Inside there is a heartbeat. A loud whoosh-a-whump. It's not the engine itself. Brophy explains that it's a giant refrigerator pump that produces a space-like vacuum, making it possible to test ion engines here on Earth.

That's a limit NASA wants to find.

So before Deep Space 1 launched in October 1998, the twin engine here in Building 148 was fired up. It is monitored and controlled by a dinosaur of a computer, a Macintosh Quadra 650. On a head-high rack that has probably been around since the 1950s, modern LED readouts glow in red next to ancient-looking black-and-white dials that resemble a car's fuel gauge.

The sound of the heartbeat grows as Brophy opens a door near the rack leading into a dimly lit and cramped room with dripping overhead pipes. The sausage-shaped vacuum chamber fills an area covered but open to the outdoors on two ends.

Brophy raises his voice to speak over the whoosh-a-whump and explains the dripping pipes. "They're really cold, and the insulation is not perfect, so it condenses the water from the air and it drips."

He points at the vacuum chamber, which was assembled from three pieces picked up from a mothballed government facility. There is a porthole window at eye-level. A peek through the glass is like a glimpse of science fiction. An otherworldly blue glow is all that's visible of the ion engine. The glow represents electrically charged atoms shooting out the back of the engine.

A spacecraft propelled by an ion engine uses the Sun as its energy source. Solar cells convert sunlight into electricity to run the spacecraft's scientific instruments and to power the engine.

"The fuel is gas that occurs naturally in the atmosphere, called xenon," Brophy explains. "It's an inert gas, so it's environmentally safe. We run an electric current through the gas, and that tends to charge it up positively. So you get charged particles, which are called ions. We apply a voltage to the engine, which is also positive, and so the positive engine pushes the positive ions out the back of the device at a very high speed, typically 35,000 meters per second, or about 78,000 miles per hour."

Because the thrust is gentle, it takes months or years to get a spacecraft to full speed. The test engine, anchored inside the vacuum chamber, isn't going anywhere.

While the ion engine on Deep Space 1 is unquestionably an advanced technology, the basic idea behind it is nothing new.

Scientists figured out in the early 1900s that the faster you push exhaust out of a rocket engine, the less propellant you need to carry. The equation is crucial for space flight, where weight equals cost, and propellant equals more weight.

With an ordinary chemical rocket, on a space shuttle for example, there's a fundamental limit to how fast the exhaust can be pushed out. The limit is related to how much energy is contained in a given type of fuel.

Ion propulsion redefines this limit, making it easy to get high exhaust velocities, Brophy says.

Scientists had this much figured out by 1960, when the first ion engine was built and tested at NASA's Glenn Research Center (GRC). Early attempts involved substances that required heating to turn them into gases. After exiting the ion engine, the atoms from these gases would cool and condense on the exterior of the spacecraft.

"What turns out to be very hard with an ion engine is to make it last long enough to be useful," Brophy says. Engineers at GRC eventually worked out the kinks, leading to the engine on Deep Space 1, which has run for more than 13,000 hours.

What turned out to be harder was convincing mission managers that they should use the technology, and that they should therefore give money to Brophy and his colleagues at GRC and JPL so they could develop the engines.

Until the past decade, a NASA mission manager was not accountable for the cost of the launch vehicle that would send a spacecraft on its way. So no manager was inclined to waste his precious budget and take on the added risk of developing a new but cheaper overall method of reaching a destination.

This left ion propulsion research languishing for decades.

"Because [ion propulsion] wasn't available, future missions couldn't plan to use it," Brophy said. "But management said they weren't interested in developing the technology, because no future missions were planning to use it."

All that changed when Dan Goldin, who became NASA administrator in 1992, required that deep space missions account for their full cost, Brophy said. Suddenly, ion propulsion sounded better. Now instead of an expensive, powerful initial thrust that would allow a spacecraft to coast to its destination, a cheaper initial boost can be given to a craft that then gradually builds up speed.