DESIGN AND TEST OF A POWER MANAGEMENT SYSTEM IN A PICOSATELLITE

Transcription

1 DEIGN AND TET OF A POWER MANAGEMENT YTEM IN A PICOATELLITE Wade G. Tonaki Department of Electrical Engineering University of Hawai i at Mānoa Honolulu, HI ABTRACT The initial design of a picosatellite power management system is presented. The system consists of solar cells, batteries, voltage converters, battery charger, battery monitor, and switches. ome of these parts were also verified, including the battery charger, which displayed the ability to balance individual cells in the battery pack. INTRODUCTION The vast majority of NAA-related satellites take years of research and development before being launched and becoming fully functional in space. Costs incurred during those years quickly add up well into the millions of dollars. In addition, due to their large size, they cost millions more to launch. Picosatellites offer a smaller, cheaper alternative that also takes considerably less time to develop and can also maintain the same functionality. The goal of this project was to efficiently generate and distribute sufficient and continuous power for a picosatellite bus and payload. This could be accomplished by meeting different subsystem requirements. One of these was to acquire usable power from exposed solar cells to effectively run on-board electronic components while recharging the onboard batteries during the sun-on period. Another was to maintain battery power operation during sun-off, or eclipse periods. The batteries had to provide sufficient power to the essential electronics, and if possible, nonessential electronics. There is, however, a limited supply of energy creating the need for a power budget. This report covers the initial flow design, power budget, selection of parts to be used, and testing of parts to be used in the electrical power system. Research was done in collaboration with the University of Hawaii s mall atellite Program, which is currently developing its sixth generation of small satellites, building off of previous designs and incorporating the latest in technology. POWER BUDGET The first task was determining the minimum amount of electronic components on the satellite and, with the knowledge of the size of the mechanical structure, calculating the corresponding power budget. The initial list of components is as follows: 67

2 Energy consumption/orbit of major subsystem components (Typical case) Component Mode Voltage Current Power Duration Energy Energy (V) (A) (W) (minutes) (Wh) (mah) MHX-2400 Radio Transmit MHX-2400 Radio Idle Retrodirective Array active PicAxe 28X1 active UB to erial Converter active Fuel Gauge active Gumtix Microprocessor active User Inputs Total Table 1 - Electronic Components Assumed low earth equatorial orbit duration is 96 minutes, where the sun-on time is 60 minutes and the sun-off time is 36 minutes. The retrodirective array and Mircrohard transceiver (MHX-2400) both work dependently, and for a typical case, operate for 20 minutes in a single orbit. The PicAxe, used for beaconing purposes, consumes minimal power along with the fuel gauge and UB to erial Converter. The main onboard microprocessor, the Gumtix, is operating at all times. The next step was to calculate the amount of energy stored in the battery packs and the amount of energy produced by the solar cells in a single orbit for a typical case. The final mechanical structure is a cube with 12-inch sides. The pectrolab ultra triple junction solar cells have high efficiency ratings of 28.3% and measure 1.57 inches x 2.76 inches. This allows 7 parallel sets of 4 cells in series on 5 sides of the satellite for a 140-cell configuration. There are bypass diodes placed between each cell in series to protect series connection power drain. Blocking diodes prevent power drain caused by unpowered sets. This configuration maximizes the amount of power that can be gathered while also meeting the minimum specification voltage for the battery charger. Blocking Diodes Bypass Diodes Figure 1 - olar Cell Configuration for 1 side 68

3 An estimate for the amount of power produced could then be determined with the knowledge of the solar array. A conservative 1.5 sides of solar cells is exposed to sunlight; conservative since multiple sides are exposed to direct sunlight and the albedo effect. Each cell provides 2.31V and 0.436A. The lithium polymer battery pack has a voltage rating of 14.8V and capacity of 6350 mah. Energy Available / Orbit (Typical Case) User Inputs Component olar Cells Batteries Number of batteries NA 1 Capacity (mah) NA 6350 Voltage (V) Average Current (A) var/w load Average Power Produced (W) NA Duration of Production (minutes) 61 NA Effective Capacity (Wh) Efficiency Transmission 28.30% 90.00% Available Capacity for subsystems (Wh) Available Capacity for subsystems (mah) Table 2 - Energy Available The final power budget can then be determined. Of note are the multiple stages of voltage conversion as well as charging that carry an efficiency rating where power is lost. The number of voltage converters from solar cells to electronic components was minimized. The charging, step-up voltage conversion, and step-down conversion were all conservatively estimated to be 90%. Picosat Energy Budget / Orbit (Typical Case) User Inputs Energy Available from solar cells and batteries (Wh) Energy Used by ubsystems (Wh) 5.55 Charging efficiency 0.9 econdary converter efficiency 0.9 Battery Capacity Used by subsystems (Wh) Net Battery Capacity Change (mah) Actual D.O.D. (%) Table 3 - Energy Budget The net battery capacity change was determined by calculating the amount of power provided by solar cells into the battery after charging and voltage up-conversion efficiencies were accounted for, while accounting for the amount of power consumed in a typical orbit accounting for after a 10% loss due to voltage down-conversion. Ideally, a negative depth of discharge is desired since its indicative of power being added into the satellite per orbit. In the typical case, approximately 250 mah, or 4% of the maximum capacity, is added in a single orbit. 69

4 POWER YTEM The power system begins with the solar cells, which give an output voltage of 9.24V that have to be up-converted since the battery charger requires a minimum voltage of 12V. The battery charger then connects to both the battery monitor and battery pack by cells. The battery monitor is in charge of gathering vital statistics such as battery capacity, voltage, current and temperature readings. The battery pack is then connected to multiple down-converters to the appropriate voltages for the devices listed in Table 1. 8V, -8V DC tep witching Network 12V DC tep Up Converter olar Cell Array Battery Charger Battery Monitor Battery Pack 5.4V DC tep 5.0V DC tep 3.3V DC tep On-board Electrical Components CDH Figure 2 Basic block diagram of a small-satellite power system TEP-UP CONVERTER The DC voltage step-up converter immediately following the solar cell array is the Maxim It has a preset output of 12V, perfect for the battery charger, and is 90% efficient over 30mA to 2A loads; the calculated current load is 3.18A after the conversion loss, so two step-up converters in parallel is necessary to maximize efficiency. BATTERY CHARGER The battery charger of choice is from FMA Direct and is the Cellpro 4s, 4A cell balancer. The charger can support charge currents up to 4A at 12V and automatically takes care of cell balancing, a crucial aspect of the power system. Without cell balancing, a cell in the pack may exhibit different characteristics than the others, possibly resulting in a broken battery pack in the future. The charger is accessible via serial where cell data such as voltage, capacity and current 70

5 readings can be graphed and plotted over time. An example of the cell balancing data was gathered: Battery Charger B1Volts B2Volts B3Volts B4Volts :57:36 13:04:48 13:12:00 13:19:12 13:26:24 13:33:36 13:40:48 13:48:00 Time Figure 3 - Battery Charger Data In the battery pack being charged, the second cell was at a considerably lower voltage than the other three cells; however the battery charger was capable of siphoning the charge into the required cells to keep them balanced and the pack healthy. The battery charger is capable of notifying when the pack has reached a full charge. BATTERY MONITOR The battery monitor/fuel gauge chosen was the Texas Instruments BQ2060A. The original model, BQ2060, was used in previous Apple laptop batteries. They are capable of measuring individual cell voltages, temperature, current and capacity readings. The battery monitor is necessary since the data acquired from the battery charger is not available when the satellite is not in a charging state. It uses MBus protocol to communicate with the microprocessor. BATTERY PACK The chemistry of the battery will determine how much space and weight is allocated in the satellite. Ideally, the batteries chosen should have the lowest gravimetric and volumetric energy densities to preserve more space and weight for more important components specifically needed for the mission. It is for that very reason that lithium polymer chemistry batteries were picked for this satellite over the more popular choices nickel-cadmium (NiCd) and lithium ion (Li-Ion). The nickel-cadmium battery is the oldest of the three and has an energy/weight ratio of Wh/kg. It is known for having lower series resistance, allowing it to supply high surge currents; this makes them a favorable choice for high power applications like camera flash units. NiCd batteries are generally also very low in cost and less toxic than other chemistries. They 71

6 also maintain a relatively constant voltage while in use and have the best cycle durability out of the aforementioned chemistries. A disadvantage of the NiCd is it has a self-discharge rate of 20% per month. In the case of a satellite mission, this is not a considerable factor since the battery will never sit idle for more than a day, but this can factor in while waiting for a launch date. It also suffers from the memory effect, which, when a battery has begun recharging before fully discharged, remembers that particular voltage in the charge cycle. This decreased voltage in the charging cycle would prevent the satellite from getting as much energy out of the battery and subsequently reduce its capacity. Lithium ion batteries were developed significantly in the early 1990s and are advantageous in many ways. They have an energy/weight ratio of 160 Wh/kg, have a low self-discharge rate of 5%-10% per month and do not suffer as badly from the memory effect as NiCd batteries. They also come in an abundance of shapes and sizes, making it easy to meet the physical requirements of the satellite. A drawback to the Li-ion battery is its lifespan depends upon the time of manufacturing in addition to the number of charge/discharge cycles. They are also extremely sensitive when being discharged; if the battery voltage drops below a certain level it will result in considerable permanent damage. This creates a need for an external protection circuit that shuts off the battery when a low voltage threshold is broken. Li-Ion batteries are also more expensive and for more unstable than most other battery chemistries. Temperature, pressure, overcurrent, overcharging, and undervoltage are all important aspects of the battery that must be monitored to ensure its longevity. The most recent developments in battery technology include the lithium polymer batteries, which have evolved from the lithium ion chemistry. Lithium polymer batteries share the same advantages as the Li-Ion, but with a slightly higher energy/weight ratio of Wh/kg. Li-Poly batteries are also lighter since they do not require a cell casing to apply external pressure like the Li-Ion batteries need. This also makes them favorable since they can be shaped for a specific need. The solid polymer electrolyte is also not flammable, unlike its predecessor. A negative of the Li-Poly chemistry is it has the lowest cycle durability of the considered chemistry and is known for having problems with internal resistance, which limits the maximum discharge rates. For the purposes of this project, the most appealing choice was the lithium polymer chemistry since it seems to be at the forefront of technological development and offers superior performance and safety. Despite its advantages, the issue of low cycle durability would have to be seriously considered if this were a space-ready model. The battery of choice is manufactured by Tenergy and has an overall voltage of 14.8V with 6350mAh. It can continuously discharge at 15x the cell capacity in amp Figure 4 Tenergy 14.8V 6350mAh pack hours, which is more than sufficient for this mission. 72

7 TEP-DOWN CONVERTER The DC voltage step-down converter following the battery pack to the specified voltages, 8V, 5.4V, 5V, -8V is the Maxim 1627; the 3.3V line required solely for the battery monitor can be taken directly from the serial to UB converter. This step down converter is adjustable by using a voltage divider with resistors. It has efficiencies greater than 90% for loads between 30mA and 2A. POWER WITCHE In case of power conservation, certain devices need to be capable of being turned off. For the step-down converters with single devices attached, they can easily be managed by connecting a digital I/O pin to the HDN pin of the MAX1627 that determines whether the chip operates or floats. However, for the converters with multiple devices attached like the 5V line, a power switch is necessary. The switch, FDC6331L, is manufactured by Fairchild emiconduc- It can handle input voltages from 2.5V to 8V with load currents up to tors. 2.5A. CONCLUION The initial design of an electrical power system of a picosatellite was successfully completed. A fairly accurate power budget was determined and can be further verified in the future by testing each individual component and its efficiencies. In addition, research was done to find the most effective and efficient parts available to meet the needs of the subsystem. uccessful testing began with the battery charger and its abilities to balance a multiple cell battery. In the future, the rest of the components step-up and step-down converters, battery monitor, and switching network need to be tested. Further development would involve the printed circuit board design integrating all of the subsystem components. Another round of testing would verify the components functionalities, whereupon a final round of testing can be conducted with all subsystems fully integrated to form the satellite. ACKNOWLEDGEMENT I would like to express my gratitude to my mentor, Dr. Wayne hiroma, for providing me with numerous opportunities to experience engineering directly and create a positive learning environment for me since my freshman year of college. I would not be in this wonderful position if not for his undying efforts towards the betterment of all his students. I would also like to offer my sincere appreciation to the Hawai i pace Grant Consortium for supporting our small satellite program. Last, but not least, I would like to thank my fellow peers who have donated countless hours towards the betterment of our projects. 73