorithm we implemented was very similar to

this. In our algorithm, voltage and power readings from the solar panel were taken every cycle and the
readings from the previous two cycles were stored in registers. If three consecutively lower power
readings were taken, then the solar panel has moved past its peak power point and the controller either
enables or disables the converter based on the three voltage readings. A summary of the code is given

if(POWER_2 > POWER_1)
                if(POWER_3 > POWER_2 // P3 > P2 > P1
                      if(VOLT_1 >= VOLT_2)
                              if(VOLT_2 >= VOLT_3) // V3 < V2 < V1 V rising
                                    output_high(PIN_C0); //enable converter
                      else // V2 > V1
                              if(VOLT_3 > VOLT_2) //V3 > V2 > V1 V dropping
                                    output_low(PIN_C0); //disable converter

Like most MPPT algorithms, ours had the potential to get “tricked” and end up moving towards either
open or short circuit conditions on the solar panel. One particular case in which this may occur was
when the solar panel current dropped below the lower limit that the current sense amplifier could detect.
In this situation the current measured by the microcontroller would remain constant and the controller
would simply maximize the voltage on the solar panel in its effort to maximize power. Thus, the voltage
on the solar panel would rise and the current would drop until the solar panel reached an open circuit
condition and the controller would hold it in this state. In order to prevent this, a timer was implemented
to keep track how long the microcontroller stayed in a particular state (converter enabled or converter
disabled). If the microcontroller spent too much time in a particular state, the microcontroller switched
to the other state and waited for a short period before returning to the normal MPPT algorithm.

   2.1.6 Rechargeable battery

For our design, a rechargeable battery was desired at a voltage level close to desired final output of 3.3V
and the output of the solar panel, which was also about 3.3V. It was also desired to supply power to a
load under low light conditions or during nights. The nominal battery voltage was 3.6V and was rated at
a capacity of 700mAH. This way it required a drop of only 0.3V for the final output during normal
operation. The output of the dc/dc converter was set to 4.3V, which would charge the battery if there
was enough input power, whenever the voltage across the battery dropped to a value lower than 4.3V. It
was a good design threshold and a variable since it protected the battery from overcharging.

The 700mAH was sufficient to provide a maximum load of 50mW plus the internal circuit consumption.
The internal circuit consumption was measured to be 20mA during normal operation with the solar
panels and the battery connected. We had options of choosing a larger mAH capacity for the battery.
However, it would cost more to buy a larger capacity battery. Our design charged the battery at low
light or cloudy conditions if it was empty. Therefore, the only time we needed the sole usage of the
battery was during night time. We can compute the amount of time the 700mAH capacity battery
operates at a maximum load conditions. First step is to determine the maximum current limit at 50mW.

                                       Power 50 x103W
                                   I                       13.88mA                              (2.4)
                                         V         3.6V
Therefore, the 50mW load required 13.88mA of current for operation. The next step is to compute the
amount of time in hours using the rated capacity and the result form Equation (2.4) as well as the
internal circuit consumption.

                                 Capacity           700mAH
                        Time                                      20.66 Hours                    (2.5)
                               Consumption (13.88mA  20mA)
From Equation (2.5), we see that the battery can supply the load and the whole system by itself for
approximately 21 hours straight without needing to be charged. Of course, we assumed 12 hours of
night period for our design. Then during the daytime, even with low light condition the battery could be
charged slowly, while supplying constant power to the circuit and the load until it ran out.

We had options of choosing one of the Li-ion, NiCd, and NiMH rechargeable batteries. We chose the
nickel metal hydride (NiMH) batteries for our design. We found through research that NiMH batteries
were non-toxic compare to the toxic cadmium material, so it was environmentally friendly.
Furthermore, NiMH battery was light in weight and more efficient than the NiCd battery. Li-ion was
also an option however, due to trickle charging of the battery in our design, the Li-ion battery would
cause memory-effects. The memory-effects are damaging in the long run for a battery, if it is trickled
charge often. The NiMH battery can withstand a very high rate of trickle charge and thus has no

   2.1.7 Gas gauge IC

The gas gauge IC was the only option for us in this design to constantly keep track of the available
charge in the battery. The gas gauge IC Bq2012 monitored the charge or discharge activity across a
sense resistor and counted up or down the nominal available charge respectively. The sense resistor was
directly connected between the Vbat- terminal and the ground. The gas gauge IC used an internal
temperature sensor to account for differences in charging and discharging rate at different temperature
steps of 10  C. The gas gauge IC was an important design choice because we had the capability to know
the available charge at any given instant through the internal register NAC. Another benefit of using the
gas gauge IC to monitor the battery in our design was that the end-user has the capability to set a
threshold on when to put the whole system in stand-by mode. For instant, the end-user can decide to put
the system in stand-by mode at 20% of the initial battery charge.

We also had the ability to program the initial capacity of the battery to the gas gauge IC. The value of
the sense resistor used in the design was 0.15  to maintain a small voltage drop across it. The initial
programmed full count, PFC, in mVH was computed using Equation (2.6).

                    PFC  Capacity * Sense Re sistor  700mAH *0.15  105mVH                         (2.6)

PFC in the units of mVH as well as the PFC as the actual number of counts is shown in the look-up table
in the data sheet. Taking the next available PFC(mVH) of 106mVH, the actual programmed counts are
33792. The programming pins were pulled high or low based on this PFC value. When we pulled
programming pin 6 high and connected the fully charged battery to the gas gauge IC, then the initial
condition PFC = NACH was met. Furthermore, as recommended in the application note of the gas
gauge [3], we first fully discharged the battery until it reached the „end of voltage‟ threshold of 1.05V
and then fast charged it using the dc power supply. This way the gas gauge IC was correctly and
properly synchronized with the actual capacity of the battery.

There were several options for the microcontroller to communicate with the gas gauge IC: Empty
output, LED display, and DQ serial communication I/O. A single-cell battery, SB, voltage was
monitored for the purpose of the reset as well as pulling the Empty pin high or low. This was
determined by a resistor divider across pin SB. The calculation of the resistor values chosen are shown
in Equation (2.7).
                                                     n 1  3 1  2                                     (2.7)
                                              Rsb 2
Therefore, Rsb1  2 Rsb 2 . Here, n is the number of cells in the battery. Based on the current limit into pin
SB, the apparent choices were Rsb1  10k  and Rsb 2  5k  . Based on the „end of voltage‟ threshold,
when the voltage Vsb fell below 1.05V, the Empty pin output went high. This method prevented us from
keeping an accurate record of the available charge when the Empty pin was low. Therefore, it was not
our design choice. We also eliminated the LED display function since we wanted to implement the
battery monitoring system for the internal control instead of the user‟s knowledge of the available
charge. The final option was to implement serial communication with the gas gauge IC. The maximum
bit rate at which the gas gauge can communicate was 333 bits/second. The microcontroller first sent a
BREAK signal to the DQ pin on the gas gauge IC and thus pulled it low at least for 3ms. After the
BREAK signal, the DQ pin is driven to a logic high state. The return-to-one data bit frame was used to
transmit 8-bits to the gas gauge IC to read 8-bits in the NACH register in the gas gauge IC. The
communication with the gas gauge IC was always done with the least-significant bit transmitted first. A
MOSFET was used to pull the pin RC7 low on the microcontroller to send the 8-bits to the gas gauge IC
[4]. Then the pull up resistor of value 10k  was used to pull pin RC7 high, so it can receive 8-bits
back from the gas gauge IC as shown in the diagram in Figure 2.2

The values of the programming pull-up and pull-down resistors of 180k  were directly recommended
in the application datasheet for Bq2012 as well as of the RC-filter, 100k  and 0.1  F, before the sense
resistor. A complete diagram for the circuit used in the design is shown in the final circuit schematic in
Figure 2.3.

   2.1.8 Switching Regulator

As mentioned in the performance specifications, the final step of our design was to regulate the final
output voltage to be 3.3V. The first approach was to just use a linear regulator to drop the voltage from
3.6V to 3.3V. However, there was a drawback to this implementation. The voltage drop across the
battery may drop down to 3.0V at 20% remaining capacity, although the actual stand-by threshold was
the choice of an end-user. Therefore, our design needed a buck-boost operation that provided an output
voltage that was smaller and larger than its input voltage. The linear regulator can only provide output
voltage that is smaller than its input voltage.

In detail, the switching regulator TPS61131PW was able to operate at sufficient input voltage range and
convert it to the fixed output of 3.3V. The input voltage range was 2V to 5.5V and the output of 3.3V
provided a maximum 1300mA of current capacity. Our output current at a maximum of 50mW was
about 14mA as computed in Equation (2.4). This output voltage did not only supply power to the load
but also our internal circuit ICs: gas gauge IC and the PIC microcontroller. The general schematic of the
switching regulator is also shown in Figure (2.3). We see that the input comes in as Vbat+. Pin SWP
was the dc/dc rectifying switch input and SWN was the dc/dc switch input. This regulator/converter
required three major passive components for storing energy during the voltage conversion: two 22  H
inductors, a flying capacitor of 10  F, and a storage capacitor at the output. At the output, a ceramic
capacitor of value 22  F with low ESR was desired to minimize the ripple. The output signal can be
enhanced using a larger value of the capacitance or a using a smaller capacitor in parallel to reduce ESR.
                                      3. DESIGN VERIFICATION

This section provides solid technical data to verify system functionality as well as performance
characteristics. Tests were completed to verify solar panel characteristics under various light and load
conditions, implementation of the peak power tracking method, serial communication with the gas
gauge IC and efficiency of individual components including the switching regulator and the dc/dc

3.1 Testing

   3.1.1 Photovoltaic

Given that our initial idea was to determine the constant voltage fraction, k, tests included determining
the voltage at which maximum power occurred at different light conditions. The circuit used to perform
the test can be seen in Figure 3.1 in the appendix. Under a certain light condition, the resistance, R, was
varied and the voltage, V, across the PVs was measured. Given these two values and using the following
two equations, the current, I, and power, P, was determined for each load.
                                                 I                                                     (3.1)
                                                 P                                                     (3.2)
Typical data taken when performing the test can be seen in Table 3.1. Note that in this case VOC was
determined to be 4.33 V. Given the data of Table 3.1, plots of V versus I and V versus P can be seen in
Figure 3.2 (a) and (b).

As can be seen from Figures 3.2, the current is low, approximately 1.7mA, resulting in a very small
power of about 3mW. In addition, given the narrow frequency band of indoor light, the energy
delivered by the PVs was small compare to what we desired. Therefore the final design was
implemented for the use in only outdoor conditions.

Having taken enough measurements of V at which maximum power occurs, Vmax, a plot of VOC versus
Vmax was plotted in order to determine k. The plot can be seen in Figure 3.3 in appendix. Given this
graph, k (the slope) was determined to be 0.74.

After changing our initial ideas, the same test procedure was followed in order to determine general
operating characteristics of the new PVs. The main goal was to determine the range of operating current
in order to obtain the best current amplifier. The range of current at which the PVs would operate ranged
from 0.01 A to 0.1A. As can be seen, the current range was somewhat large, therefore choosing a
current amplifier proved to be very challenging. Typical data obtained using the new solar panels can be
seen in Table 3.2. Given the data of Table 3.2, plots of V versus I and V versus P for these new solar
panels can be seen in Figure 3.4 (a) and (b).

The PVs that we were going to be used in the final design required a test to determine the variance of
VOC under cooler and warmer temperatures. In order to do this, a hot pack and a cold pack were placed
on top of the solar panels for 3 to 4 minutes. After the packs have been removed, VOC was measured and
compared to the original VOC before the packs were placed. Testing showed that when a hot pack was
placed on top of the PVs, VOC was reduced from 2.837 V to 2.685 V. On the other hand, when a cold
pack was placed on top of the solar panels, VOC increased from 2.870 V to 3.035 V. Possible reasons for
these could be the solar insulation and the electrons and holes of the semiconductor. At a higher
temperature the electrons and holes move faster, therefore less voltage I needed for the carriers to move
across the p-n junction. Another possible reason for the effect on the PVs at different temperature was
the internal diode of the solar panels. The current through the diode has strong temperature dependence
as can be seen in Equation 3.3 below.
                                           I d  I 0  eqVd / kT  1                              (3.3)
 Id is the current through the diode, I0 is the saturation current, q is the electron charge (=1.602 x 10-19C),
Vd is the voltage through the diode, k is the Boltzmann‟s constant (=1.381 x 10-23J/K), and T is the
temperature in K.

A final test that was performed with the PVs was the variance of VOC throughout a day. The test results
can be seen in Table 3.3. Given the results of Table 3.3, a plot of the time of the day versus VOC was
plotted in Figure 3.5. We see from Figure 3.5 that VOC increased rapidly when the sun rose while it
decreases rapidly when the sun set. Although data was not taken around noon, it would be expected that
VOC would increase a bit more.

   3.1.2 Current Sense Amplifier

The current sense amplifier was tested to determine the measurable range of currents. This was
accomplished by providing the amplifier with 3.3 V from the Agilent E3631A Power Supply. 3.3 V was
also provided to the positive side of the sense resistor and an adjustable load resistance was placed
between the negative side of the sense resistor and ground. The load resistance was varied and the input
current and amplifier output voltage were measured on Keithley Multimeters. Various combinations of
gain and sense resistor were tried and the combination of 20 V/V gain and 0.2 ohm sense resistor offered
a measurable current range of approximately 0.004 to 0.75 A, resulting in a output voltage range of
approximately 5 mV to 2.3 V, as shown in Figure 3.6. This should be appropriate considering testing of
the solar panel outdoors indicated a maximum output current of approximately 0.8 A, but this was under
short circuit conditions and would not be achieved regularly.

   3.1.3 Dc/dc converter

The dc/dc converter was tested for proper output voltage and efficiency. The converter was provided an
input voltage from the Agilent DC Power Supply and a load resistor was placed between the output of
the converter and ground. The input voltage was varied as the output voltage was measured with the
Keithley multimeter. The converter was able to maintain the desired 4.3 V output down to an input
voltage of approximately 1.4 V. For efficiency, the input current and voltage was measured using the
Keithley multimeter and the output voltage measured across a known load resistance. The input voltage
was varied for a given load resistance and input and output power calculated. Efficiency was then
calculated for each input voltage and the procedure was repeated for different load resistances. The
results revealed an efficiency of around 85% as shown in Figure 3.7.

   3.1.4 Microcontroller

Basic tests were performed for microcontroller functionality(A/D conversion, input/output, etc.). In
addition, a test was performed to determine the measurable voltage range of the Analog to Digital(A/D)
conversion pins of the microcontroller. The microcontroller was set to display the contents of a register
with the results of an A/D sample on LEDs. The voltage on the A/D pin was then varied and the voltage
observed for when the LEDs stopped changing their display. The measurable range was found to be
about 0.174 to 3.17 V. The MPPT algorithm was tested for functionality. The current sense amplifier,
dc/dc converter, and microcontroller were connected together. A solar panel was connected to the input
of the dc/dc converter and a load resistance connected between the converter output and ground. The
solar panel was tested under a lamp and the voltage at which maximum power output occurred was
determined. The MPPT hardware was then tested with the solar panel under the same light conditions.
The voltage across the solar panel with the MPPT implemented was measured using the Agilent 54642A
oscilloscope and the data collected was loaded from the scope to Excel. The average of the voltage
readings was then taken and compared to the previously determined maximum power voltage. This was
repeated for three different light conditions. The results are shown in Table 3.4. A typical waveform is
shown in Figure 3.8. There was a somewhat large variation in the voltage as the converter was enabled
and disabled. This was probably due to the converter pulling a somewhat large amount of current right at
start up.

   3.1.5 Battery Charging and Discharging Characteristics

Basic tests were done on the battery to determine the charging and discharging characteristics. The
charging characteristic was not critical for us to know since the battery charged at any given current.
We utilized a phone charger for the battery to determine the voltage that the battery reached when the
charging was completed through the internal phone charge control. This value of the voltage was about
4.3V. Therefore, we decided to use the 4.3V as the output from the dc/dc converter.

The discharge characteristics were important since we needed to know how long the battery can supply
power with a full charge and how the voltage varies across various load conditions. Taking 3-cell
individually and charge-discharge it across various loads would take a long time. Therefore, to save
time we used individual cells and connected each across one of three different resistors. This would
discharge the battery with respect to the current that each resistor drew. The measured discharging
characteristic is shown in Figure 3.10. We see that up to 80% of the charge capacity, the voltage
remained approximately constant to about 1.1V and thus for three cell it would be 3.3V.

   3.1.6 Efficiency for Switching Regulator

To determine the efficiency for the switching regulator, we connected three different resistors across the
output. The dc power supply was used as the input to the switching regulator and the KEITHLEY
multimeter was used to measure the input current. We computed the input power using the input voltage
times the input current. The output voltage was measured across a fixed resistor while changing the
input voltage. The output power was computed using the output voltage and the resistor. Then, the
efficiency was computed as the ratio of output power to input power. The test repeats for two other
values of the resistors. The efficiency as we see in Table 3.5 was on average 85%. The curve for
efficiency for each load versus the input voltage is shown in Figure 3.9.

3.2 Conclusions

Once all subcomponents were tested for functionality and performance, the entire system was integrated
and testing started on overall system function and performance. Testing was performed outdoors by
taking an oscilloscope outside with an extension cord and utilizing outside power outlets. Functionality
was tested by confirming that the tracking system was performing appropriately and not getting stuck by
observing the solar panel voltage, current and power on the oscilloscope. This was performed at varying
times of the day and under different conditions (cloud cover, etc.).

Once basic functionality was confirmed, attempts were made to quantify any power advantages or
disadvantages that power tracking offered versus just leaving the dc/dc converter enabled. This proved
difficult as we were unable to download actual measurement numbers from the oscilloscope without
having connected to a computer with the appropriate software, which we did not have in a portable
package. We discovered that there was probably no advantage to MPPT in full sunlight conditions at
mid-day (11:00 – 2:00), as the system (load and battery) could not absorb all the power available and
thus it was impossible to achieve peak power conditions. Tests were performed indoors under a lamp
while measuring the dc/dc converter output voltage and output current with a current sense amplifier.
The data for these measurements indicated a small reverse current from the battery to the converter
when the converter was disabled, resulting in a lower output power for the MPPT configuration. These
losses may be offset by potential power gains MPPT may offer in an outdoor setting, but we were
unable to confirm this conclusively. However, the system was able to provide the desired power levels
as our test circuit actually consumed roughly 50mW driving the troubleshooting LEDs that we used.
This was using either MPPT or just leaving the converter enabled at all times.

                                                 4. COST

This project contained five ICs as well as the external resistors, diodes, capacitors, and inductors and a
MOSFET. The five ICs include the microcontroller, the current sense amplifier, dc/dc converter, the gas
gauge IC and the switching regulator. The external components are shown in the final schematic Figure

4.1 Parts

Some parts of the final circuit required specific types of inductors and a MOSFET. The Sumida CR54-
220 inductor used was a specific coil type inductor required for the dc/dc converter. The DR74 inductor
was a cross-coupled inductor desired for the switching regulator operation. We tabulated the individual
component price in Table 4.1. The total unit cost was $100.84. However, approximately $20 dollars
were spent on the other external components such as 17 resistors and 9 capacitors, which were gathered
from the electronic shop.

                          Table 4.1 – Major Parts and the Whole Unit Costs

                             Part               Price ($)           Description
                          MAX1675                 5.06            dc/dc converter
                          MAX4173                 1.88        current sense amplifier
                         PIC16F876A               4.71                controller
                      Sumida CR54-220             1.17                 inductor
                Cooper Bussmann PB-5R0H474        5.13                capacitor
                           MPF102                 0.37                MOSFET
                        HD01DICT-ND                0.7                   diode
                          13FR200E                1.55             sense resistor
                Edmund Scientifics #3039811(x6)   53.7              solar panels
                           Bq2012                 3.81    gas gauge IC to monitor battery
                        TPS61131PW                4.61          switching Regulator
                            DR74                  2.19                 inductor
                   SNN5542 Battery Pack          15.96    3-cell 3.6 V NiMH battery pack
               TOTAL                               100.84        Single Unit Price

4.2 Labor

The desired ideal hourly rate for each electrical engineer at this level would be about $25/hr. Each
individual spent 20 hours per week including the shop and machine shop hours and thus contributing to
the labor cost as shown below.

Individual labor cost = ideal salary x actual hours spent x 2.5
Ideal salary = 25/hr
Individual hours spent = 20hr/week * 16 weeks = 320 hours

Individual labor cost = $25/hr x 320 hr x 2.5 = $20,000/individual

Total = 3 individuals x $20,000 = $60,000

                                          5. CONCLUSIONS

Our design successfully functioned and met the performance specifications. The circuit provided
necessary power output at 3.3V as well as monitored the battery charge constantly for the purpose of
internal control. The design of six solar panels in series provided enough power for load and internal
circuit consumption as well as was able to charge the battery when it was more or less discharged.
When the solar panel did not provide enough power, the battery operated for about 21 hours at
maximum load condition alone. Finally, at a very unlike case, if the battery ran out, the system was also
able to go into a stand-by mode.

One of the uncertainties in this design was that we were unable to perform outdoor tests on solar panel
to actually verify the significance of the peak power tracking method. Our design could improve the
overall efficiency by using another type of solar panel. The semiconductor material GaAs used in solar
panels provide higher efficiency and larger power output per one square meter of solar panel at room
temperature compare to the lower efficiency and lower power output of the Si material used in solar
panels. However, GaAs solar panels are mostly used in satellite applications, and they are not
commercially available for this type of low power system design.

The future work would involve verifying and modifying the use of the peak power tracking method,
mostly in outdoor sun-light conditions. This would make the MPPT unnecessary. Outdoor testing
indicated that during high sunlight hours of the day, the solar panel configuration used produced more
power than the system (load and battery) could absorb, suggesting the solar panel could be reduced in
size and cost. The MPPT configuration could be improved by designing a dc/dc converter that would
always keep the output voltage higher than the battery voltage, thereby eliminating reverse currents and
associated losses.

                          APPENDIX – TITLE

        Figure 1.1 - General Block Diagram and Interconnections

            Figure 2.1 – Series Connection of Six Solar Panels

Figure 2.2 – Serial Communication Implementation Circuit on PIC16F876A
                                                                                                                        + Solar panel
                                                                                                                                          Vcc = 3.3V

                                                               1       Vcc                    RS+       8

                                                               2       N.C.          N.C.               7
                                                                          MAX4173FESA                   0.2
                                                               3       GND           RS-                6                                             0.1uF
                                                                                                                                                              1        (MCLR)'                     RB7       28
                                                               4       OUT                    N.C. 5
                                                                                                                                                      10k     2        AN0                         RB6       27

                                                                                                                                                              3        AN1                         RB5       26
                                                                                                                                                              4        -Vref                       RB4       25

                                                                                                                                                              5        +Vref                       RB3       24
                                                                                                                                                              6        C1OUT                       RB2       23
                                                                                                                                                              7        C2OUT                       RB1       22     Vcc = 3.3V

                                                                                                                                                              8        Vss                         RB0       21

                                                                                                                                                              9        CLKI                        VD0       20
                                                                                                              1       14       8
                                                                                                                       FOXF1100F                              10       CLKO                        Vss       19                  10k
                                                                                                                  22uH1        7                              11       RC0                         RC7       18
                                                                                                                                                                                                                                   Gas gauge pin-13
                                                                                                                                                              12       RC1                         RC6       17
                                                                                                              2                                               13       RC2                         RC5       16
                                                           1       FB                    OUT        8
                                                                                                                                                              14       RC3                         RC4       15
                                                           2       LBI      MAX1675          LX     7                              47uF

                                                           3       (LBO)'                GND        6                                                                                               1                  2               1
                                                                                                                                                                                                                                           SWD    TPS61131PW   Vout    16
                                                           4       REF                 (SHDN)' 5                                                                                                                           10uF
                                                                                                                                                                                                                                       2 SWN                    FB     15             100uF
                                                   0.1uF                                                                                                                                                 2
                                                                                                                                                                                                                                       3 PGND                  PGood   14

                                                                                       Lcom         Bq2012           Vcc    16                                                                           22uH                          4 Vbat                   LBO    13

     Figure 2.3 – Final Schematic
                                                                                                                                                 Vcc = 3.3V
                                                                   1         2       2 Seg1                          Ref    15       0.1uF                                                                                             5 LB1                    GND    12

                                                                   1         2       3 Seg2                        (CHG)' 14                                                                             1                             6 SKIPEN                LDOin   11
                                                                   1         2       4 Seg3                           DQ    13                                                                               10uF      1M              7 EN                  LDOsense 10
                                                                                                                                          PIC pin-18
                                                                       1         2 5 Seg4                           Empty 12                                                                                                           8 LDGEN               LDOout    9
                                                                                                                                                                   + Vbat
                                                       470                                                                                      10k
                                                                       1         2 6 Seg5                             SB    11                                                                                         390k
                                                                   1         2       7 Seg6                       (DISP)' 10            0.1uF
                                                                                     8 Vss                            SR9
                                                                                                                                                              - Vbat
                                                                                                                                                      0.15                                 Title
                                                                                 0                                                                                                                       Fianal Schematic - battery and solar panel not shown

                                                                                                                                                                                           Size          Document Number                                                        Rev
                                                                                                                                                                                              A          Doc 1                                                                    1
                                                           Vcc = 3.3V
                                                                                                                                                                                           Date:             Saturday , April 29, 2006               Sheet      1      of   1
                                                                                    Figure 3.1: PV test circuit

                                                                                     Table 3.1: PV Test Data

                                                  Resistance (kΩ) Voltage (V) Current (A) Power (W)
                                                       0.120         1.19       0.0099      0.012
                                                       0.219         2.19       0.0100      0.022
                                                       0.275         2.65       0.0096      0.026
                                                       0.327         3.13       0.0096      0.030
                                                       0.466         3.67       0.0079      0.029
                                                       0.908         4.05       0.0045      0.018
                           -3                                                                                                      -3
                    x 10                  V versus I with a Lamp                                                            x 10                        V versus P with a Lamp
              1.8                                                                                                      3

Current (A)

                                                                                                          Power (W)

               1                                                                                                       2


              0.2                                                                                                      1
                1.6             1.8   2   2.2       2.4     2.6               2.8      3      3.2                      1.6              1.8        2        2.2       2.4     2.6   2.8   3   3.2
                                                Voltage (V)                                                                                                       Voltage (V)

                      Figure 3.2(a): V versus I of PV                                                                                         Figure 3.2(b): V versus P of PV
                                                                                                Determining k

                                                                         y = 0.74*x - 0.094

                                                   Vmax (V)



                                                                                                                                               linear fit

                                                                     0       0.5        1       1.5       2                   2.5              3            3.5
                                                                                                  Voc (V)

                                                                                    Figure 3.3: VOC versus Vmax

                                                          Table 3.2: Typical test data

                                        Resistance (kΩ) Voltage (V) Current (A) Power (W)
                                                1         0.654        0.654      0.428
                                               1.8        1.155        0.642      0.741
                                               3.6        2.268        0.630      1.429
                                               4.7        2.589        0.551      1.426
                                               6.8        2.806        0.413      1.158
                                               8.2        2.925        0.357      1.043
                                               11         3.021        0.275      0.830
                                              12.2        3.045        0.250      0.760
                                              16.7        3.121        0.187      0.583
                                               18         3.130        0.174      0.544
                                              20.2        3.150        0.156      0.491
                                              29.9          3.2      0.107023 0.342475
                                                  33.3           3.217                       0.096607     0.310783
                             V versus I at Voc = 3.301V                                                  P versus V at Voc = 3.301V
              0.7                                                                    1.6

              0.6                                                                    1.4

              0.5                                                                    1.2
Voltage (V)

              0.4                                                                     1
                                                                         Power (W)

              0.3                                                                    0.8

              0.2                                                                    0.6

              0.1                                                                    0.4

               0                                                                     0.2
               0.5    1      1.5         2          2.5     3      3.5                 0.5       1       1.5         2          2.5   3   3.5
                                    Current (A)                                                                 Voltage (V)

                     Figure 3.4(a): V versus I of PV                                                 Figure 3.4(b): V versus P of PV

                                              Table 3.3 - Variance of VOC throughout a day

                                                                Time                   Voc (V)
                                                                 6.45                   2.827
                                                                 7.16                   3.081
                                                                 7.45                   3.205
                                                                 7.55                   3.242
                                                                 9.13                   3.402
                                                                 14.1                   3.410
                                                                14.43                   3.423
                                                                 15.2                   3.430
                                                                  16                    3.390
                                                                 16.3                   3.380
                                                                17.03                   3.240
                                                                17.35                   3.000
                                                                17.55                   2.770









                                               6               8            10            12              14          16             18

              Figure 3.5: Time versus VOC Corresponding Table 3.3

                                                                   MAX4173 w/ Rsense = 0.2 ohms


                      Vout (V)




                                       0                 0.5            1              1.5            2          2.5             3
                                                                                   Current (A)

Figure 3.6 - Output Voltage vs. Current Curve Showing Saturation at High Currents




                                                                                                                   R=      100
                                 0.7                                                                               R=      90
                                                                                                                   R=      68
                                                                                                                   R=      48

                                1.4                1.6     1.8      2            2.2    2.4     2.6        2.8    3        3.2       3.4
                                                                                 Input Voltage (V)
                                 Figure 3.7 – Efficiency for the Dc/dc Converter
                          Table 3.4 – MPPT Algorithm Verification Values

                Voc                                        Vmax power                                         Vavg                          % error
               2.25                                         1.76                                              1.84                             4.5
               1.96                                         1.51                                              1.65                             9.3
               1.71                                         1.28                                              1.33                             3.9

                                                                                     Current Signal

                                                                                                                 V*I             Voltage

               Figure 3.8 - Typical Solar Panel Characteristics for MPPT Algorithm

       Table 3.5 – Measured Values to Compute Average Efficiency for Switching Regulator
    Input Voltage (V)      3.6                         3.8              4            3.6               3.8          4             3.6       3.8        4
   Input Current (mA)     84.9                        82.8            80.3          76.06             71.95       68.25          65.46     61.8      58.4
         Pin (W)          0.306                       0.315           0.321         0.274             0.273       0.273          0.236     0.235     0.234

  Output Voltage (V)      3.15                         3.2            3.24          3.33              3.33        3.33           3.33      3.33       3.33
measured resistance (  ) 38.7                        38.7            38.7          47.6              47.6        47.6           55.2      55.2       55.2
       Pout (W)           0.256                       0.265           0.271         0.233Average      0.233       0.233          0.201     0.201     0.201
       Efficiency         83.888 84.096 84.451 85.079 85.205 85.333 85.245 85.542 85.996 84.982
                                                    Efficiency versus Input Voltage for the Switching Regulator TPS61131PW
                                                            56 ohms
                                                            47 ohms
                                                            39 ohms
                            Efficiency (%)




                                                3.6       3.65     3.7      3.75      3.8      3.85     3.9     3.95         4
                                                                               Input Voltage (V)

                      Figure 3.9 – Efficiency Curves for the Switching Regulator
                                          Battery Voltage Discharge Vs. Time




        Voltage (V)
                      0.8                                                              V1 w / 2.3Ohms
                                                                                       V2 w / 5 ohms
                      0.6                                                              V3 w / 3.4 ohms


















                                               Time (mins)

                      Figure 3.10 – Battery Discharge Characteristics

 Figure 3.11 – Serial Communication Example with Gas Gauge IC

Figure 3.12 – Load and Input Variation Tests on Switching Regulator


[1]    J. Jordan, “Design and Implementation of a Stochastic
Wireless Sensor Network,” M.S. thesis, University of Illinois, Champaign, IL, United States of America,

[2]   J.A Jiang, T.L. Huang, Y.T. Hsiao, and C.H. Chen, “Maximum Power Tracking for Photovoltaic
Power Systems” Tamkang Journal of Science and Engineering, vol. 8, No 2, pp. 147-153, 2005.

[3]   “Using the bq2010 – Tutorial for gas gauging,” Texas Instruments Incorporated, U-541, 1990

[4]   “HDQ Communication Basics for TI’s Battery Monitor ICs” Texas Instruments,
       Application Report, SLVA101, May 2001.


To top