Changes

4,150 bytes added ,  16:14, 27 August 2016
no edit summary
Line 114: Line 114:     
Oscar has three mounting holes, designed for #8 bolts, for mechanical integration into HABHIVE.
 
Oscar has three mounting holes, designed for #8 bolts, for mechanical integration into HABHIVE.
 +
 +
== Milestones ==
 +
Oscar is the first-ever PCB developed for the Balloons Core Avionics Suite. It is also one of the most feature rich and modular avionics systems produced within SSI at the time of its creation. Every system on Oscar is functional and flight-proven -- the only subsystem that was not proven was the FET current sensing, which was removed.
 +
 +
== Problems ==
 +
Oscar, as the first in the line of avionics suites, was riddled with problems in its production, design, execution, and testing.
 +
 +
Key problems included:
 +
* Initial reflow of Oscar resulted in many shorts, most notably on the MCP23017 and LTC4218.
 +
* Very small grid spacing in Altium resulted in multiple pins on the I2C multiplexer being unconnected to anything. All three address selection pins were short-soldered together, and greenwired to the backup battery ground pad.
 +
* Vigorous whicking removed a trace & via leading to FET 1, resulting in FET 1 being unusable.
 +
* Poor reflow of pressure sensors resulted in the removal and subsequent re-soldering of all pressure sensors. Only one of the two BMP280's and one of the MS5803's is functioning at final flight time. The MPL3115A2 is functional for some period of time, and non-functional by time of flight.
 +
* Improper design of FET LED circuit.
 +
* The GPS breakout TX/RX lines are swapped; they are later swapped on the breakout itself, to allow for the PCB to remain unmodified.
 +
* Despite having SPI pinouts, there are no chip select pinouts.
 +
* The thermocouple is connected backwards, the ferrite beads were of the improper frequency rating, and the diodes had to be removed.
 +
 +
== Flights ==
 +
 +
=== SSI-39 ===
 +
SSI-39, the inaugural flight of Oscar, was a flight plagued with unfortunate mistakes, and some poor decision making. Flight software for Oscar was completed approximately 4 hours before departure from campus, fully functional. En route to launch site, additional components were added, including ascent rate calculation,  as well as GPS flight mode and a few other features.
 +
 +
At launch site, extensive debugging and checking took place to verify the system. GPS lock was seldom obtained initially, but was later regained with high frequency.  RockBlock communication was very sparse, often losing lock and communicating on average every 15-20 minutes rather than the programmed 2 minute interval. Debugging, configuration, and testing took place on-site until 1 pm -- two hours past the 11 am expected launch time. At this point, the system was performing nominally outside of habhive, but, upon insertion, worked sporadically, until the flight software terminated randomly. At one point, it was determined that the time was getting late and the payload must be launched. A communication was received, and it was determined that the payload was ready for flight. After release, no more communications were received.
 +
 +
Upon return to campus, root cause of failure was analyzed. The problem lied in the GPS antenna being loose -- contact with pressure sensor housing of the BMP280 or MPL3115A2 caused flight software to crash. The flight provided no data logs, and no communications.
 +
 +
Aside from the antenna problem, the system was nominal.
 +
 +
=== SSI-40 ===
 +
[[File:ssi40oscar.jpg | right| thumb | <center> Oscar, during SSI-40 </center>]]
 +
SSI-40, taking place shortly after SSI-39, was an exact re-flight of Oscar, with the modification of having the GPS antenna taped to a habhive wall. The system performed nominally from prep to launch and throughout the entire flight. Communications were frequent, coming in often 30 seconds apart. The payload successfully recorded and transmitted flight critical parameters for the entire duration of the flight, including internal and external temperatures, altitude, ascent rate, power state, and more. A successful cutdown was also triggered at 25 km using a custom FET and nichrome wire. A heater was included but did not activate during the flight.
 +
 +
The payload was recovered floating atop algae in an irrigation canal. All flight electronics remained in nominal state at time of recovery. The Micro SD card contained thorough data from throughout the entire flight.