Original article here

In an era of satellites, Doppler radar, and supercomputers, a weather balloon seems like a quaint artifact of the 1950’s, an icon that keeps company with vacuum tubes, black & white TV, and Sputnik. In the hands of a new generation of Berkeley Lab scientists, however, the venerable gas bag is a tool for the future.

During spring break on the UC Berkeley campus, a small group of Berkeley Lab researchers led by atmospheric chemist Odelle Hadley conducted the first test run for a project to measure layers of soot in the atmosphere — fingers of air laden with particles of black carbon that are part of the complex puzzle of human impact on global climate. On a stretch of green grass at Memorial Glade, these researchers literally floated a trial balloon.

Expecting a day of Fun Science, I found not only the Fun — it was a balloon launch, after all — but also an afternoon of insights about how difficult, detail-oriented, frustrating, and ultimately rewarding the simplest scientific experiment can be.

Weather balloons, I learned, were never retired. At meteorological stations all over the globe, these goofy latex orbs are still used for direct observation of atmospheric conditions. Hadley’s team — which includes her boss, climate scientistThomas Kirchstetter, and engineering grad student Daniel Wilson — hopes eventually to launch more than a dozen a year to measure soot concentrations from ground level to the stratosphere. They are recipients of an initial $140,000Laboratory Directed Research and Development (LDRD) grant, a small part of an internal Berkeley Lab program that this year allotted $19 million to seed innovative science and new research directions.

With direct measurements, the balloon data can validate computer modeling of sooty layers and the remote sensing data from satellites. It can do so at a fraction of the cost of aircraft-borne equipment.

Wilson, whose engineering credits already include design of brake actuators for the Boeing 787, became the self-taught weather balloon set-up man. “I watched a lot of YouTube videos,” he told me, “but this is the first time I’ve ever done this.” Wilson also designed the payload; a package of electronics carried in a white Styrofoam cube that otherwise might have been used to ship a fresh lobster via FedEx. Inside the cube, the 1950’s are left behind. If the computer power packed into it were launched fifty years ago, the balloon would have had to lift a small building. The entire weight of this package: about four pounds.

The white cube is equipped with a radiosonde, which collects and transmits atmospheric data. It is wedded to a circuit board that stores more data on a flash memory card and decides when to cut the payload loose with a small gunpowder charge. The package is designed to fall from about 7 miles on a small, colorful parachute. Also on board: a palm-sized, store-bought GPS tracking unit, which will signal the chase team where the payload lands; and a pair of digital cameras to record the flight.

This first launch was designed to test components and protocols; from filling the 4-foot wide balloon with helium to setting off a small firecracker’s worth of gunpowder to release the payload. This balloon would remain tied to some heavy-duty cotton string, and would fly only 100 feet above Memorial Glade.

Wilson’s 30-minute procedure to fill the balloon from a heavy tank of helium outside a UC Berkeley engineering building went flawlessly. A white plastic bucket filled with water simulated the weight of the payload. When the balloon lifted the bucket, it was ready to be carried to the launch site — to the delight of the handful of bystanders. “There’s something about balloons,’’ said Hadley. “People love them. Just talk about them, and people perk-up and start smiling.”

Then the team encountered Murphy’s Law. At Memorial Glade, the tracking program was failing to write test data onto the memory card. Had this been a data-gathering flight, it would have been scrubbed. For this trial, it wasn’t needed. Wilson made a note: find the bug.

The launch was perfect, and Kirchstetter carefully played out more than 100 feet of line. Hadley paced nervously. “I don’t like explosions of any kind,” she confessed. Then, with a soft POP up in the sky, the white package dropped….about three feet.  Somehow, it was still hanging on to the balloon. Flabbergasted, the team reeled down the entire dangling contraption. A connector made from a pair of nylon eyebolts, each screwed into an end of a small plastic tube filled with a few grains of gunpowder, had properly snapped-in-two when a model rocket igniter fired. But the team discovered that they had accidently installed the connector upside down. So after the successful firing, payload and balloon remained tethered to the thin detonator wire. Another note: wrap a ring of electric tape on connector to mark “this end up.”

Round 2. As Wilson begin to wire a new connector to the circuit board, POP! A premature detonation. The charge had been secured at a safe distance before the wire was touched, but the failure startled everyone. Wilson immediately guessed what had happened: a power transistor had switched to the on position after the first test.  Another note: test the transistor with a voltmeter first.

Round 3. A smooth set-up, a voltage check, and the balloon successfully launched. The connector was set to blow. Fifteen minutes of silence, and a disappointed team pulled the balloon back down again. Diagnosis: Dud. A faulty ignition wire. Another note recorded.

Round 4. Wilson was visibly tense: Engineering pride was on the line. The balloon rose, high above the green grass.  Wilson called out, “One minute to go.”  Spectators watched transfixed. Then, POP! and the package started to fall. The chute deployed. Everyone cheered, and the payload bounced lightly on the lawn. Wilson was whooping, dancing in air, as if Cal had just beaten Stanford.

It had seemed at times like a comedy of errors. Yet here was a valuable lesson in how real science actually works.  This is why procedures are tested methodically in the first place. Their day done, the team members packed up their gear, armed with experience and a list of changes.  “I think it was great,’’ said Hadley, three days later. “It really gave us an idea how to set up for next time. We already solved the software bugs. We have a new protocol. We know our instruments can survive. We won’t make the same mistakes twice.”