Time flies like an arrow; fruit flies like a banana. 2010-02-08 22:19:00

Declaring a week’s advance notice of a balloon launch to the edge of space when we hadn’t even bought most of the equipment, let alone built it, was probably an act of pique, if not madness. Remarkable how well it worked out, though – just look at the pictures!

The plan was simple: a ham radio broadcasting an APRS position beacon, a GPS that was known to work at high altitudes, a camera hacked for time-lapse photography, and an Android cellphone that we’d program to scream out its own GPS co-ordinates via SMS whenever it caught a glimpse of a cellphone network.

We were buoyed by the unexpected help from all three members of another Space Balloon Project, one of whom Christie had met in line for the checkout at a Ham Radio Outlet, who then showed up at the fateful meeting when the timeline was decided. They’d done it before in style, and even – very bravely! – volunteered their camera and parachute (a donation from rocketchutes.com) for the experiment.

So the work began.
12 hours before our planned launch, Icarus, as Andrew’s phone-home code for the Android had been dubbed, was working beautifully. Everything else was a shambles.
The APRS beacon seemed to be sending, but it wasn’t being received by either our own radios or the public repeaters. Our receiving radios couldn’t pick up either our own or any other APRS beacons. The weather forecast for the Central Valley was miserable, and the FAA says you can’t launch balloons into more than 50% cloud coverage. We hadn’t worked out how to pack the payload safely, nor how to rig it and the parachute to the balloon. Not even the known-working camera was working!
But by 3am of the launch day, the electronics seemed to be functional, and since we had an extra tank of helium and two spare balloons, there was enough of a margin for error to make the rest of it up as we went along.
On arriving in a largely overcast Central Valley in the early afternoon, I rechecked the FAA regulations and found that the 50% cloud cover restriction didn’t apply to payloads as light as ours, so our initial gas station rendezvous point instantly turned into our first launch location.

The gas station was well equipped with many things that we’d forgotten to bring, including miniature weather balloons for testing before the real launch (available in packs of 3, with optional lubrication). The staff were also related to the people who owned the adjacent field, and told us that although they weren’t giving us permission to enter, they also weren’t going to call the police immediately. We were committed.

More problems soon showed up – the radios hadn’t taken a charge overnight, and were dangerously low on battery. The ground station battery failed altogether after receiving a single APRS packet. Our attempt to use expandable foam to hold the components in place in the cooler also failed, leaving us with a soggy mass in a plastic bag with no rigidity. We broke the attachment points on the cooler.

Nonetheless, after jury-rigging a gas regulator out of high-school chemistry supplies (inexplicably present in Christie’s van), we started to fill the balloon. We turned on all the electronics, confirmed that we were receiving both SMS position updates every minute from Icarus and APRS position updates (via Brian’s APRS-capable radio) every 30 seconds, and then, with the balloon almost completely filled and yearning to fly, we proceeded to debate the best way to attach the payload, parachute and balloon together. After filling our quota of rope burns, we were ready to fly – and launch we did, complete with countdown!

Partying ensued, while we watched the balloon soar skywards at a furious pace, and was soon lost in the clouds. The ebullience muted a little when we discovered that the APRS data wasn’t being seen by any of the area digipeaters, although Brian’s radio was still receiving updates. Only one further position SMS was received after the balloon was released, and then Icarus fell silent. After a few minutes, the APRS beacon failed as well. The balloon was lost.

We cleaned up the site in a more somber mood, with the burners automatically falling into a line to scour for discarded trash, and decided to drive down to Los Banos, about 50 miles south, in the hope of being closer to the vicinity of any eventual landing. Almost an hour after launch, when halfway to our destination a single APRS packet appeared on the digipeaters, which told us we’d achieved our mission by ascending to 60,901 feet, but also spelled out the doom to which we’d sent our faithful electronic payload.

The automatic beacon reports not only GPS position, but also the temperature inside the cooler and input battery voltage. That single packet – the only one that APRS digipeaters received for the whole flight – told us that the temperature was critically low, at -21C – water vapour would have condensed onto the electronics and frozen, and the internal chemistry in the batteries would have slowed to the point of failure. We had a 9V lithium battery for power, and the tracker (which also powered the GPS) required a 6.8V minimum to maintain its internal voltage. The reported voltage was only 6.7V – and the balloon was still climbing.
We continued south, with each car’s occupants adding to their own ever-growing list of flaws with our process and equipment, desperately hoping for another update from the balloon, but to no avail.
Eventually we arrived at a Denny’s, and stopped. The sun was setting, and temperature falling quickly. We plugged in the depleted battery to charge in the hope that another working radio might help find a signal, and sat down to start listing all of the things that we’d need to fix for the next attempt. Just as the waitress came by to take our order, my phone beeped.
Icarus, amazingly, had survived our abysmal failure at providing environmental isolation, and reacquired a wireless signal. Its first SMS told us that the payload was plummeting through 5,500 feet, and less than 20 miles away. With rushed apologies to the staff, we tore outside and back into the cars, as Icarus kept us updated every minute with its slowing descent and final resting place.

Half an hour later, in pitch darkness, we were deep inside a goose farm, most of the team standing watch for farmers with shotguns while the intrepid few bravely ventured forth through fertile fields, ankle-deep mud, and throngs of potentially man-eating fowl to successfully retrieve the payload – in pristine condition.

So many things went wrong – the effort that we put into the APRS side vastly outweighed the results. The camera failed partway down due to ice melting inside the electronics. Nothing inside the cooler was secured, and only luck kept us out of turbulence sufficient to have components damage one another. But Icarus was a blinding success, and the sensor data, photographs and video that we captured are just amazing. Accelerometer readings indicate a peak altitude somewhere between 67,000 and 70,000 feet.

We’ll work out the insulation, and the APRS gear. Icarus will soon be in the Android Marketplace for others to use. Later, we’ll send up solar panels, model rockets and radiation detectors, control balloon altitude to steer across countries and oceans with prevailing winds, and descend to a specific spot with a UAV.

But right now we have another full tank of helium.

Next launch is mid-March.

Photos from Andrew, Christie, and Mikolaj; join the spacebridge mailing list to help out with future adventures. Also available is raw sensor data from the GPS, accelerometer and magnetometer on the G1, 180MB of photos taken by the camera (including the image above), and the KML track for use in Google Earth. The complete video taken by the downwards facing camera is also available.

Categories: Spacebridge, Syndicated