Light fuse. Step away.

…But not necessarily in that order, when you’re dealing with the world’s largest solid rocket motor. In fact, engineers who tried last Thursday to light the ATK five-segment motor planned for NASA’s Ares I rocket, were in an underground bunker half a mile away when ignition was to occur at a quar…

...But not necessarily in that order, when you're dealing with the world's largest solid rocket motor. In fact, engineers who tried last Thursday to light the ATK five-segment motor planned for NASA's Ares I rocket, were in an underground bunker half a mile away when ignition was to occur at a quarter past 1:00 in the afternoon (15 minutes past the planned time). I was standing above ground about a mile away under a brilliant Utah sun with temperatures in the mid-90s and the wind blowing away from us. This was to be the first static firing of the five-segment booster derived from the space shuttle's twin, four-segment boosters. The new booster will loft NASA astronauts to orbit after the shuttle retires.

Instead, a hold was called with just 20 seconds to go, and an announcer shortly informed a crowd of 50 journalists and 500-plus NASA and ATK brass assembled at Promontory, Utah, two hours northwest of Salt Lake City, that the test was canceled. The problem was a faulty valve that feeds an auxiliary power unit that spools up a few tens of seconds prior to ignition and drives a hydraulic system in the booster's aft skirt that gimbals the exhaust nozzle, which is how the rocket steers itself in flight. It was a critical part of the test, as engineers were going to run the nozzle, which is 18 inches longer than a shuttle booster nozzle, through some vigorous motions to see how it holds up under 24 percent more thrust than that produced by a shuttle booster.

In the press conference an hour later, one journalist asked a slightly weary-looking Pat Lampton, NASA's chief engineer on the program, about all the "bureaucrats" who would be involved in the Monday-morning quarterbacking. He calmly replied, "Fortunately, at NASA, most of the bureaucrats are engineers."

What it should have looked like. Credit: ATK

By 5:00, ATK dashed any remaining hopes for a test the next day. There were plenty of murmurs throughout the press corps about the timing of the test: A matter of days before the Augustine panel's end-of-the-month  deadline to the White House, with a printed report ready by late September. The panel's conclusions were becoming an open secret: Not only will NASA's paltry 18-billion-dollar budget fail to return astronauts to the moon by 2020, but we may need a cheaper rocket, such as Elon Musk's Falcon Nine, or Deltas and Atlases that have been building a strong record lofting satellites. That's not what NASA and ATK want to hear after developing Ares I for the past five years.

As luck would have it, the space shuttle Discovery blasted off the next day, just before midnight on the 28th. On the 29th, ATK distributed a press release stating, "More than 100 RSRM (reusable solid rocket motor) flight sets have been launched to date, marking a two-decade track record of flawless performance." Blake Larson, ATK Space Systems President, was quoted in the release, saying, "The launch of the Space Shuttle Discovery and the upcoming fall launch of the Ares I-X , highlight the capabilities and progress ATK and NASA have made in developing the most reliable and affordable family of solid rocket motors ever produced....They will continue to be instrumental to the success of the remaining shuttle flights as well as the future human spaceflight programs."

Earlier that day, we'd been piled into a bus and carted up the hill for a closeup look at the 12-foot-diameter, 154-foot-long booster, bolted horizontally to the ground for the static test. These ground connections keep it still while the real anchor, a house-size block of concrete weighing several million pounds, endures 3.6 million pounds of thrust via an Erector-Set-like assembly of steel beams sandwiched between the nose of the motor and the concrete. We wouldn't have seen the rocket move forward, but it would have, about an inch at most points along its frame. If that flex weren't designed in, the motor might rip itself apart. Its five reusable segments are durable—they've flown in space a total of 48 times. The rear segment flew on the very first shuttle mission in April 1981.

Twenty yards behind the exhaust cone, a mirror roughly eight feet square awaited obliteration atop a metal pole that emerged from its own block of concrete set firmly in the ground. Cameras forward of the motor are pointed at the mirror and record the first milliseconds of the ignition as the gases roar out the nozzle. Milliseconds later, the mirror gets annihilated, its pole and concrete pedestal uprooted and flung 50 yards up the barren hill where others lay after earlier tests. A large, foil-wrapped swing arm with a CO2 delivery system stood ready to insert a probe into the nozzle immediately after the test to cool the motor's interior. This substitutes for frigid altitude and seawater that cools an operational booster. A six-inch blanket of gray sand covered the entire concrete terrace behind the motor. Some of the sand would be blown up onto the hillside, but much of it would remain and turn to glass in the 6,000-degree heat.

All I saw on Thursday was the sleeping rocket, not the one belching fire and smoke. I was still impressed. This hydraulic glitch had not caused a scrub of a static test or a shuttle launch that anyone could remember. Murphy's Law had followed me to Utah.

But I had to remind myself that the one shuttle launch I've witnessed, a night launch with a full moon hanging above the Atlantic in a cloudless November sky, had gone off without a problem. I'll admit, I'd rather see the shuttle fly any day than a rocket motor chained to the desert floor.

Get the latest stories in your inbox every weekday.