How to Build an ”Affordable” Space Plane: Design, Test, Repeat

Take off from a typical airstrip, fly a nearly vertical maneuver to 200,000 feet at twice the speed of sound, make a slow backflip, take a few snapshots of the earth while you linger a few minutes at zero G, and spiral down safely back to the runway.  

Refuel and repeat.

That’s the 30-minute flight plan for the Lynx, a two-seat suborbital air/spacecraft developed by XCOR Aerospace ( www.xcor.com ; Mojave, CA), one of a handful of companies chasing the dream of private space travel.  Their reusable craft will fly several times a day, taking sightseers to the edge of space--for a $95,000-ticket.

Wind tunnel testing is currently underway, and plans call for a dress-rehearsal flight in mid-2010.  The first tourist flight hasn’t been scheduled.  But how does a relatively new private company--funded with venture capital investors, and government and commercial contracts--possibly compete with the would-be titans of space tourism, including mogul Richard Branson and his Virgin Galactic effort?  To hear XCOR’s Jeff Greason tell it, it all starts with a completely different kind of product development mindset than what he sees as the prevailing aerospace culture: “We wanted to think of the rocket engine as you would a car engine,” says Greason, XCOR CEO and co-founder. “Rocket engines should not have to require 500 guys with German accents and white coats crawling all over them to work well.”

Greason says his vision for a lean rocketship development company wasn’t the industry norm.  In fact, his foray into aerospace was something of a culture shock compared with his time at Intel, where he managed a team developing new generations of chips designed specifically to shorten the time from architecture to production to customer.  In 1997, he left Intel to manage the propulsion team at the now-defunct Rotary Rocket, which was developing a single-stage craft that used a helicopter rotor to take off and a rocket to break through the atmosphere.  At Rotary, one group handled structural design, another was on fabrication, another on thermal dynamics, and another did testing.  Often, these silos were geographically disbursed, flight tests were largely done off-site, and communication was limited to interface documents instead of face-to-face discussions.  Standard operations included mountains of data before any testing could take place and months of post-mortem work on what those tests ultimately revealed, Greason notes.

After Rotary disassembled its propulsion team in 1999 and ultimately ceased operations a few months later, Greason and three other Rotary engineers started XCOR, where he vowed to follow an operations model closer to Intel than to his predecessor. “We got rid of all the title distinctions,” he says. “We have one design team.”  

That means during a rocket firing test, a thermal engineer developing the chamber sits next to the machinist who will cast the parts.  Gone are the data silos, which even in a company of XCOR’s size – 25 workers – can crop up given the scale of their task.  This strategy has enabled XCOR to build 10 from-scratch rockets in 10 years and conduct more than 3,600 hot fire tests on them.

“You never want to be stupid, but ultimately you have to experiment and see what the result is.  Because no matter how carefully you plan, there are always surprises,” he says.  “By doing that, we could get the cycle of design, build, test fire and learn down from months to days, or in some cases less than a day … Even with all the CAD tools available, rocket engines are devices with a lot going on.  So why spend three months on something you can only know for certain in a test?  Design the thing within 25% accuracy and then tweak the model.”

XCOR’s longest-running rocket underwent 2,000 test firings before the team deliberately killed it by shutting off its cooling to see how it would have responded without it.  That 2,000 firing number is important to keep in mind, considering the Lynx rocket isn’t a throwaway propulsion device, but one that’s integrated directly into the craft, which the company intends to launch multiple times a day once it’s operational.

In December 2008 XCOR conducted the first test fire of the Lynx rocket, known as the 5K18, which generates 2,500 to 2,900 lb-ft. of thrust by igniting liquid oxygen and kerosene.  The Lynx’s final engine will be fed using the company’s design: a cryogenic piston pump for liquid oxygen and a similar mechanism for the kerosene, which they say are critical to affordable space journeys.

Because the rocket design dictates just about everything else on the flight vehicle, getting it right helps to minimize the risk of numerous craft design iterations, Greason says.  Unlike the rocket engines, in many respects, the craft design harkens closer to the supersonic pilot-operated jets of the past than the multi-stage, Mission Control (and controlled) rockets that would come later.  “We’ve actually had to go back a generation and unearth design techniques where you to focus on the flying qualities of the vehicle,” Greason says. “It has not been an easy process … because the Lynx has to fly through a range of speeds and altitudes and angles of attack.”

The Lynx is the third-generation spacecraft from XCOR.  The first proof-of-concept rocket jet from the company, the EZ-Rocket, modified a home built “Long-EZ” aircraft with the addition of two 400-lb. thrust rocket engines.  The second-generation, called the “Rocket Racer,” was modified to carry the 1,500-lb. thrust rocket engine, which burns liquid oxygen (LOX) and kerosene.  The rear seats were removed to accommodate a LOX tank near the center of gravity.  XCOR then integrated a new rocket propellant piston pump in the Rocket Racer, which eliminates the need for the pressurized fuel belly tank seen on the EZ-Rocket.  Overall, the company has flown these vehicles 66 times from its Mojave flight center, which according to XCOR, accounts for more than half of all manned rocket-powered flights in the 21st century.

All of that early testing influenced the Lynx layout, which initially mimicked a two-seater F-16 – with the pilot up front and the passenger directly behind.  That was scrapped in favor of a side-by-side configuration (with the tourist flying co-pilot), but that made the craft too wide.  The final design is a staggered arrangement, with the pilot seated somewhat more forward, but still next to the passenger.  Apart from its sharply concave fuselage, the Lynx closely resembles a private jet.  In fact, XCOR plans to manufacture and sell them, along with the ground control services to manage spaceflights, at a price lower than Gulfstream 5, the going rate for which is $60 million.