On June 30th 2013 PSAS conducted our tenth launch "L-10" with our LV2.3 launch vehicle.
- PSAS Launch-10 — June 30, 2013
- Launch Analysis
We have an Analog Devices ADIS 6 axis + magnetometer IMU that needs integration into our flight computer. We have opted to use an STM32 ARM based microcontroller to handle low level communication with the sensor, while reporting it's values over Ethernet to the FC.
We want to use the same STM32 architecture to do other low level work, in this case send updates to the roll control servo. This board does also doubles as the hardware launch detect (using a shorted tether).
Over the last year we rebuilt our Launch Tower Computer (LTC). This computer serves as the link to the rocket when it's on the pad, providing power and communication as well as safely controlling the ignition. This also meant rebuilding the software to talk to it.
We validated Roll Control on our October 2010 launch. But this was using a one-off, completely embedded solution. We very much would like to fly the same algorithm using our new flight computer framework, Ethernet communication, etc.
This winter/spring CS Capstone team built a Flight Computer Framework. This allows us to write modules in c and then define the connections and program flow between them using simple configuration files. The framework then handles setting up the build targets, events and callbacks for the computer. This is the flight test using the new code.
The new LTC has an additional level of safety on the firing sequence. The flight computer must assert a line high to let a relay close and allow the rocket to be fired. This lets us check for conditions on the rocket (GPS lock, sane sensors, etc.) and automatically safe the launch circuitry if something goes wrong.
We have a new HD camera pointed out the side of the rocket, as well as a new 3D printed shroud for the downward facing camera.
A special payload is flying on this launch. A 360° camera module made up of 5 GoPro's will sit just above the roll control module and will be stitched after flight into a singular, extreme HD full view of the flight!
One of the small innovations on this launch are small, locking connectors for the ethernet and power that are partly COTS and partly 3D printed.
The nosecone has the top half of the recovery system. A pusher presses against the recovery module below it while the two are held together by the nosecone separation ring (NSR). When the NSR is fired the nosecone separates and the pusher pushes it away from the rest of the rocket. This carries the drogue chute into the air-stream.
The rocket falls under drogue until about 300 meters altitude when the drogue and nosecone are cut away by a pyrotechnic line cutter. The action also pulls out the main chute.
The payload module contains two commercial off the shelf 'flight computers' that are simple devices that track the height of the rocket and ignite the charges that open the parachutes at the right altitudes. The payload module also contains the downward facing camera and a side facing HD camera. The downward camera sends down a live image through an amateur TV band transmitter. All other cameras on the rocket record to SD cards.
The 360° camera module houses 5 HD GoPro cameras that recorded a complete, overlapping cylindrical view from the rocket.
The roll control module uses small canards to spin the rocket clockwise and counterclockwise as it flies. This is our test stabilization system.
The flight computer (FC). This is the main computer, Ethernet hardware battery and power supply, etc. It has a special shell it drops into that has 3 linealry polarized clydrical patch attennas. One for GPS, one for live video broadcast and one for WiFi.
The motor casing also has a set of fins for stability. But ours are special The 3 fins are on a free-rotating bearing so rotational forces on the fins do not overpower the small roll control canards.
Here is the timeline circa May 21st, 2013:
|Sat 6/08||Integration day 1 @ Dave's house Raise Tower|
|Sat 6/22||Final prep and integration test day @ Dave's house / Final Checklist|
|Tue 6/25||Go/No Go decision, final logistics, etc.|
|Thu 6/27||Weather last call|
|Sat 6/29||Drive to Brothers|
We mostly stuck to this timeline, but the actual hardware button up didn't happen until Wednesday before the launch. The RocketHub was not finished in time to be meet the cutoff and was not flown this launch.
|Fri 6/28||Evening||Prep equipment and pack up toy hauler at Dave's house.|
|Sat 6/29||Morning||Depart from Dave's house for Brothers.|
|Sat||Afternoon||Arrive in Brothers, set up launch tower, prep rocket for flight, watch other people's rockets.|
|Sat||Evening||Final tower and LTC assembly and testing, final rocket assembly. Astronomy, after the moon goes down.|
|Sun 6/30||Early||Rocket to pad, final system testing before flight.|
|Sun||Afternoon||Recovery, debrief, pack, drive back to PDX|
|Sun||Evening||Unload rocket stuff at Dave's house.|
Note, launch actually occurred at 12:30, an hour and a half late.
See the Procedure Book for this launch
At aproximatly 12:30pm local time (19:30 UTC) we sent the ignite command to the rocket.
All recovered flight data can be found in this git repo:
Download all the data as a zipfile:
Some things actually worked!
- WiFi worked very well
- FCF enabled fast, independent development
- 360° worked perfectly
- New camera attachment worked very well
- Recovery and 8 month old NSR worked
- Ethernet, ChibiOS, worked
- Cool LTC control software
- More people deeply involved in more pieces of the rocket then ever before
- FC Reboot at boost
- ATV Receiver didn't come up
- USB bus failed, dropping GPS
- LTC connection and phidgets server
- Ground video at tower (ran out of time)
- TeleMetrum version mismatch (old firmware)
- ATV and LTC WiFi collision
- Couldn't hear radios, 2m was spotty
- People didn't have roles or practice
- Needed a way to drag rocket in desert
- Recovery team organization
- Roll Control
- Battery charger stopped working (again) leading to discharged battery and panic about battery SOC
Planning for the next flight.
No data was recovered from the rocket during the boost phase of the launch. The logs start again sometime after apogee when the rocket was already on parachutes.
FC Launch Failure Timeline
|FC System Clock [mm:ss]||Activity|
|18:50||Rocket Ready command sent|
|19:45||FCF logfile last modified on disk (CF card)|
|20:16 (aprox)||last telemetry over WiFi (4 g)|
|—||Missing 49 seconds|
|21:05||First CF logged data from FCF|
|52:56||Last timestamp in /var/log/syslog|
|53:10||FCF logfile last modified on disk (CF card)|