We launched a weather balloon into the path of the annular eclipse. This is my favorite of the zillions of pictures. It was taken shortly after the maximum eclipse.
We launched at 5:33pm on May 20, 2012 from Westwood in Northern California. The balloon went to roughly 90000'. We got really lucky with the winds and picked it up only about 38 miles away.
There is a solar filter on the right side of the image so that the sun is not so over-exposed.
The attitude control system described below managed to stop the capsule from spinning but ran out of gas before it could point the box well.
Inside the Styrofoam box was:
Balloon and Descent
We used a Kaymont 1kg balloon with around 150 cubic feet of helium for the ascent. Roughly 5m/s ascent rate?
We used a rocketman 3' parachute which seems like it was probably too small. The ziptie of the SPOT broke, and the radio batteries came loose so that the radio didn't transmit.
The final trajectory was unknown, only a handful of ground points (no altitude) are known, sent by the SPOT. Ignore the first point. The 2nd point is the launch site in Westwood.
FAA was notified 2 days ahead. I set up a notice to airmen (how sexist!) by calling 877-487-6867
Spoke with KW
High ball balloon
11 mi ENE of Rogers airport (code O05)
0030 Zulu (5:30 pm) launch, for 2.5 hours. East bound. To 90000’.
The final trajectory prediction from 24 hours before launch looked like this, with the final point also plotted:
To estimate where the path might be, I used data from the atmospheric balloon soundings that have been done from Reno for the last 17 years. They are published here: http://weather.uwyo.edu/upperair/sounding.html
Mark helped me glom them all into a tab spaced file and I crunched them in excel and converted them to google earth format and plotted them.
We set out to find the capsule at around 7pm after enjoying the eclipse from the ground and packing up our picnic. We staged our 6 year old and his mom and the minivan at the start of the dirt road and set out on the dirt road and parked near Biscar Reservior. The hike was 1.5 miles as the crow flies but it felt much farther because we started in the dark. I used my phone and followed the numbers till they got pretty close and Emily and I saw it nearly simultaneously. It stood out in the starlight among the sage and rocks. The sky was super dark with tons of stars. My 11 year old was way past his bedtime!
The camera had broken off its mount, the puffer had broken off its mount (both hot-glued to the styrafoam), the battery was loose and the SPOT had broken free. The GoPro had a vacuum inside and some serious force was needed to open it. The balloon was spaghetti.
The Puffer (attitude control)
Sometimes we all need a little attitude control...
We need a way to orient the capsule. It only needs to work for a few minutes. A few ideas were suggested and rejected:
1) Have a motor orient the camera. The problem is the balloon only has one attachment point so there's nothing to push against. A motor on the string has minimal effect.
2) Have a net over the balloon to get more than one attachment point and then use a motor. The problem is that the balloon gets really big so the net might get too heavy. Furthermore, it may pop the balloon through abrasion. The risk here was unknown but maybe something could be made to work.
3) Have a motor orient a mirror in front of the camera. The camera could point straight down and the mirror could be at 45 degree so the view is facing out. This would probably work but has a few problems.
a) The pictures would be sideways or really any orientation.
b) It would be hard to waterproof.
c) It's a little tricky to get the mirror to spin 360 degrees. There will be an obstruction in at least one direction.
4) Have a mass that gets spun in the opposite direction as the capsule so that the capsule stops. The problem with this is subtle. If there is a static force (due to, say, some twist in the line), then the mass would have to spin faster and faster to compensate. Eventually the top speed would be reached with no further control possible. It might be a bit on the heavy side. This could possibly be made to work for the brief time required.
5) Have two capsules with a motor between. Put the camera in one and everything else in the other. This has the exact same control problem as #4, even though it's disguised a bit.
6) Put a long stick hanging horizontally from the bottom. Drive it with a motor. This is the same as #4.
What we finally ended up with was this.
It uses a CO2 cartridges and pneumatics from paint ball guns.
Here's a parts list so far:
Pneumatics from airsoldier.com:
PPS-MICROROCKREGPalmer Micro Rock LPR 1 $45.00 $45.00
FINISH: MATTEBLACK (pressure regulator)
MAC-33A-AAA-RDUA-0BLMAC 33A-AAA-RDUA Solenoid Valve 2 $22.00 $44.00
CCI-12GRAMCHANGERCCI Phantom 12g Changer 1 $22.00 $22.00
FINISH: BLACK (12 g CO2 cartridge holder)
RAWBOTTOMLINE Raw 7 Degree Bottomline ASA 1 $15.00 $15.00
PNEUMADYNE-EB-BARBSPneumadyne M3 Hose Barbs 1 $8.00 $8.00
BESWICK-TUBING Beswick LP tubing 4 $1.50 $6.00
1/8-27 brass plug Plug from home depot
12g CO2 cartridge from any sporting good store with bb guns.
From Pololu: 1 x #1265 MinIMU-9 Gyro, Accelerometer, and Compass (L3G4200D and LSM303DLM Carrier) = 49.95
Arduino Nano OSEPP-NAN-81 (~$35) (Frys)
two NTE2987 power mosfets (Frys)
two power flyback diodes. (Frys)
Sparkfun Real Time Clock DS1307 64 x 8 Serial Real-Time Clock (~$15)two resistors from gate to ground so the fet gates don't get overvoltaged. 550k
The program for the Arduino controller uses the compass for position angle and the gyro for rate-of-change of position angle. The output parameter is the number of milliseconds that the valve is open. The sign indicates which valve is open (the one that causes clockwise motion or the other one). The algorithm is pretty simple for now. The gyro scaled to 256 overrides the compass beyond a 50 deg error. Pulses under 10ms are ignored.
I will probably have to tweak the algorithm when I move from Easter basket to the real capsule. I've attached the arduino code which is fairly readable. It includes code to start at a certain time. Sequence1 is the good stuff. [The final code, hab2 is also included]
Two issues right now:
1) The pneumatics leak; a fresh cartridge is empty in 24 hours. I'm hoping to improve this.
2) When the basket swings back and forth, the algorithm tries to compensate for the changing compass reading. This will waste a lot of gas. I may be able to use the accelerometer to only correct position when it's level. We'll see.
Thanks to all who helped with this project including but not limited to Emily, Josh, Sean, Mike and Grace, MarkC, JohnA, MikeT, MikeF, DavidS, DanM, JohnH, Pololu, AirSoldier, and the folks on the arduino and chdk groups.
And especially, my wife, who's been supportive along the way.
Questions can be sent to firstname.lastname@example.org