SELF-SUSTAINABLE SOLAR STREET LIGHT CHARGING

Transcription

1 SELF-SUSTAINABLE SOLAR STREET LIGHT CHARGING By Anirban Banerjee Priya Mehta Surya Teja Tadigadapa Final Report for ECE 445, Senior Design, Fall 2017 TA: Zipeng Wang December 2017 Project No. 4

2 Abstract This paper constitutes the final report for the Self-Sustainable Solar Street Light Charging project undertaken by Team 4 for the Fall 2017 session of ECE 445: Senior Design Laboratory. The project focuses on the charging circuit required for the street light, and is intended to serve as one of three subsystems required for the complete, functional streetlight (the other two subsystems, discharging and user monitoring, have been undertaken by different teams). This report explains in detail the Maximum Power Point Tracking (MPPT) approach used as the basis for the charging circuit, the different circuit designs examined by our team including the final implementation actually used for the project, and the performance of the different modules in the circuit during testing.

3 TABLE OF CONTENTS 1. Introduction Idea Project Overview Design Block Diagram Design Progression Design I Analog Design II Atmega328p Design III C Physical Design Buck Converter MCU Voltage Monitor Current Monitor Battery Charger Component Equations Code Algorithm Implementation Requirements and Verification Buck Converter MCU Voltage Monitor Current Monitor Battery Charging Costs Labour Costs Parts Cost Conclusion Accomplishments Uncertainties Future Work and Improvements Ethical and Safety Considerations References...16 Appendix A: R&V Table.17

4 1. Introduction 1.1 Idea The goal of our project was to design and create the charging circuit for a portable, self contained, self sustaining and user friendly streetlight that could be set up anywhere with minimal effort, independent of the power grid, and would function similarly to a commercial streetlight in providing light during periods of darkness. The full design of the streetlight would consist of three subsystems: a charging circuit (solar panel to battery), a discharging circuit (battery to bulb) and also feature some user monitoring (web/android app that gives details on state of charge and charging efficiency). These subsystems and their interconnections have been shown in Figure 1 below. The focus of our project was be the MPPT solar charge controller block present between the solar panel and lead acid battery in the charging circuit (shown as green in the Figure 1). Figure 1: Block diagram for solar street light 1.2 Project Overview In order to fulfil the self contained and self sustaining requirement of the streetlight, the efficiency of the charging circuit needed to be maximised. Solar panels have IV characteristics that differ from those of most other electronics, and supply varying levels of current and voltage depending on the level of sunlight that falls on the panel s surface. This causes the power output from solar panels to be very inefficient if directly supplied to a battery without an appropriate conversion process. Our project focused on creating an efficient charging system that will maximise the amount of power delivered to the battery from the solar panel. This was achieved by using a method known as Maximum Power Point Tracking (MPPT), which has been discussed in greater detail in section below. Our high level requirement for this charging circuit was that it should charge the lead acid battery at 80% efficiency. 1

5 2. Design 2.1 Block Diagram Figure 2: Block diagram of actual charge controller Figure 2 shows the different modules present within the charging circuit, and can be thought of a zoomed in version of the MPPT solar charge controller block from Figure 1. The electricity generated by the solar panel is passed through a buck converter before being output to the lead acid battery. The input from the solar panel and the output to the battery is both monitored for current and voltage. These readings are sent to a microcontroller unit that operates the buck converter so as to implement MPPT (discussed in more detail in section 2.3.2). 2.2 Design Progression Design I Analog Our first major pivot was changing from creating a chain of systems to creating just the MPPT solar charge controller. Since then during the design review we had decided to build an analog MPPT Solar Charge Controller. After considerable research on the working of MPPT, we decided to move away from this method for several reasons: firstly, it would have required us to hand wind our own inductor, an extremely difficult and time consuming task. Secondly, would have been be very difficult to debug the circuit if every component did not work flawlessly. Thirdly, the circuit was designed assuming impractical ideal case scenarios. Lastly, soldering and removing components every time we want to change something would have make it a very labour intensive process. Figure 3 below shows the schematic for the analogue MPPT charge controller. 2

6 2.2.2 Design II Atmega328p Figure 3: Analog MPPT circuit We then moved to a digital design. We decided that in the interest of points and ease of design, it was more suitable for us to use a microcontroller to power our PWM signals instead of relying on analog components. The main addition to this phase outside of the microcontroller itself was the use of the INA 270 current monitor by TI. Finding this component made our lives significantly easier. It allowed us to capture both voltage and current readings, using which we could capture readings both at the start and end of the circuit and feed them into our microcontroller to drive the PWM signals accordingly. The INA 270 used an I2C based digital connection to communicate with the MCU (atmel atmega328p). After meeting with our TA, we that this design had a few downsides. Firstly, the atmega was not the appropriate choice of MCU for driving PWM signals. Secondly, the atmega chip did not allow us to debug in real time if we needed to. Despite ultimately abandoning this implementation, it was still a very crucial stage in our design process since it gave us a solid understanding of the MPPT circuit and what we were trying to build. We also found a majority of our components and established key requirements that we needed to achieve in order to build our circuit. Figure 4 below shows the schematic for the Atmega328p based MPPT Charge Controller. 3

7 Figure 4: Atmega328p Circuit Design III C200 We eventually settled on the C2000 by Texas Instruments as the MCU for our MPPT Solar Charge controller. The C2000 is well suited to the production of fast PWM signals to run the mosfets in our circuits. Another major advantage offered by it is the ability live debug our circuit. We use header pins for communication between the PCB and MCU and banana cables for communication between the solar panel PCB and battery PCB. One major change it caused, however, was that it was not very well suited to handle I2C connections and this rendered our previous current monitor useless in this schematic. While the microcontroller has I2C ports it requires a considerable amount of complex wrapper code written on top of it to be able to communicate properly with our current monitor. This led us to change our current monitor to INA 138 that sent out analog connections. These analog connections are then deciphered by the ADC on the Microcontroller. We also realized we lacked gate drivers and made important adjustments to our proposed PCB design after making the switch to the C2000. Figure 5 below shows the schematic for this design. Figure 5: Actual eagle schematic of the circuit with C

8 2.3 Physical Design Energy is produced by the solar panel which is rated at 100W and a Voc of 22.5V. The current produced by the solar panel is directed via a banana cable to the PCB and through two diodes connected in parallel and rated at 3.5A each Buck Converter Our buck converter, seen in Figure 6 below, consists of the input capacitors, gate driver, p mosfet, n mosfet, inductor and output capacitors. The two mosfets together replace the diode that is part of a usual buck converter. The input capacitance is rated at about 330uF. The p mosfet and n mosfet (switches) get their signals from a gate driver that amplifies the PWM signals coming from the MCU. The current then moves through the inductor which is rated at 47uH before going through the output capacitance rated at 22uF MCU Figure 6: Buck converter schematic We used the C2000 Launchpad with the Piccolo TMS320F28027 chip, pictured in Figure 7 below. The advantages for using the launchpad instead of just the MCU chip are that it allows for easy testing and live debugging. The MCU is powered by the battery and its primary responsibility is to run the MPPT code that generates PWM signals and take in the monitor readings. We utilize two epwm ports to output the PWM signals and utilize four ADC ports to read the input and output current and voltage readings for the MPPT code. Figure 7: C2000 Launchpad 5

9 2.3.3 Voltage Monitor Voltage is monitored before and after the buck converter, for gauging the efficiency of the buck converter. Voltage monitoring is important as it allows the MPPT algorithm to function, which relies on input voltage and current to modify duty cycles accordingly and also because it is supplied to the team that is responsible for the user monitoring app. The voltage monitoring circuit is shown below in Figure 8. It relies on voltage division and then sends this output to an op amp used for stabilizing. This final output, after passing some additional components, is sent to the ADC pin of the microcontroller Current Monitor Figure 8: Voltage monitor schematic Current is monitored for similar reasons to the voltage monitoring: for MPPT and for supplying to the user monitoring app. The current monitoring circuit is shown in Figure 9 below. A current monitoring chip, the INA138, is utilized for this purpose. The INA138 takes in a current and converts it to voltage using internal resistance in the chip. The voltage goes through an op amp, and, after passing some additional components, is sent to the ADC pin of the microcontroller. Figure 9: Current monitor schematic 6

10 2.3.5 Battery Charger A crucial component of the circuit was the charger IC that connected the output of the buck converter to the battery. The reason we had to eliminate this component on the hardware level was that it caused too big a voltage drop to be practical in our case. Since our design called for us to control the duty cycle of the of the mosfets, we decided to use our code to control the charging of the battery instead. The website of the charge controller by TI UC3906 (one of the most common lead acid battery charge controller) recommends a npn darlington transistor array when we actually need a pnp. Not having caught this early on lead us to design with the NPN transistor array, but on realising that the PNP transistor array causes a huge voltage drop we abandoned this route Component Equations Many of our passive component values were derived from datasheets of the ICs in our circuit. Some had to be derived using fundamental equations: Input voltage monitoring derivation: Output voltage monitoring derivation: R1 22.5R1+R2 2 R (R1 + R 2) 0.911R R R1 R2 (1) R3 15.6R3+R4 2 R (R3 + R 4) 0.872R R4 6.81R3 R4 (2) Based on these derivations and taking base values of 10kΩ for R2 and R4 gives us a value of 102.4kΩ for R1 and 68.1kΩ for R3. Buck Converter Equations: L = (V output (V input V output ))/(I ripple * V input * Switching Frequency) (3) D = V output * Efficiency/V input (4) P input = V input * I input (5) P output = V output * I output (6) Efficiency = (P output / P input ) * 100 (7) 7

11 2.4 Code Algorithm Maximum Power Point Tracking (MPPT) is an algorithm that will find, within the given conditions, the maximum power from a solar panel. The maximum power point is the voltage at which the solar panel produces this maximum power and varies with solar insolation and temperature. Using MPPT allows for the maximum current from a solar panel to be acquired without changing the voltage of the solar panel. A buck converter is utilized for eventually modifying the input current into the battery. The Perturb & Observe method allows us to modify the operating voltage and operating current of the solar panel until the maximum power is acquired. The MPPT charge controller compares the output of the solar panels with the battery voltage. After this, it calculates the best power that the panel can produce to charge the battery and then converts this into the best voltage to get maximum current into the battery. Figure 10 shows the IV curve of the solar panel with the maximum power point. Figure 10: IV curve of solar panel and the Maximum Power Point 8

12 2.4.2 Implementation We implemented the MPPT algorithm through the Perturb and Observe approach in code (Figure 11 below shows a flowchart of how this algorithm works). We had many functions, including several initialization functions that set up the microcontroller properly. Our main functions were Data_Update() and Adj_PWM(). Data_Update() is responsible for retrieving the values from the ADC pins of the microcontroller and accordingly calculating the new power in. In particular, it retrieves input voltage and input current. Adj_PWM() is the function that modifies duty cycles according to the calculated power, as per the MPPT algorithm. Figure 11 shows the flowchart of the MPPT algorithm. Figure 11: MPPT algorithm flowchart 9

13 3. Requirements and Verification A summary of R&V is provided in this table below. The complete list of the requirements and verifications has been attached to this document in Appendix A and individual results in the sections below. Table 1: Requirements and verifications summary Buck Converter MCU Voltage Monitor Current Monitor Battery Charging Buck converter fully functional at low power levels but untested at high power levels. MCU generates required PWM signals and MPPT Code fully functional. Voltage monitoring fully functional. Current monitoring inaccurate because conversion factor needs to be tuned. Battery charging has not been tested. 3.1 Buck Converter When the gate driver is supplied with the same input voltage as the power trace (solar panel output) it amplifies the input PWM signals to the required gate voltages to switch the mosfets. Both the P Mosfet and N Mosfet switched according to the PWM signals sent to them from the gate driver. The output voltage was stepped down from the input voltage according to the duty cycles of the input PWM signals. The buck converter was unit tested at low power levels we were unable to test them at high power levels since our PCB lacked ground plains. Figure 12 shows a plot of the buck converter efficiency and demonstrates it increasing with power. Figure 12: Input Power vs Efficiency 10

14 3.2 MCU In order to test the functionality of the microcontroller code, we connected the C2000 to an oscilloscope and were able to monitor generated PWMs. The MCU successfully generated PWM signals which could be tracked using an oscilloscope (an example can be seen in Figure 13). The MCU was also able to calculate the maximum power point based on input readings and was able to change duty cycle as desired. Figure 13: PWM signals generated by MCU Code. A Top: P Mosfet (50% duty cycle); Bottom: N Mosfet (25% duty cycle). 3.3 Voltage Monitor Our voltage monitoring was fully functional, and close to 2% error on average. As seen in Figure 14 below, there is a very linear relationship between the two axes. The equation on the right is what we used to successfully convert an ADC reading to a value close to the input voltage. In particular, two conversions were needed: one to convert the ADC reading to a digital code, and then one to convert the digital code to a value near the input voltage. The voltage monitoring having a low error value allowed us to have a more efficient MPPT implementation since voltage was utilized in calculations. Figure 14: ADC Results vs Input Voltage 11

15 3.4 Current Monitor Unfortunately, our current monitoring was not functional by the time of the demo. This could have been due to the actual current monitor or op amp not working as expected, but more likely could have been caused by and incorrect conversion from the ADC reading to a current value. Attempting to use a conversion similar to the ones that we successfully used for voltage monitoring gave us a periodic graph that is likely sinusoidal. The graph should have been linear, allowing for a straight mapping between the reading and the converted voltage, which would convert to current. Figure 15 below shows our data for ADC readings vs input supply. Figure 15: ADC Reading vs Input Supply 3.5 Battery Charging In order to actually use our circuit to charge the battery, we would have first needed to completely integrate the three circuit modules (buck converter, current monitoring, and voltage monitoring). Since we were unable to complete integration in time, we were unable to test battery charging using our PCB. 12

16 4. Costs 4.1 Labour Costs Table 2: Labour Costs Labourer Salary/Hour Hours Anirban Banerjee $ Priya Mehta $ Surya Teja Tadigadapa $ Parts Cost Table 3: Parts Cost Description Manufacturer Quantity Cost Solar Panel Renogy 1 $ Lead Acid Battery Renogy 1 $ Capacitors Various 19 $1.05 (Avg) Current Monitors TI 2 $1.78 Diodes Diodes Inc. 2 $0.53 Gate Driver IXYS IC Division 1 $1.89 Inductor Bourns Inc. 1 $7.58 LDO Analog Devices Inc. 1 $3.09 P Channel MOSFET Diodes Inc. 1 $0.67 N Channel MOSFET Vishay Siliconix 1 $1.70 Op Amps TI 4 $2.54 Resistors Various 15 $1.30 (Avg) PCB & Stencil PCBWay 1 $30 Total Cost = Total Parts Cost + Total Labor Cost = $ $ = $

17 5. Conclusion 5.1 Accomplishments We were able to successfully complete some of the major components of our project: microcontroller generation of PWMs via implementation of the MPPT algorithm in code, the buck converter, and the voltage monitoring. The MPPT algorithm successfully generated PWMs that we were able to monitor in an oscilloscope. We were also able to modify the duty cycles, generating different PWMs. The buck converter was fully functional at low power levels and was a major success in the scope of the project: without it, it is unlikely we would have had any meaningful implementation of MPPT as a whole. Our voltage monitor was able to retrieve a correct voltage reading. This was important as it allowed us to have accuracy when calculating input power to modify duty cycles, allowing for efficiency in the Perturb and Observe approach of MPPT. In all, we were happy with our accomplishments in this project. We are computer engineers with very little hardware and power electronics experience. In the beginning of the semester, none of us knew much about solar panel and battery usage, and indeed had never as much as heard of the term MPPT. It was only through this project that we were able to gain exposure to the working of a solar charging circuit, the functionality of MPPT, hardware testing, PCB design, and the use of tools such as the hot air gun for solder paste. We have most definitely learned a great deal from this project and are proud of everything we have achieved. 5.2 Uncertainties The main potential source of error in our circuit is noise in the current and voltage monitors which can cause significant errors with the MPPT algorithm. The internal resistance of battery increases with age and this can cause errors in determining state of charge which affects our charging sequence. If the system is operated at high ambient temperatures, our components may either burn out or have different electrical characteristics from specifications. 5.3 Future Work and Improvements We had several major components of this project working individually but we were not able to integrate the components in the time frame of this project. Next semester, another team may be able to start where we left off in the integration. 14

18 The first step the team would have to take is to make some small changes that have great effect in the success of the Solar Street Light project. The current monitoring conversion, as mentioned earlier, needs to be fixed. Secondly, it is best if a new PCB with ground planes is ordered. The PCB layout is already created; it just needs to be reordered, making sure the files sent to PCBWay and the PCB sent back from PCBWay contain ground planes. This will enable high currents and voltage levels to be tested, bringing the project to the level of a final product. The next step would be for this team to integrate our working components on our PCB and make all the components (buck converter, MPPT algorithm, monitoring) work together. This would take a great deal of time, as unexpected errors can always occur when bringing together sections that were only tested separately. Once this is completed, the charging circuit needs to be connected to the specific solar panel and battery Professor Reinhard had purchased. Once this is done, it can be confirmed that the charging circuit has completed functionality. The final step would involve connecting the charging circuit to the remaining subsystems: the discharging circuit and user monitoring app. Once this is done, we will have a final, working Solar Street Light. We have kept documentation and videos for a future team to follow when picking up our project and integrating our features. 5.4 Ethical and Safety Considerations Of all the components in our project, the lead acid battery has the greatest number of safety considerations. Lead acid batteries are contain lead, a toxic metal, and sulphuric acid, a corrosive agent; and hance must be disposed of carefully. If lead acid batteries are discharged too much or too fast, they can potentially be damaged. Certain temperatures and overcharging also cause lead acid batteries to fail, requiring preventive measures to be taken. In some older models, extreme conditions can even result in the battery exploding. Lead acid battery leakage has the potential to cause serious damage clothing and human skin, as well as other components such as wiring. Hence proper care needs to be taken when dealing with the battery. Another issue is that of the weather effects on our streetlight. It is intended for the streetlight to be self sustainable and be placed outside without being attached to another device. If the attachment of the solar panels and components on the streetlight are not housed properly, it is plausible that high winds can loosen or even completely detach these solar panels or components and cause damage to nearby items or people (after all, the streetlight is expected to function in populated areas with people nearby). We must keep IEEE Ethics Code #1 ( to accept responsibility in making decisions consistent with the safety, health, and welfare of the public... ) and ACM Code #1.2 in mind as our streetlight can affect people s safety if not constructed properly. 15

19 5. References 1) Coder Tronics. (2017). C2000 Solar MPPT tutorial covering the electronics & C code. [online] Available at: tronics.com/c2000 solar mppt tutorial pt1 [Accessed 13 Dec. 2017]. 2) Coder Tronics. (2017). C2000 Solar MPPT tutorial covering the electronics & C code. [online] Available at: tronics.com/c2000 solar mppt tutorial pt2 [Accessed 13 Dec. 2017]. 3) Coder Tronics. (2017). C2000 Solar MPPT tutorial covering the electronics & C code. [online] Available at: tronics.com/c2000 solar mppt tutorial pt3 [Accessed 13 Dec. 2017]. 4) Coder Tronics. (2017). C2000 Solar MPPT tutorial covering the electronics & C code. [online] Available at: tronics.com/c2000 solar mppt tutorial pt4 [Accessed 13 Dec. 2017]. 5) Electronic Circuit Projects. (2017). Simple Solar MPPT Circuit. [online] Available at: circuits.com/simple solar mppt circuit using ic555/ [Accessed 13 Dec. 2017]. 6) Acm.org. (2017). ACM Code of Ethics and Professional Conduct. [online] Available at: acm/acm code of ethics and professional conduct [Accessed 20 Sep. 2017]. 7) ECE 445, (2017). [online] Available at: [Accessed 4 Oct. 2017]. 8) Ieee.org. (2017). IEEE IEEE Code of Ethics. [online] Available at: 8.html [Accessed 20 Sep. 2017]. 9) Solar electric.com. (2017). What is Maximum Power Point Tracking (MPPT). [online] Available at: electric.com/learning center/batteries and charging/mppt solar charge c ontrollers.html [Accessed 29 Sep. 2017]. 10) Ti.com. (2017). TIDA Solar MPPT Charge Controller TI.com. [online] Available at: [Accessed 29 Sep. 2017]. 16

20 Appendix A: R&V Table Buck Converter Requirements A. Gate Driver Amplification Table 4: R&V for Buck Converter Verifications A. B. P Mosfet Switches C. N Mosfet Switches D. Output Voltage is stepped down E. Efficiency is over 80% 1. Power the gate driver. 2. Connect input and output signals to an Oscilloscope. 3. Verify output waveform meets specification. B. 1. Connect Oscilloscope to P Mosfet input and output. 2. Verify the waveform meets specification and is at the correct duty cycle. C. 1. Connect Oscilloscope to N Mosfet input and output. 2. Verify the waveform meets specification and is at the correct duty cycle. D. 1. Check output Voltage across a resistor with a Voltmeter 2. Verify that output voltage is close to ((P Mosfet duty cycle * (1 N Mosfet Duty Cycle)) * Input Voltage) E. 1. Calculate output power (Vo^2/R) and divide by input power (Vi * Ii) 17

21 Voltage Monitor Requirements A. Output voltage monitoring data as a analog signal B. Convert the analog signal to a digital signal via ADC on the MCU C. Send the analog signal to Team 35 Table 5: R&V for Voltage Monitor Verifications A. 1. Test output at 3 input voltage points and ensure linearity. 2. Ensure output is no more than 3.3V for max input voltage. B. 1. Test output at 3 input voltage points and ensure linearity. C. 1. Test output at 3 input voltage points and ensure linearity. 2. Ensure output is no more than 3.3V for max input voltage. Current Monitor Requirements A. Output current monitoring data as a analog signal B. Convert the analog signal to a digital signal via ADC on the MCU C. Send the analog signal to Team 35 Table 6: R&V for Current Monitor Verifications A. 1. Test output at 3 input current points and ensure linearity. 2. Ensure output is no more than 3.3V for max input voltage. B. 1. Test output at 3 input current points and ensure linearity. C. 1. Test output at 3 input voltage points and ensure linearity. 2. Ensure output is no more than 3.3V for max input current. 18

22 Microcontroller Unit Requirements A. Generate PWM signals B. Calculate Maximum Power Point C. Vary duty cycles of the PWM signal based on incoming data Table 7: R&V for Microcontroller Unit Verifications A. 1. Generate 2 PWM signals of varying duty cycles at 15kHZ. 2. Connect MCU pins to Oscilloscope and verify waveforms meet specifications (Duty Cycle). B. 1. Get 2 different input power data from voltage and current readings for 2 cycles of the code. 2. Calculate power for both cycles. C. 1. Based on current power and new power from B, vary the duty cycle. Battery Charging Requirements A. Charge the battery at the 3 different charging states. Table 8: R&V for Battery Charging Verifications A. 1. Connect Buck converter output to a 10% SOC lead acid battery through an ammeter and voltmeter. 2. Observe state of charge of battery as system progresses through charging cycle. 19