Graham Warwick/WASHINGTON DC

Lockheed Martin's venture to develop and operate a fully reusable launch vehicle has always been acknowledged as risky; just how risky was brought home by the news that engine manufacturing problems have delayed the first flight of the X-33 technology demonstrator by six months, to December 1999.

Although Boeing Rocketdyne's linear aerospike rocket motor will now have to be installed on the launch pad at Edwards AFB, California, if the goal of a first flight by the end of next year is to be met, Lockheed Martin and NASA continue to express confidence in the X-33. They have to.

Lockheed Martin still plans to stand up early next year the limited-liability company that will attempt to raise the $5 billion required to build and operate the follow-on VentureStar reusable launch vehicle (RLV). NASA is counting on the VentureStar to reduce the burden on its budget of supporting the International Space Station (ISS), as the goal of the RLV programme is to reduce launch costs by a factor of 10.

Before Lockheed Martin can raise the commercial finance for Venture Star, and NASA can begin saving on launch costs, the X-33 must demonstrate that critical technologies for a single-stage-to-orbit RLV can be developed with acceptable risk.


News of engine problems cuts to the heart of the programme. Rocketdyne's linear aerospike rocket motor is one of the key design features of the VentureStar. Because it is integrated into the lifting-body vehicle, spreading the thrust load and essentially eliminating base drag, it reduces structural weight and helps makes an SSTO RLV feasible.

The X-33 will be the first vehicle to be powered by a linear aerospike engine. The full-scale VentureStar will require development of a much more powerful version of the engine, with more than double the thrust-to-weight ratio. Difficulties building the relatively low-performance X-33 powerplant, therefore, could concern potential investors.

Rocketdyne's latest problems came when it tried to produce the curved nozzle ramps which form the structure of the engine. The aerospike engine has several small combustion chambers mounted in rows at the forward ends of the upper and lower ramps. These thrusters exhaust on to the ramps, and the plumes adjust automatically with altitude, expanding as the ambient pressure decreases.

This results in greater propulsive efficiency than a conventional rocket motor with a bell nozzle, in which the plume is under-expanded at low altitudes and over-expanded at high altitudes. In an aerospike engine, the ambient pressure influence matches the plume expansion to the altitude and automatically optimises performance, says X-33/VentureStar programme manager Jerry Rising.


Although the aerospike engine is described as "medium risk" based on ground-test experience, producing the powerplant has proved problematic. In the X-33's 190,000lb-thrust (845kN) Rocketdyne XRS-2200 engine, the nozzle ramps are made from copper.sheet. Grooves are milled into the back side of the ramp and a steel-alloy sheet brazed on to create passages through which liquid hydrogen is pumped to cool the ramp.

Difficulties encountered bonding the steel backing on to the copper resulted in the scrapping of the first ramp early this year. Rocket-dyne successfully produced its first ramp in September, and although eight ramps are required for the X-33's four engine modules, the company believes it can now deliver the powerplant by September next year. This represents a delivery slip of four months, on top of an earlier delay of three months.

Rising says Lockheed Martin still plans to roll out the X-33 in July, but will now install the engine at the Edwards AFB launch site. A mobile hangar which can be rolled over the launch pad will allow the installation to be performed on site. Ground vibration testing required after the vehicle is delivered to Edwards will be accomplished using mass simulations of the engine, says Rising.

Other testing required before the first flight is also running. Ground tests of the turbomachinery "powerpack" for the X-33 engine finally got under way at NASA's Stennis Space Center, Mississippi, in October. These involve the liquid hydrogen and liquid oxygen turbopump pair that supply the banks of thrusters in each engine module. The XRS-2200 uses turbomachinery originally developed for the Saturn J-2 rocket motor.

Another test that has been delayed is the first airborne firing of a linear aerospike engine, mounted on the back of NASA's Lockheed SR-71 testbed. This subscale engine has been "cold fired" in flight, when propellants were pumped through but not ignited. "Hot fire" tests have been delayed by "persistent" liquid oxygen leaks, says Rising. Safety reviews are under way, and he is hopeful of an inflight test firing before year end. Although not critical, these tests are needed to validate performance predictions for the linear aerospike engine.

As part of the $1.53 billion RLV technology demonstration programme agreed by NASA and Lockheed Martin in July 1996, a prototype of the larger VentureStar engine will be ground tested. The programme costs are being shared 80:20 by NASA and the Lockheed Martin-led industry team, but the space agency's share is capped at $941 million. Lockheed Martin has already had to increase its investment in the X-33 by more than 50%, to $231 million, and is now looking to its partners to cover the costs of any further problems.

Rocketdyne says the engine problems have increased its costs by $36 million, and the company wants to take that money from the 20% of the overall RLV programme budget that is allocated to development of critical technologies for the VentureStar. This includes the building and testing the higher-performance engine, and any move to transfer funding from this part of the programme will affect development of the VentureStar, which is currently scheduled to become operational in 2004.


The RS-2200 engine planned for the full-size VentureStar represents substantial step forward from the X-33 powerplant. Thrust is increased to 400,000-500,000lb and thrust-to-weight ratio is more than doubled, from 35:1 for the XRS-2200 to almost 80 for the RS-2200 (the Space Shuttle main engine has a thrust-to-weight ratio of 68). This will require a substantial effort to reduce the engine's weight and improve its performance.

Key to achieving the required thrust-to-weight ratio will be the composite nozzle ramp planned for the VentureStar. Rising says Rocketdyne is in the final stages of selecting a supplier for the ramp, which will be the main contributor to reducing the weight of the engine. Efforts under way to redesign the turbomachinery and thrusters "...are taking weight out of the engine," he says.

Another critical technology which was not mature enough to incorporate in the X-33, but is essential if the VentureStar is to succeed, is a composite liquid oxygen tank. The X-33 has carbonfibre composite liquid hydrogen tanks and airframe structure, to minimise weight, but an aluminium liquid oxygen tank had to be installed because "...there are currently no composite materials that are compatible with liquid oxygen," says Rising.

Plans call for a composite VentureStar liquid oxygen tank to be built and ground tested. Concerns over leakage, which led to use of an aluminium tank in the X-33, have also delayed production of demonstrator's liquid hydrogen tanks, now being assembled by Lockheed Martin from fibre-placed composite sections produced by AlliantTechsystems.

The task facing Lockheed Martin in scaling up the X-33 to produce the VentureStar is substantial. The demonstrator is a 53% model of the full-size vehicle, but gross lift-off weight (GLOW) will increase almost tenfold, to 1.18 million kg (2.6 million lb), and cost by a factor of four (the X-33 will cost $360 million to build, Rising says).

Empty weight of the full-scale vehicle will be 9.8% of the GLOW: 6.5% for the structure, 2.6% for the engine, and 0.7% for subsystems. Payload will be just 2% of the GLOW, so reducing weight is critical. "Going from the X-33 to the VentureStar we have to get the weight down," says Rising. Design goals are a 56,000lb payload to low earth orbit; 25,000lb to the ISS; and 18,000lb to geostationary transfer orbit (with an upper stage).

Lockheed Martin has scheduled a decision on whether to proceed with the VentureStar for 2000, but this depends on when - and how - the X-33 flies. "Multiple flights" of the demonstrator will be required before the company is in a position to make a decision, Rising says. The plan is to complete two flights in December 1999, but that will depend on when the first flight takes place, he admits.

The vehicle is now in final assembly at Lockheed MartinSkunkWorks in Palmdale, California. Assembly tooling is 90%, the liquid oxygen tank is in place and composite support structure for the metallic thermal protection system is being attached. The liquid hydrogen tanks are scheduled to arrive late this year, and subsystems deliveries are to be completed in January. This includes the triplex avionics and dual-redundant fly-by-wire flight controls.


Assembly is to be completed in July and, the completed vehicle (minus engine) will be transferred to the Edwards launch site for ground testing. The site will be completed by the end of this month, Rising says,The X-33 will be launched vertically from the Edwards pad to make suborbital test flights to two landing sites: Michael Army Air Field in Utah, a distance of 720km; and Malmstrom AFB in Montana, a distance of 1,500km. As the tests will be overland, narrow flight corridors over lightly populated areas have been selected.

A total of 15 flights are planned, but only seven have been defined: the first five will be to Michael and the next two to Malmstrom. On flights to Michael, the engine will operate for 145s to boost the X-33 to Mach 9 and 165,000ft before it glides to a landing 14min later. Flights to Malmstrom will last 24min, the engine operating for 195s to boost the vehicle to M13.8 and 300,000ft. This compares with the M25 maximum speed planned for the full-scale VentureStar and is "...a little more than half the way to LEO," says Rising.


Eight more flights are planned "...to fill out the flight test matrix," he says. The programme calls for two key demonstrations of the X-33's "aircraft-like reusability, maintenance and scheduling". Lockheed Martin is required to perform two consecutive seven-day turnarounds between flights and one accelerated two-day turnaround. The vehicle is equipped with a health monitoring system, and Rising says the metallic thermal protection system is designed to require inspection and periodic maintenance only.

Demonstrating rapid, routine turnarounds is a key part of validating the VentureStar business plan, which is now being drawn up. The plan is based on performing at least 20 flights a year per vehicle at a launch cost to LEO of $450/kg - almost a tenth that of the Space Shuttle and about half that of the new Evolved Expendable Launch Vehicles that Boeing and Lockheed Martin will begin flying in 2001.

The $5 billion programme cost covers two VentureStars (each with a 100-flight life), two spaceports and the first year of operations until revenues begin to flow in. Commercial launches are scheduled to begin in 2004, if Lockheed Martin gives the go-ahead in 2000 as planned. That decision is predicated not only on successful flights of the X-33, but also on successful negotiations with partners and investors on financing the VentureStar.

Lockheed Martin plans to complete is business plan by year-end and stand up Venture-Star LLC in January to begin negotiations with risk-sharing suppliers and venture-capital providers. The company acknowledges that, compared with developing the technology, constructing a business case for the VentureStar is a task with "...an equal degree of difficulty."

One of the bigger hurdles still to be overcome is getting legislation passed that would provide US Government-backed loans for the programme. This would dramatically reduce interest rates compared with venture capital, and Rising cites as a precedent the fact that similar loans have been made available to US commercial shipbuilders.

During 1999, Lockheed Martin will continue with preliminary design of the full-scale VentureStar, as well as selecting the two launch sites, one in the east and one in the west. Rising says 14 locations are bidding to host the spaceports. The company hopes to start operations with 33-37 launches a year and is aiming for a "modest share" of the projected market.

Source: Flight International