Post by Sean Curry on Jun 22, 2017 10:33:14 GMT -7
The North Medford (OR) High School Tornado Team had a successful launch and recovery on June 20th, in spite of a few technical problems. I thought I'd share a few key learnings:
1) The cable from the IMU to the Arduino broke - 2 wires became disconnected. This is likely the result of the new calibration process which involves manually rotating the IMU. Although we were able to re-solder the wires, we could not get the IMU to work after that, and had to track manually (see below). We strongly suggest that all teams inspect this cable, and reinforce it by hot-gluing it to the board, or some similar fix. Also, be sure to take a soldering iron to the August launch.
2) Because of the IMU problem, we could not track automatically with the dish. We pointed it it manually by powering down the servos, and the students took turns manually pointing the dish to the estimated location of the balloon. They then used the signal strength lights to optimize the pointing. As a result, we were able to stream video out to 20km or so, although intermittently at times.
3) ffmpeg worked, but intermittently stalled and crashed at times. We kept a command window open, and when this happened, killed it with CTRL-C and restarted the process. The stalls and crashes seemed related to intermittent signal from the video payload.
4) The saved video segments on the Pi had incorrect date and time. We got an error on startup: "Cannot Access Hardware Clock via any known method". There does not appear to be an RTC on the streaming video payload (only the still image). Should we add an RTC, or should the Pi pick up time from the PC via the network connection?
5) Tracking via Iridium worked very well, although there was a 3-4 minute dropout at high elevation. We saved the Iridium data from the BOREALIS tracking page by Cut and Paste, and put it into Excel for detailed analysis of the flight. What do the balloon icon colors indicate on the BOREALIS page?
6) Is there any way that Stream can add some kind of indicator on the eclipse.stream.live website to show the status of their video server? Our sense was that the server stalled at times, but we couldn't tell if the problem was at our end or the server end. Also, there seems to be a 20-30 second lag in the video stream - why is this?
7) On our first launch attempt, the thin line from the balloon to the cut down system snapped, causing a loss of the balloon, and we had to start over. That thin line passes through some sharp edges in the cut down box, and we are re-evaluating how this will work in August. Of course, we should have had a safety tether on the balloon, and will do so in the future.
8) Recovery was a real hoot. Be sure to have a student who can climb trees, a long pole saw, and a friendly rancher with an ATV who can unlock the irrigation ditch gate, and drive through the thick blackberries!
We hope some of this is useful to all the teams, and we would like to hear feedback from everyone on the Tuesday launch.
Sean, Colin, and the entire NMHS Tornado Team
1) The cable from the IMU to the Arduino broke - 2 wires became disconnected. This is likely the result of the new calibration process which involves manually rotating the IMU. Although we were able to re-solder the wires, we could not get the IMU to work after that, and had to track manually (see below). We strongly suggest that all teams inspect this cable, and reinforce it by hot-gluing it to the board, or some similar fix. Also, be sure to take a soldering iron to the August launch.
2) Because of the IMU problem, we could not track automatically with the dish. We pointed it it manually by powering down the servos, and the students took turns manually pointing the dish to the estimated location of the balloon. They then used the signal strength lights to optimize the pointing. As a result, we were able to stream video out to 20km or so, although intermittently at times.
3) ffmpeg worked, but intermittently stalled and crashed at times. We kept a command window open, and when this happened, killed it with CTRL-C and restarted the process. The stalls and crashes seemed related to intermittent signal from the video payload.
4) The saved video segments on the Pi had incorrect date and time. We got an error on startup: "Cannot Access Hardware Clock via any known method". There does not appear to be an RTC on the streaming video payload (only the still image). Should we add an RTC, or should the Pi pick up time from the PC via the network connection?
5) Tracking via Iridium worked very well, although there was a 3-4 minute dropout at high elevation. We saved the Iridium data from the BOREALIS tracking page by Cut and Paste, and put it into Excel for detailed analysis of the flight. What do the balloon icon colors indicate on the BOREALIS page?
6) Is there any way that Stream can add some kind of indicator on the eclipse.stream.live website to show the status of their video server? Our sense was that the server stalled at times, but we couldn't tell if the problem was at our end or the server end. Also, there seems to be a 20-30 second lag in the video stream - why is this?
7) On our first launch attempt, the thin line from the balloon to the cut down system snapped, causing a loss of the balloon, and we had to start over. That thin line passes through some sharp edges in the cut down box, and we are re-evaluating how this will work in August. Of course, we should have had a safety tether on the balloon, and will do so in the future.
8) Recovery was a real hoot. Be sure to have a student who can climb trees, a long pole saw, and a friendly rancher with an ATV who can unlock the irrigation ditch gate, and drive through the thick blackberries!
We hope some of this is useful to all the teams, and we would like to hear feedback from everyone on the Tuesday launch.
Sean, Colin, and the entire NMHS Tornado Team