Sometimes it feels like there have been so many false starts, so many cancelled projects or ‘paper planes’ over the last 60 years that you could almost pile them up and walk past the Kármán line. But few have got quite so far and failed quite so frustratingly as the X-33.

In the latter years of the last millennium, it seemed like the X-33, or it’s full size follow-on the VentureStar, were everywhere. Visually synonymous with NASA’s next steps in space, the great new hope for cheap reliable access to space appeared in all it’s CGI glory in every pop-science magazine, every book on space travel, every Discovery Channel documentary.

The X-33 flight article was approaching 90% complete and the launch facilities were ready when the project ground to a halt in 2001 with NASA announcing there would be no further funding from it’s side.

During the early 1980s there was a recognition within NASA that a successor to the Shuttle would be needed early in the new millennium to offer the routine, reduced cost access to Low Earth Orbit (LEO) that the Shuttle had never been able to deliver. A reusable launch vehicle was the preferred option, but one that could avoid some of the problems that had dogged the Shuttle as a result of design compromises during it’s development.

Moving beyond the Shuttle, the lure of Single Stage to Orbit (SSTO) has always been strong, promising a craft that is fully reusable, requiring minimal servicing and flying to and from orbit regularly – it just seems a more elegant solution than dropping huge chunks of expensive metal into the ocean. But SSTO comes at a price. By definition, the entire craft needs to be lifted into orbit, empty tanks and all. It must then return in one piece, so a Thermal Protection System (TPS) is needed for the entire craft. Developing the right materials and carrying that weight poses a far greater challenge than the traditional expendable ‘staging’ route. There are two main SSTO approaches:

  1. Design a craft with a combined cycle propulsion system, meaning that it can ‘breathe’ air as it flies through the atmosphere operating like a jet/ramjet before closing it’s intakes at high speeds and altitudes to become a rocket relying on internal oxidiser for the final push into space. This has the advantage of needing less physical space for propellants and lower takeoff weight for oxidiser and it’s associated tankage, but the disadvantage of being extremely demanding technically in terms of the propulsion technology.
  2. Design a craft that is structurally incredibly lightweight but strong so it can carry all the propellant it needs and rely on rocket power all the way to orbit. This approach is potentially simpler from the point of the propulsion system, but extremely complex in terms of the structure and materials technology required – typically the craft’s structure should account for no more than 10% of takeoff weight for this approach to be viable.

The X-30 NASP
In 1986, President Reagan proposed an Orient Express – an aerospace plane capable of reducing travel times around the globe by flying at hypersonic speeds beyond the atmosphere. This concept became the X-30 National Aerospace Plane (NASP), but had it’s roots deep within the classified budgets of programmes such as Copper Canyon. The X-30 programme aimed to develop a craft that could take off from a runway, reach LEO delivering a variety of payloads and land again like a normal aircraft. To do this it would employ the first approach detailed above: an advanced combined cycle propulsion system fuelled by hydrogen slush. Various designs were proposed, but as the programme progressed a ‘wave-rider’ shape developed with most of the vehicles lift being provided by the body shape.

Rockwell International's concept for the X-30 NASP [Img: NASA]
Rockwell International’s concept for the X-30 NASP [Img: NASA]
The X-30 NASP project was extremely ambitious and was some way beyond the state-of-the-art at the time both in terms of materials and propulsion technology. It also never really shook off it’s ‘Black World’ origins, with some claiming it was simply a public front for military developments in support of Reagan’s Strategic Defence Initiative or even the fabled Aurora spyplane. The programme was eventually cancelled in 1993 without entering the flight hardware phase.

The road to the X-33
Beyond the X-30, NASA continued to look for a way to ensure the United States would have an effective and economical launch system post-Shuttle, while responding to congressional concerns regarding the size of their budget. In 1993, NASA Administrator Dan Goldin initiated the Access to Space study as a way to examine options and propose a path forward that would be within the technical and financial means of the agency, while ensuring the demands of major space infrastructure projects such as the proposed Space Station Freedom could be serviced.

One option was creating a next generation reusable launch vehicle that could utilise new technologies, including some of those developed under the auspices of the NASP. Within Access to Space an Advanced Technology Team developed recommendations pointing towards a practical Reusable Launch Vehicle (RLV) with airline-like operations – echoes of the early shuttle programme – and suggesting a fast-track development process creating technology demonstrators under X-Plane designations to rapidly mature the required technologies.

During the summer of 1994, NASA initiated co-operative agreements with industry partners to examine options for a RLV. Funding would be made available until 1999 for the technology demonstrator, to be known as the X-33. It was proposed that this would then lead to the immediate development of a follow-on, optionally-crewed RLV operated commercially by the private sector with NASA leasing flights. This commercial partnership would, it was hoped, allow NASA to develop it’s next generation launcher without having to shoulder the entire operational burden as it did with the Shuttle.

The Phase I designs from (l to r) Rockwell International, McDonnell Douglas and Lockheed Martin [Img: NASA]
The Phase I designs from (l to r) Rockwell International, McDonnell Douglas and Lockheed Martin [Img: NASA]
Phase I of the project began in March 1995 with teams led by Lockheed Martin, McDonnell Douglas and Rockwell International developing proposals for un-crewed demonstrators. McDonnell Douglas’s proposal looked to develop on the success of their DC-X prototype. This vertical take off and landing craft had impressed during it’s test programme and paved the way for NASA’s interest in composite materials to reduce structural weight. The Rockwell proposal resembled a scaled-down Shuttle Orbiter with a cylindrical fuselage and the trademark double-delta wings.

Lockheed Martin’s proposal presented the most challenging approach. They took a flattened triangular lifting body fuselage and combined this with revolutionary Linear Aerospike rocket engines. It would take off vertically and land horizontally but most of the key technologies it required would be unproven at the outset of development.

And the winner is…
On July 2 1996, Lockheed Martin were announced as the winners of the Phase I competition by Vice President Al Gore. The fact they had chosen to showcase so many significant new technologies within their design seemed to have worked in their favour. It wasn’t the first time NASA chose to do something “..Not because it is easy, but because it is hard…” but it certainly represented an ambitious choice given the short timescale and constrained budgets envisioned for Phase II.

A side by side comparison of initial designs for the X-33 and the VentureStar [Img: NASA]
A side by side comparison of initial designs for the X-33 and the VentureStar [Img: NASA]
The Lockheed Martin X-33 was to be a 53% scale prototype for a proposed operational craft, named the VentureStar. With a first flight date in mid-1999, it would seek to demonstrate and build confidence in three key technologies:

  • The use of lightweight alloys and carbon composite materials for internal tank structures.
  • A metallic tiled TPS which it was hoped could be far more resilient than the Shuttle’s silica tiles
  • Linear Aerospike engines, providing more desirable thrust characteristics across all altitudes than conventional ‘bell’ nozzles.

Although the design and final construction would be carried out by Lockheed Martin’s Skunk Works division at Palmdale, California, much of the other work was sub-contracted to other divisions or third party suppliers around the United States. The Linear Aerospike engines would be developed by Rocketdyne (a division of Boeing), the TPS was to be constructed by B.F. Goodrich and the lightweight carbon composite liquid hydrogen tanks were to be developed by Alliant Techsystems. Developmental support would be provided by NASA centres throughout Phase II and a launch site for the X-33 was to be constructed within the boundaries of Edwards AFB.

A Linear Aerospike engine undergoing testing [Img: NASA]
A Linear Aerospike engine undergoing testing [Img: NASA]
Developmental issues
As Phase II construction began, it wasn’t long before a number of technical issues began to beset the programme. Extensive wind tunnel testing had shown that the craft’s stability would be marginal at low speeds. These poor flight characteristics were further exacerbated when the advanced alloys required for the ramps of the aerospike engines weighed in heavier than expected resulting in a rearward shift in the X-33’s centre of gravity. Re-design work followed, but by now concerns were beginning to grow about the final weight of the craft and the effects this would have on performance.

In May 1997 a joint NASA/Lockheed Martin Tiger Team recommended a weight reduction programme to keep these problems in check, but at the cost of delays and consequent increases in manufacturing and project costs. With these problems addressed, the X-33 passed it’s Critical Design Review in October 1997 and NASA seemed happy with progress.

Work began on the launch facilities at a site near Haystack Butte at Edwards AFB, and plans were devised for the flight test phase which would consist of a series of sub-orbital flights building-up to speeds of Mach 15 and altitudes exceeding 50 miles. These full-speed flights would use Malmstrom AFB, Montana as their landing site, but this decision caused concerns regarding the wisdom of allowing sub-orbital flights over populated areas. These concerns led to enquiries and risk assessments, again causing delays and increasing costs.

Work on the fabrication and construction of the prototype continued apace throughout 1998-9 with most of the issues with the TPS and aerospike engines being ironed out. Lockheed Martin and Rocketdyne agreed to absorb some of the additional costs arising from developmental issues, but it was by now becoming apparent that there would be no first flight before at least July 2000. Elsewhere, reorganisations within NASA’s Office of Aeronautics and Space Transportation Technology saw Project Head and strong SSTO advocate Gary Payton moved to a more limited role. This set off ripples of concern within Congress and Dan Goldin had to step in to re-assure the NASA Authorisation sub-committee that the RLV initiative remained on track.

The completed X-33 launch site at Haystack Butte, Edwards AFB [Img: NASA]
The completed X-33 launch site at Haystack Butte, Edwards AFB [Img: NASA]
Tank troubles
With most of the key technological advances progressing and the launch site now completed it had seemed that maybe the X-33 had got over it’s early teething troubles, but 1999 saw a series of failures with the carbon composite liquid hydrogen tanks that looked like bringing the project to it’s knees. In January and more seriously November 1999 these complex tanks failed during testing with de-lamination occurring in the composite structures under representative flight conditions.

It began to look like this may be one revolutionary technology too far for the X-33, but engineers at the Skunk Works having foreseen this eventuality proposed switching to an Aluminium Lithium construction for these tanks. Although the tank walls would be heavier, the nature of the joins was less complex and the overall weight was actually lower than that of the composite tanks. Some re-engineering would be required as the TPS joined directly to the tank structure but Lockheed Martin’s recovery plan was accepted and a Red Team review of the programme concluded in February 2000 that there were “ obvious showstoppers” to the successful completion of the X-33. But again, these problems had caused significant slippage of timescale and budget.

By September 2000 construction on the prototype was nearing completion. The majority of the TPS had been installed and two aerospike engines were being tested side-by-side at NASA’s Stennis centre with encouraging results. But the delays were beginning to take their toll with a first flight now predicted for 2003. Although NASA remained outwardly upbeat about the X-33 and subsequently the Venture Star’s prospects it now became clear that any additional funding from the agency would need to come from the Space Launch Initiative (SLI) programme.

As late as December 2000 both NASA and Lockheed Martin were confident this would happen, but then in March 2001 the X-33 received a hammer-blow as the SLI contracts were announced. Neither the X-33 nor the X-34 had been included. NASA would provide no additional project funding beyond the existing development agreement. They did, however, leave the door open for Lockheed Martin and it’s partners to continue on with the X-33 towards an operational VentureStar but given the amount of money already invested and, perhaps, Lockheed Martin’s existing interests in the expendable booster market they decided not to pick up the option.

Some attempts were made by the U.S. Air Force to resurrect some or all of the X-33 programme in the following years, but as far as we can tell, these too came to nothing.

But the writing had been on the wall for the X-33 for some time prior to the SLI decision. There were already voices expressing a deeper concern for the way the X-33 programme, and particularly risk assessment, had been run by NASA. In his April 2000 testimony before the House Science Committee’s Sub-committee on Space and Aeronautics, Industry expert and former NASA man Ivan Bekey stated “It was recognised that the X-33 would be a high risk program, and it was designed that way. We must place the state of the X-33 program into this context of high risk-high payoff experimental flight programs: failures and major setbacks are to be expected…” continuing “…the failure of the X-33 composite fuel tank and other components should have been anticipated by well-funded parallel component developments. This was not done for budgetary reasons, and though the about $1.5 B investment in the X-33 cannot be called puny, it nonetheless was single-string, with its attendant risks that led to the result we now face.”

Noting NASA and Lockheed Martin’s intention to fly the X-33 with the aluminium lithium tank as opposed to the original composite design, he stated this would be a mistake, feeling that all key technologies must be tested together “…If that is not done, the principal reason for the flight program disappears”.

This testimony, together with the programmes requirement for additional funding from SLI to proceed to completion surely put the final nails in the X-33 coffin.

So what went wrong?
The X-33 cost a LOT of money. At the time of cancellation it is reckoned that NASA spent some $922 million, with Lockheed Martin and partners contributing a further $357 million. While outwardly being part of Dan Goldin’s ethos of “Cheaper, Faster, Better” it was failing on the first two of those and the ‘better’ was becoming debatable to many within NASA. The old engineering adage “Better is the enemy of good” so beloved of senior Apollo men like Joe Shea seemed to have been cast aside in an attempt to do too much in a new way and get it right first time, on time and on budget.

In large part, the X-33 was as much a failure in project management as it ever was a technical failure. Original Phase II budgets and timescales were extremely optimistic given the ground that needed to be covered to get the system operating as envisioned. As Bekey noted, it was a ‘high-risk’ programme, but it wasn’t necessarily sold as such from the outset.

Possibly one of the other Phase I designs would have offered a better prospect for success. In picking the most challenging design NASA must have estimated the technical pay-off would have been worth the risk but the potential for overruns was always apparent.

Some have pointed to the association the X-33 had to the Clinton Administration, particularly to Al Gore and wondered if maybe the cancellation in 2001 was politically motivated. This is possible but difficult to prove: while there was no obvious ‘better’ successor to the X-33/VentureStar system, the problems had been apparent for some time.

It has also been said that by 2000, the design for the VentureStar had diverged sufficiently from that of the X-33 that the flight test programme would have had a limited input into the full-scale vehicle, but in counter to this argument many of the systems would have remained broadly similar and still needed proving before VentureStar could begin development.

From a distance of 14 years, in a post-Shuttle world where NASA is still working to develop the commercial ecosystem to support it’s activities in LEO, it seems a tremendous shame that the X-33 wasn’t given the chance to prove some of it’s groundbreaking technologies.

Although weight gains would have impacted on it’s top speed and altitude it could still have conducted useful work in proving the Linear Aerospike engines and metallic TPS. While the use of composites for cryogenic tanks proved too challenging at the time, big strides have been made since and the prototype could potentially have been upgraded to prove these and other technologies in hypersonic flight as they became available – this type of retro-fitting had been carried out on the McDonnell Douglas DC-X.

It’s also worth noting that for the second half of it’s celebrated career, the X-15 performed as a platform to carry experiments and test materials into challenging flight regimes. Given the current push towards new hypersonic weapons systems, a research vehicle capable of routinely reaching hypersonic velocities may have proved more valuable than was imagined in 2001. With DARPA now involved in designing the XS-1 two stage to orbit system with a spaceplane first stage, could the X-33 not have performed this role, or at least provided a head-start?

Engineers who worked on the X-33 have since expressed both disappointment that it never flew and a confidence that the design as-built was sound and could have performed it’s mission. Perhaps if it had been allowed to continue in the role more traditionally associated with X Planes – for a structured research programme pushing back boundaries of flight and increasing aerospace knowledge, it would have paid back some of the investment NASA and Lockheed Martin ploughed into it.

As it was a 90% complete Flight Article and mothballed launch facilities ended up being of little use to anyone.

An early rendering of the VentureStar releasing a payload on orbit [Img: NASA]
An early rendering of the VentureStar releasing a payload on orbit [Img: NASA]