Made in-house with scrap Gallium-Arsenide cells using a documented, repeatable production process. These panels cost 30x less than the off-the-shelf panel ($80 vs. $2400) and have 80% the power output. The solar panels double as light-sensors to tell the satellite when it’s in the sun. The side panels can fit approximately 20 cells, while the top panels have been designed to include 25 cells each. We are also in the process of testing these panels to find exact efficiencies and to confirm that they will convert enough sunlight into electrical energy to supply ample power to our satellite.
- Total size must be less than 8.3 cm X 10 cm per side and 10 cm X 10 cm for top and bottom panels
- Must have >15% efficiency
- Must be available “off-the-shelf” without consumer purchasing restrictions
- Must be inexpensive
- Triple Junction Gallium Arsenide TASC cells (Spectrolab)
- 7.56 V nominal per side
- 60% Packing Factor
- 1.6-1.8 W produced per side
- cell efficiency ~27% (typical silicon cells are approx. 16% efficient)
- $2.50 per cell (~$80 per side)
One of the most important parts of construction is achieving such a high packing factor while ensuring that the cells do not touch one another (and result in a short). The placement template is a piece of acrylic which was cut using a Universal Laser Cutter. In the future these templates could potentially be made using rapid-prototyping in order to maintain the desired accuracy while minimizing costs.
During assembly, a solar cell can be placed in the cutout rectangle and pressed against the wall to ensure correct placement. The plans for the template were made by creating an Adobe Illustrator file using measurements taken from the PCB design software. The template is aligned with the PCB by sliding a dowel pin through the mounting holes in the PCB which have matching holes on the top placement template and a corresponding thick solid piece of aluminum.
Printed circuit boards have numerous advantages, the first of which is that they not only become the physical connection between the solar cells and the chassis but they also provide the electrical connection between cells which cut down on excess wiring that would be very difficult without professional processes.
The boards can also be produced in quantity for cheap prices and high precision using online vendors. When placing and arranging these cell rectangles on the PCB, multiple considerations were taken into account. The maximum packing factor was desired to covert the maximum energy possible, but was limited due to a few conditions. Due to the manual construction process, some safety factor had to be maintained between the cells in order to prevent accidental shorting due to human error. Also, the batteries’ maximum voltage is 8.4V so the panels must have solar cells in 5 cell groups in order to produce a high enough voltage. Mounting holes were placed on each of the corners to secure the PCB to the chassis.
The free program PCB Express was used to design our PCBs (seen on the right).
Two easy to use mediums exist to electrically and physically connect the solar cells to the PCB, solder paste and conductive epoxy. Solder paste is placed between two metal surfaces, such as between a solar cell and a PCB pad and then paste is heated with a rework tool to secure the connection.
Conductive epoxy on the other hand is an adhesive which usually contains silver. Epoxy comes in two tubes, resin and hardener, and when combined it begins to cure and eventually solidifies.
Trade offs: Solder Paste & Rework Tool vs. Silver Conductive Epoxy
|- no curing time- can remove broken cells- less mess, worry
|- cells can move if not cooled down- acrylic is melting
- air can blow cell out of place
- can burn PCB
The epoxy found to suit these purposes is Epoxy Technology’s EPO-TEK H20E Electrically Conductive Silver Epoxy.
- Set up Assembly Equipment: Solar Cells, Placement Template, new PCB, Epoxy, Syringe, and Alignment Dowel Pins
- Place clean, empty PCB in the Placement Template
- Mix Epoxy Resin and Hardener with appropriate amounts given on package directions
- Place mixed Epoxy in Syringe
- Dispense Epoxy onto PCB surface on the long rectangular pad
- Pick up one Solar Cell with the Handi-Vac and carefully place into position using the Placement Template
- Repeat Step 5&6 until all the cells are arranged
- Allow sit five minutes then remove Placement Template
- Allow Epoxy to cure for 24 hours
- Cut and strip top of 30 AWG Wire
- Solder Wire into holes on back side of PCB, cut to length and strip top
- Carefully bend Wire using tweezers and solder onto positive terminal of Solar Cell
- Repeat Step 12 until all positive terminals have been electrically connected to PCB
- Depending on the cure time of the Epoxy it may be necessary to only mix enough for a portion of the panel’s cells so that the Epoxy does not harden before positioning the Solar Cells
- DO NOT allow the Epoxy to harden outside the projection of the bottom of the Solar Cells; this will lead to electrical shorting between Solar Cells
- Care must be taken to avoid creating a solder bridge between the positive and negative sides of the Cell while soldering the wire into place which could short the two sides
Lithium Ion batteries, standard in many satellites, are lightweight and small. These batteries offer a low depth of discharge for increased life expectancy and low voltage drop. They are designed redundantly in case of battery failure. Recently, the power subsystem has reevaluated their choice of batteries for the satellite. This decision will be made based upon the needs of the flash system, which has also made some recent changes. The two options for batteries are the lithium ion (leave the stuff about it on there) and lithium iron phosphates, which have a different chemistry which lends itself to a different function.
The important connection of the power system is the batteries and panels, which means a different battery type may require less batteries in total.
- Allow power input/output of approx. 5.4 W
- Lightweight and compact
- Limit depth of discharge to 20% to increase longevity
- Avoid large voltage conversion between batteries and subsystems to maximize efficiency
- 4 LiFePO4 18650 Batteries
- 6.4 V nominal output voltage
- 1.1 Ah capacity
- 80 g total weight
This test will allow us to measure the depth of discharge and the lifetime of our batteries. This can be done using a relatively simple circuit that continually charges and discharges the batteries, using data acquisition and a relay switch. This data would potentially allow us to find a charge and discharge curve for our particular batteries.
Charge By Solar Panels
This will test the performance of the heart of the power system: how the batteries respond to the power from the panels. Based upon the results solar panel testing, the battery life cycle testing would occur under the light that best reflects the performance of the panel.
Given the series and parallel wiring, the ideal output for the panels is 7.56V with 210mA for the side panels and 240mA for the top panels. When testing at 3pm on April 16, 2013 in Providence, RI (latitude 42º) the panels performed below specifications. The average performance for the full side panels was 7.08V and 136mA and 7.09V and 133mA for the top panels. When one side and one top panel were wired in parallel, as they would be in the satellite, 7.18V and 290mA were produced. The expected values were 7.56V and 450mA.
The lower performance was expected as we were testing under atmospheric conditions in Providence, RI (latitude 42º). However, there may be may be an additional degradation of output due to construction methods.
Through Brown’s School of Engineering, we have access to a Solar Simulator Lamp, provided by Professor Pacifici. This lamp provides the power of a sun (100 W/m^2) that you cannot receive from a normal lightbulb. In addition, this machine can shift the wavelength of light to simulate being in space, which will be used for tests in the future. This lamp will be used to test the efficiency of our cells as well as our panels. The software of the solar simulator outputs our panel’s IV curves, which can be used to test our efficiency under different conditions, such as at an angle or at a higher temperature.
Control Tests: Before doing special cases tests, we ran a control test to set benchmarks on the performance of each panel individually. Each panel under perfect conditions should output 7.56V. However, the sun simulator was outputting the power of light at the equivalent of Earth, not space, so some decrease in voltage was expected.
Results: As you can see in the graph, the two side panels performed close to this standard, while the top panels underperformed at 6.2V and 5.3V. This also gave us a better idea of which panels has issues during construction. Another value you can see on the graph is the voltage at max power output of the panels. This is a lower number, due to the shape of the IV curve of the panels and cells. The max power values pointed us in the direction of considering powerpoint tracking for our panels to optimize their power output.
Lost Cell Analysis
An additional test case that our panels were analyzed for was that of a ‘Lost Cell’. A lost cell could be the result of mechanical damage to a single cell, wiring or shorts, or a degradation in performance due to localized thermal conditions throughout the panel. For this study, test panels were wired with only two TASC cells in series or parallel and performance was recorded using a Sun Simulator while one of the cells was covered (blacked out).
As can be seen in the included tables, in the parallel configuration ‘eliminating’ one cell resulted in approximately have the current and power, as expected. In the series case, covering a single cell resulted in have the open circuit voltage, as expected, but did not produce consistent maximum voltage or current changes.
Additional testing is required. Coming soon…
According to the Spectrolab data sheet for the TASC cells, the peak power voltage of an individual cell decreases by 6.2mV/ºC increase in temperature. This means, for example, that a 100ºC increase in temperature of the panels would decrease the open circuit voltage from 7.56V to 5.7V for the sets of three TASC cells in series.
Results: Coming soon…