For the Love of Physics (33 page)

Once we were ready to go, we needed to have winds under 3 miles per hour in a steady direction for three or four hours, which is how long it took to get the balloon off the ground (the inflation alone took two hours). That’s why we mostly launched at dawn, when there was the least amount of wind. But it could happen that our forecast was wrong, and we just had to wait, and wait, and wait some more, until the weather cooperated.

We were in the middle of a launch one time in Mildura—we had not even started inflation—and the wind came up, contrary to the weatherman’s forecast. The balloon was destroyed, but thank goodness the telescope was safe! All that preparation, and $200,000—gone in a few seconds. Talk about painful. All we could do was wait for better weather and try again with our spare balloon.

The failures stick with you. On my last expedition to Alice Springs we lost two balloons in a row right at launch, because the launch crew made some tragic mistakes. Our expedition was a complete failure—but at least our telescope wasn’t damaged. It never left the ground. On my last expedition (in 1980), in Palestine, Texas, the eight-hour flight was successful, but when we terminated the flight by radio command, we lost our telescope; the parachute never opened.

Even today, balloon launches are far from a sure thing. In an attempted NASA launch from Alice Springs in April 2010, something went wrong and the balloon crashed while trying to take off, destroying millions of dollars worth of equipment and nearly injuring onlookers. You can see the story here:
www.physorg.com/news191742850.html
.

Over the years I must have launched about twenty balloons. I had only five that failed during launch or didn’t get to altitude (they may have been leaking helium). That was considered a good success rate (75 percent). In the insert you can see a picture of the inflation (with helium) of a balloon and also a picture of a balloon launch.

Months before going to the launch site, we would test the payload at a firm in Wilmington, Massachusetts. We put the telescope into a vacuum chamber and brought the air pressure down to the same we’d have way up high, about three-thousandths of an atmosphere. Then we cooled it down to minus 50 degrees Celsius (–58°F) and ran it—turning on all the X-ray detectors and monitoring for ten seconds every twenty minutes X-rays from a radioactive source for twenty-four hours straight. Some of our competitors’ telescopes—yes, we did feel like the other teams doing the same kinds of things were our competition—would fail sometimes because their batteries would lose power at low temperatures or quit altogether. That never happened to us because we had tested them so thoroughly. If we saw in the testing period that our batteries were going to lose power, we figured out how to heat them up if necessary and keep the power going.

Or take the problem of corona discharge—sparking from high-voltage wires. Some of our equipment ran on high voltage, and very thin
air, where the pressure is very low, is an ideal environment for sparks, from wires into the open air. Remember the buzz around transmission lines I mentioned back in
chapter 7
? That’s corona discharge. Every experimental physicist who works with high voltage knows you can get corona discharge. I show examples of these sparks in my classes. There, corona discharge is fun. At 145,000 feet, it’s a catastrophe.

In lay terms, the equipment would start to sputter, and you would get so much electronic noise that you couldn’t pick out the X-ray photons. How big a disaster would this be? Total and complete: you would get no usable data at all on a flight. The solution was to coat all of our high-voltage wires in silicon rubber. Other folks did the same thing and still got corona discharge. Our testing and preparation paid off. We
never
had corona discharge. This was just one of dozens of complex engineering issues involved in building these intricate telescopes—that’s why they took so long to build, and cost so much money.

So, once we got the telescope high up into the atmosphere, how did we detect X-rays? The answer to this question is not simple, so please bear with me. To begin with, we used a special kind of detector (sodium iodide crystals), not the proportional counters (filled with gas) the rockets used, but something that was able to detect X-rays with energies higher than 15 keV. When an X-ray photon penetrates one of these crystals it can kick an electron out of its orbit and transfer its X-ray energy to that electron (this is called photoelectric absorption). This electron in turn will produce a track of ions in the crystal before it comes to a stop. When these ions get neutralized, they release energy mostly in the form of visible light; thus a flash of light is produced—the energy of the X-ray photon is converted into a light flash. The higher the energy of the X-rays, the stronger the light flashes. We used photomultipliers to detect the light flashes and convert them into electric pulses: the brighter the light flash, the higher the voltage of a pulse.

We then amplified these pulses and sent them to a discriminator, which measured the voltage of the electric pulses and sorted them according to magnitude—which indicated the energy levels of the
X-rays. In the early days we recorded the X-rays at only five different energy levels.

So that we would have a record of the detections after the balloon flight, in the early days we recorded them on board, by energy level and the time they were detected. We wired the discriminator to send these sorted impulses to light-emitting diodes, which created a pattern of flashing lights at those five distinct energy levels. Then we photographed those flashing lights with a camera running continuous film.

If a light was on, it would make a track on the film. All together, the film of an observation would look like a series of dashes and lines, lines and dashes. Back at MIT we would “read” the film with a special reader designed by George Clark that converted the lines and dashes to punch tape: paper tape with holes in it. Then we read the punch tape with light sensitive diodes and recorded the data on magnetic tape. We had written a program on computer cards in Fortran (I realize this sounds prehistoric) and used it to read the magnetic tape into the memory of the computer, which—finally!—gave us X-ray counts as a function of time in the five different energy channels.

I know it sounds like a Rube Goldberg machine. But think about what we were trying to do. We were trying to measure the counting rate (the number of X-rays per second) and energy levels of X-ray photons, as well as the location of the source that had emitted them—photons that had been traveling for thousands of years at the speed of light, spreading through the galaxy and thinning out continuously by the square of the distance they traveled. And unlike a stable mountaintop optical telescope whose control system can keep the telescope trained on the same spot for many hours and can return to the same spot night after night, we had to make use of whatever time we had (at most once per year)—always measured in hours—while a fragile balloon carried our thousand-kilo telescope 145,000 feet above the Earth.

When a balloon was in flight I followed it in a small plane, usually keeping it in sight (in the daytime, that is—not at night), flying at just 5,000 or 10,000 feet. You can imagine what that was like, for many hours
at a time. I’m not a small man. It was easy, all too easy, to get sick in these little four-seater planes, flying for eight, ten, twelve hours at a time. Plus, I was nervous the whole time the balloon was up. The only time you could relax was after the recovery, when you had the data in hand.

The balloon was so enormous that even though it was nearly 30 miles up, when sunlight hit it, you could see it very clearly. With radar, we could follow it a long way from the launching station until the curvature of the Earth would make that impossible. That’s why we outfitted the balloon with a radio transmitter, and at night we had to switch exclusively to tracking the balloon by radio beacon. No matter how hard we worked getting articles in the local newspapers about the launch, the balloons could drift hundreds of miles, and when they were aloft we’d get all kinds of reports of UFOs. It was funny, but it made perfect sense, really. What else were people supposed to think when they caught a glimpse of a mysterious entity in the sky of indeterminate size and distance? To them it really was an unidentified flying object. You can see a picture taken with a telescope of a balloon at 145,000 feet in the insert.

Even with all our planning, and weather forecasts, and even in turnaround, the winds at 145,000 feet altitude could turn out to be unreliable. Once, in Australia, we had expected the balloon to head north from Alice Springs, but instead it took off straight south. We followed it visually until sunset and kept radio contact with it through the night. By morning it was getting too close to Melbourne, and we were not allowed to enter the air space between Sydney and Melbourne. No one was going to shoot it down, but we had to do something. So when our wayward balloon was just about to reach forbidden air space, we reluctantly gave the radio command that cut the payload loose. Separating the telescope from the balloon would shatter the balloon—it could not survive the shock wave caused by the sudden release of the payload—and the telescope would start to fall, the parachute would open (except in 1980) and slowly float down, bringing the telescope safely back to Earth. Huge pieces of the balloon would also hit the ground, usually spread out over an acre or more. This occurred sooner or later in every balloon
flight, and it was always a sad moment (even though it was always necessary), because we were terminating the mission—cutting off the data flow. We wanted the telescope to be aloft as long as possible. We were so hungry for data in those days—that was the whole point.

Recovery in the Outback: Kangaroo Jack

We put cardboard crash pads on the bottom of the telescope to soften its landing. If it was during the day, and we had visual contact with the balloon (which would suddenly disappear when we sent the cut-down command), we would soon spot the parachute; we did our best to follow it all the way down, circling it in our little airplane. Once it landed we would mark its location on a very detailed map as accurately as possible.

Then the really bizarre part started: because here we were, in an airplane, and our payload, with all our data, the culmination of years of work, was lying on the ground, almost within reach, but we couldn’t just land in the middle of the desert and get it! What we had to do was to draw the attention of local people, and the way we usually did this was by flying a plane low over a house. Houses were pretty far apart in the desert. Residents knew what the low-flying plane meant and usually came out of the house and made contact by waving. Then we would land at the nearest airstrip (not to be confused with an airport) in the desert and wait for them to show up.

During one flight, there were so few houses in the area that we had to hunt for a while. Eventually we found this guy Jack living in the desert 50 miles away from his nearest neighbor. He was drunk and pretty crazy. We didn’t know that at first, of course. But we made contact from the air and then flew to the airstrip and waited; after about 15 hours he showed up with his truck, a battered old jeep-like thing with no windshield, just a roof on its cab, and an open bay in back. Jack liked to tear around the desert at 60 miles an hour, chasing and shooting kangaroos.

I stayed with Jack and the truck and one of my graduate students, while our tracking airplane directed us to the payload. The truck needed
to go across unmarked terrain. We kept in radio contact with the plane. We were lucky with Jack. From all that kangaroo hunting he really knew where he could drive.

He also had this awful game I hated, but we were already depending on him, so there wasn’t much I could do; he gave me a demonstration just once. He put his dog on the roof of the jeep, accelerated up to 60 miles an hour, then slammed on the brakes, and the dog catapulted through the air onto the ground. The poor dog! Jack laughed and laughed and then delivered his punch line: “You can’t teach an old dog new tricks.”

It took us half a day to reach the payload, which was being guarded by a six-foot-long iguana—a really nasty-looking creature. To tell the truth, it scared the hell out of me. But of course I didn’t want to show that, so I said to my graduate student, “There’s no problem. These animals are harmless. You go first.” And he did, and it turns out that they
are
harmless, and during the entire four hours it took us to recover the payload and get it on Jack’s truck, this animal never moved.

The Balloon Professor

Then we went back to Alice Springs, and of course we were on the front page of the
Centralian Advocate
with a great photograph of the balloon launch. The headline read S
TART OF
S
PACE
P
ROBE
and the article talked about the “balloon professor” having returned. I had become a sort of local celebrity and gave talks to the Rotary Club and for students at the high school, even once in a steak house, which earned me dinner for my crew. What we really wanted to do was take our film back home as quickly as possible, develop and analyze it, and see what we’d found. So after a few days’ cleanup we were on our way. You can see just how demanding this kind of research was. I was away from home for something like two months at least every other year (sometimes every year). And there’s no question about it that my first marriage suffered a lot because of it.

At the same time, despite all the nervousness and tension, it was
exciting and great fun and I was proud of my graduate students, notably Jeff McClintock and George Ricker. Jeff is now senior astrophysicist at the Harvard-Smithsonian Center for Astrophysics and won the 2009 Rossi Prize (named for guess who?) for his work measuring the masses of black holes in X-ray binary star systems. (We’ll get to that in
chapter 13
.) George, I’m happy to say, still works at MIT. He is brilliant at designing and developing innovative new instrumentation. He is best known for his research in gamma-ray bursts.