Rockwell Flyer: The Story of the X-33 and Beyond - Timeline in a Post

Rockwell Flyer: The Story of the X-33 and Beyond

American Rocket News

January 2, 2006

Written by Arnold Holmes

With the start of Congressional hearings on the factors relating to the decisions involved in the Exploration Systems Architecture, we here at American Rocket News feel it is worth revisiting the program which has spawned such debate as to the future of NASA's human spaceflight programs: the Rockwell X-33 Sprite.

The first steps toward what would become the X-33 occurred when the White House, in 1994 selected Option 3B from the NASA Access to Space Study. This decision--to publicly assist with developing the technologies and systems necessary to reduce the private risks in developing reusable space vehicles--lead directly into the X-33 Competition.

At the start of the competition, there had been three leading options. The first, from McDonnell-Douglas PhantomWorks, was a conical DC-X derived design using vertical take-off and landing on a single Space Shuttle Main Engine (SSME). The second, from Lockheed-Martin Skunkworks, was a flattened triangular lifting body design using a linear aerospike, visually similar to their LS-200 alternative to the final Space Shuttle design from the 1970s. The third was Rockwell’s, a vertical-takeoff, horizontal landing design using a winged cylinder with integral cylindrical tanks with an immediate visual resemblance to their winning Space Shuttle concept. Rockwell’s design was selected among the three, with NASA’s initial announcement on July 2nd, 1996 stressing the importance of a low-risk base design to serve as the testbed for other technologies later down the road. NASA’s internal evaluations, revealed in documentation exclusively made available to American Rocket News’ Stage 2, reveals that this drive for a vehicle which could fly first and foremost and test additional technologies second was key to Rockwell’s victory. NASA determined that Lockheed’s multi-lobe composite propellant tanks, with their numerous complex shapes and joints, might be extremely problematic to fabricate, while the air-start of the SSME for landing on the McDonnell Douglas design would require a successful ignition of an engine, which could be temperamental even when lit on the ground, within a margin of seconds on every single flight or the loss of vehicle could result.

Compared to these risks, Rockwell’s domed cylindrical tanks and simplified propulsion arrangements--a single burn from the RS-25 Space Shuttle Main Engine and two RL10-A-5’s from the DC-X/XA program, as opposed to McDonnell Douglas’s re-lighting engine or Lockheed’s all-new linear aerospike--proved to be the right level of conservative development for the X-33 program. Furthermore, while Rockwell’s design could test an advanced Thermal Protection System (TPS) such as Lockheed’s metallic tiles, it did not need such a system to work, and relied on high density ceramic tiles in hot and high impact areas, and shuttle- derived thermal blankets for most of the vehicle. While such advanced technologies (along with others like supercooled or even slush propellants) could be retrofitted onto the vehicle for their own demonstrations after initial proving flights, they would not be critical to have ready from day one--a major reduction in the overall cost and risk of the program.

Progress on the X-33 program continued at a steady, but reasonable case through the end of 1999, when the first major component test of the vehicle took place. The large forward hydrogen tank, made out of carbon fiber composites, had to be pressure tested prior to vehicle integration. If the tank had failed, there were forces both inside and outside of NASA that would take it as a sign that the program was doomed to fail, and would push for cancellation. On Tuesday, November 2nd, their fears were allayed, and the tank passed a full-up pressure test that took it well above flight pressures. It was then transported from the test stand at Marshall Space Flight Center, Alabama, to US Air Force Plant 42 in Palmdale, California - the same facility where the Space Shuttles were built in the decades prior.

While the test vehicle was being built Rocketdyne, another division of Rockwell, was assembling a modification of the SSME (Space Shuttle Main Engine) intended for use on the X-33. This version of the SSME (Block IIA RS-25C) was supplemented by two Pratt & Whitney RL10-A5 engines. These small expander cycle engines were of the same design as the engines used on the McDonnell Douglas DC-X test vehicle, and had an exceptionally low expansion ratio - only 4:1 - that allowed them to operate at at sea level to provide a minor power boost (each engine would provide only about 7% of the total lift-of thrust of the vehicle), and a high degree of control. The only change to all three engines relative to their original vehicle applications was that the RS-25C would not have any gimbal actuation, and the RL10s would have the electro-mechanical gimbal actuators from the Atlas IIIA’s RL10-A4-1s engines. These actuators, combined with similar actuators on all flight surfaces, eliminated the need for hydraulics and an Auxiliary Power Unit - a system whose hydrazine powered implementation had proven to be problematic on the Space Shuttle.

Furthermore, Rocketdyne was in discussions with Energomash NPO to bring to the US the RD-704 engines just as Pratt & Whitney had worked with Energomash to bring the RD-180 for the Atlas III and V families, and Aerojet’s deal with the Kuznetsov Design Bureau to bring the Nk-33 and Nk-43 engines to the US, and specifically to Kistler to use on their K-1 vehicle under the designations AJ26-58, 59, and 62. Unlike the oxidizer rich staged combustion engines that P&W and Aerojet were bringing to the US, Rocketdyne was attempting to bring in something even more exotic, for the RD-704 was a tri-propellant engine, burning oxygen with both kerosene and hydrogen over the flight profile. For initial burns early in the accent, the RD-704 burns the oxygen with both kerosene and hydrogen at the same time to offer a high-thrust and high-mass flow mode. Once sufficient altitude is reached however, the kerosene feed is cut, and the engine becomes a conventional hydrogen-oxygen (or hydrolox) engine for maximum specific impulse or efficiency. By using kerosene to assist with the launch thrust, the tank volume and thus mass can be minimized relative to a pure hydrogen-oxygen vehicle. During this program to bring the engines to the US, the engine was assigned the designation RS-74 for engines built in Russia and imported to the US, and RS-2100 for engines produced in the US under license from Energomash. Rockwell’s planned configuration was to use an array of six RS-2100 engines on the tail of the full size Reusable Launch Vehicle (RLV) to provide an engine-out capability without the loss of the vehicle or payload.

Such plans would have to wait for the demonstrator to fly, however. Final vehicle assembly of the X-33A took place in the first half of 2000, with the vehicle proceeding on its official roll-out to much fanfare and celebration. The name of the vehicle, Kitty Hawk, was announced by long time NASA Administrator Dan Goldin at the roll-out on Thursday July 18th, 2000. The official reason the name was picked was “To honor the commitment of those early aviators, and their drive to make the heavens a reality.” That the name was sufficiently patriotic while being (mostly) non-political was certainly seen as a plus. It is of note, that while Kitty Hawk was being rolled out for the first time, her predecessor OV-102 Columbia was nearby, in Building 150 at Palmdale, receiving what was to be her final Orbiter Maintenance Down Period - OMDP-J3. Rockwell’s need to both prepare Kitty Hawk for tests while processing Columbia lead to issues in the middle of 2000 when wiring inspections of Columbia discovered over 4,500 discrepancies. In the end, Rockwell was forced to bring onboard even more staff than they had planned - over 500 on the shuttle, and several hundred on the X-33A.

While the X-33 program was proceeding as scheduled and performed a test series of captive carry tests followed by drop-tests from NASA’s second Shuttle Carrier Aircraft (SCA) N911NA over the remainder of 2000, the uncertainty that swirled around the American political system continued around the program long after the election was resolved. With a new administration, and with the strong possibility of a new direction in space development, the X-33 program felt under pressure to deliver progress. Fortunately, the program was on budget, and while not running around ahead of schedule, it was not behind either. In an effort to show results, the X-33 Kitty Hawk fired its engines for the first time on Tuesday, March 13th, 2001. Though short (just 20 seconds), the burn had been accelerated by nearly two months from the original schedule, thanks in no small part to some heroic work from engineers and technicians at the Haystack Butte Launch Facility at Edwards Air Force Base. With engines taken from existing programs, everything went very smoothly, and the facility stood down to prepare for the first launch later that year.

While the preparations for the first flight were not as rushed as those for the test firing (as the possible political pressure to cancel the program as a relic of the prior administration had abated), they did proceed apace, and the Kitty Hawk made its first powered flight on Tuesday, September 4th, 2001. This first flight would be a relatively short hop for the program, barely going above mach 2 and only traveling down range about 100 miles from Haystack Butte Launch Facility Edwards AFB to the Silurian Dry Lake Bed just north of Baker California. The vehicle’s first two flights in 2001 to Silurian were the only two flights of the program that saw the vehicle transported back to Edwards completely via ground transport. Based on the evaluation of these initial flights, a third flight to the dry lake (though at mach 3 peak velocity) was deemed unnecessary. Instead, the flight was omitted, saving the operational costs, hopefully to be replaced in the schedule with a later flight to a landing site further down-range, with the exact location to be driven by the needs of the program. With the basic flight testing complete, the program’s funded 15 flights would be better saved for the technology demonstration goals of the program.

As 2001 rolled into 2002, the Rockwell team began to set up processing equipment at their next landing target - Michael Army Airfield at Dugway Proving Ground, Utah. The Michael facility, located south and west of Salt Lake City, has a relatively long (nearly fourteen thousand foot) runway, surrounded by flat terrain, which was perfect for landing the X-33, and its range from Edwards was suited for landings after flights into the hypersonic regime. Prior to the first flight to Michael, a certain amount of equipment--including lift cranes and one of the two shuttle carrier aircraft--was required to be pre-positioned to support operations such as safing the vehicle, and loading it on the SCA for transport back to Edwards. The first of these preparations started in January for a planned February launch. An attempted launch on Tuesday the 12th was held due to range issues, and the flight was delayed for 24 hours, when the vehicle performed as expected, and flew downrage to Utah. Following the flight, the same “Away” team that had performed the mate-demate operations for the captive carry and drop tests went to work, and were able to mate Kitty Hawk to N911NA in under 12 hours. Over next year, the “Home” mate-demate team at Edwards would get this down to a seven hour process, though the ‘Away’ team that was deployed to Michael, and eventually Malmstrom, would on average take an hour longer. Over the following six months, Kitty Hawk performed a grand total of 7 flights to Michael, and flights Seven, Eight, and Nine took place on July 30th, August 6, and August 13th, demonstrating the back-to-back week-long turn-around goal for the program.

The final phase of the official X-33 test program was to validate the full vehicle reentry capabilities. The required velocity and altitude to create this aerothermal load would require a speed up to 30% higher than the mach 12 tested on the Edwards to Michael flights, and thus longer ranges than the nearly 400 nautical miles (over 700 kilometers) from Edwards to Michael AAF. The landing site for the last batch of tests was over twice as far down-range - Malmstrom Air Force Base, Great Falls, Montana. Malmstrom had seen the last permanent fixed-wing aviation detachment leave in 1997, but retained an eleven thousand foot runway that, after just five years, was easy to place back into active service for both the X-33 and SCA. During the landing site selection process, Malmstrom scored well despite the inactive runway because of the large tarmac areas available for parking the SCA and for mate-demate equipment and limited impact of flight tests on local military aviation. While the test did require the reconstruction of a control tower, the new facility was paid for out of the US Air Force Budget as a way of configuring the base for future use.

Site preparations at Malmstrom took slightly longer than anticipated however, and the first of five planned flights lifted off on Wednesday, October 9th. The flight, like all others before, was smooth until landing when the only major issue of the test program occured. During the vehicle’s glide into Malmstrom, one of the rudder actuators developed an issue and shut itself down. To retain the require margin for control at low altitude (under ten thousand feet), the vehicle’s automatic flight systems switched to the designated systems for that phase of flight, and relighted both of the RL10 engines to ensure proper yaw control would be maintained. The use of the engines in such a way was one of the main reasons why the low expansion ratio engines had been selected for the program originally. This event also represented the first time that a spaceplane had performed a powered landing (as the RL10s were at thirty percent throttle at touchdown).

After the vehicle was safed and returned to Edward, engineers from Rockwell and NASA did a teardown of the vehicle’s rudder systems, and discovered that the issue was not with the actuator, but the actuator health systems. When the health sensor on the actuator failed, the vehicle flight systems shut down the actuator itself, to avoid damage to the vehicle. While the Rockwell and NASA teams were relieved that nothing was physically wrong with the vehicle, the time taken to diagnose the problem ate into the calendar year, and with limited time available in the year a decision was made to try for the second turn-around target for the vehicle - the ability to launch, land, recover, process, and relaunch the vehicle inside of two days. The target launch dates for flights 11 and 12 would be the 10th and 12th of December. The weather cooperated, and the first launch proceeded smoothly. The weather in Malmstrom that morning was even warm - nearly fifteen degrees fahrenheit (9 degrees Celsius) warmer than average [0]. This allowed the ‘Away’ team at Malmstrom to complete their work quickly. Once Kitty Hawk was mated to the N911NA, the team, for only the second time in the history of the program, stayed deployed (the previous time was the campaign for flights Seven, Eight, and Nine). Flight operations continued, and by the late evening of the 10th Kitty Hawk was being towed back to her check-out hanger at Haystack Butte. The next day saw an expedited processing flow that fortunately encountered no unexpected issues, and by midnight of the 11th/12th, the Kitty Hawk was back on her launch pad. Finally, at 8:45 Pacific Standard Time, she took to the skies again - demonstrating the ability of the teams and the vehicle to fly with a 48 hour turn around. Actual time from launch to launch was a nail-biting 47 hours and 50 minutes. Time lapse video of the amazing processing flow that the Rockwell engineers conducted that week is available exclusively on American Rocket News’ Stage 2.

Having completed the rapid turn-around tests (and proving that when properly designed, rapid reuse could be possible), the engineering teams stood down for the holiday season in preparation of the launch campaign in the new year. While only two launches were officially programmed, there was hope that the original third flight to Silurian, which had been officially ‘indefinitely delayed’ would be re-introduced as a sixth flight to Malmstrom. While the first thirteen flights had been used to demonstrate the vehicle’s basic flight and operational characteristics, the remaining two or three flights would be used to test the Kitty Hawk’s capabilities as a platform for scientific and future vehicle research. The first of these tests would be to expose the vehicle to a hotter and more energetic environment by performing a pair of ‘depressed trajectory’ flights. While the prior tests had pushed the vehicle to mach 15 and over 75 kilometers (about 250 thousand feet), the last two tests would be much lower - only fifty to sixty kilometers while maintaining such a high speed. By reaching such high speeds in the thicker parts of the atmosphere, the compression heating--the heating caused by the vehicle pushing the air out of its way, the primary source of entry heating for a returning spacecraft--would be increased to a level more consistent with that of a vehicle returning from orbit. These thermal and aerodynamic tests could be conducted much easier than other proposed tests such as replacing Kitty Hawk’s main engine with a proposed aerospike.

While flights 13 and 14 were not planned to engage in the high speed turn arounds of flights 11 and 12, or even the rapid turn arounds of 7-8-9, they were to be quick - the ideal would be to get both done by the end of January. Flight 11 was launched the morning of January 14th, and subjected Kitty Hawk to the highest thermal loads she had yet to experience. Kitty Hawk was ferried back to Edwards, and the teams enjoyed the three-day Martin Luther King Jr. Day Weekend. The next week involved the most thorough inspection of the vehicle’s Thermal Protection System yet, and a validation for flight that next week. By this point, the current program budget would not support a 15th flight, so this 14th would be the last one of the current flight program. On January 28th, Kitty Hawk launched, and as usual, performed her mission beyond specifications. This 14th flight was unique in that the vehicle’s outer mold line was changed with the addition of two aerodynamic fairings, one on the side of the fuselage, and a second on the belly of the vehicle in a lower heat region, that contained a sample of Lockheed metallic shingles. The shingles, being placed over the existing TPS, were not critical for the vehicle, but the flight represented their first test in true reentry conditions. Kitty Hawk was carried back to Edwards the following day due to a minor issue when being mated to the SCA, which delayed the flight back until after sunrise. With the future of the program uncertain, a number of Rockwell engineers and technicians took a trip north to observe a rare over-land entry of the space shuttle. OV-102 Columbia was returning from her science mission (STS-107) and would pass over central California and Nevada sheathed in plasma as she continued on her way to the Cape. Columbia would never make it.

While the loss of Columbia was certainly the most visible bit of news to involve space in January of 2003, it was not the only bit of news that would have long term impacts. As the X-33 tests were winding down, the program had proved that if designed for, reusability was possible with vehicles designed to fly at speeds and in the aerothermal environments needed for spaceflight. The first major beneficiary of this was not Rockwell, but a smaller company, and one of the last survivors of the 1990s ‘New Space’ movement, Kistler. While Kistler had reached 75% complete by mass on their first vehicle by 2000, development had slowed, and not even the Chairman of the company, the legendary George Mueller of Apollo Program fame, was able to keep bringing in investments in 2001. The flights of the X-33 however did attract the attention of one group that had not been approached - the Ontario Teachers Pension Plan. The investment of over $200 million (US) into Kistler was enough to bring in extra investment from other sources, and allow Kistler to stave off what was seen as an impending bankruptcy. Furthermore, the investors wanted to see some payoff, and a target of the end of 2005 was set for a first flight. Similar to Rockwell’s X-33, the Kistler K-1 did not require the development of an all new engine, simply the certification of the Russian (actually ex-Soviet, as they had been built in the 1970s) Kuztensov Nk-33/43 oxidizer rich staged combustion engines. These engines are unlike any built in the west, offering a specific impulse fifteen percent higher than any American engines that use the same kerosene and liquid oxygen propellant combination. Three of the these engines in the vehicles first stage would ignite on the pad, and lift the stack to a little over 175 thousand feet (~50 kilometers) just 140 seconds into flight. At this point, the upper, orbital vehicle stage would separate from the first ‘Launch Assist Platform’ Stage, and boost the payload to earth orbit atop a single Americanized NK-43, designated the Aerojet AJ26-60. Following separation, the Launch Assist Platform or LAP would perform a backflip, and would relight the center engine to change the trajectory so that it would come down near the launch site. This maneuver would be called the ‘Lob Retro’ return, and simplified transport of the first stage as it required neither any additional flying vehicles to return the LAP over long land distances nor airbreathing engines on the LAP. Once the Lob-Retro trajectory was achieved, the vehicle would climb to an altitude of over 300 thousand feet (~90 kilometers) and would descend to the landing area on parachutes, followed by a touchdown on airbags. Once landed, the LAP would be loaded by crane onto a trailer, and transported back to the processing facility where it would be checked out for reflight on a two to three week time period.

The Orbital Vehicle (OV), with an integral forward cargo bay would have continued on to orbit, eventually shutting down the main engine, and circularizing on a pair of Alcohol-Liquid Oxygen orbital maneuvering engines. Once in orbit, the blunt nose of the orbital vehicle would swing open, allowing the payload contained therein to be deployed into space. Once payload was either deployed, retrieved, or both, the nose would swing back into place, and the vehicle would wait until the orbit passed back over the launch facility (this would usually take about twenty-four hours), before using the same engines used to circularize to de-orbit. As the vehicle entered, the forward nose would be the primary heat shield, and the rest of the vehicle would be protected by ceramic thermal blankets similar to those used on the Space Shuttle and the X-33. Once the vehicle was low enough, parachutes would be deployed and the OV would land on airbags under parachutes, similar to the LAP. Processing flows for the OV were similar to the LAP, though more involved as the vehicle would be subjected to greater stresses and loads. After three to four weeks, the vehicle would be processed, and ready to be loaded and relauched.

The primary launch facilities for Kistler were planned to be the Woomera Test Range in Australia which would see the first series of launches. A second launch site in the Nevada Test Range is planned for follow-up development. Both sites would offer launch to medium inclination orbits (between 45 and 60 degrees), and polar orbits (up to sun synchronous), but the US launch site would be useful for US payload that for various reasons were not allowed to fly on foreign rockets. One additional prize that Kistler was targeting was a NASA International Space Station Resupply contract that had been proposed several times, but never put out for bid. If Kistler could get flying, they would stand a very strong chance of getting the contract.

The loss of Columbia on February 1st sent a shock through Rockwell. NASA’s plans to keep the shuttle in service through 2020 were now up in the air, and with no further tests planned in the X-33 program, Rockwell was forced to start placing that program in a state of suspended animation. Staffing was planned to be cut to a bare minimum following and an extensive knowledge capture program was implemented to aid in restarting the project at some point in the future. As this knowledge capture program proceeded in early 2003, one idea started to be studied. During the development and testing of the X-33, Rockwell engineers continuing parallel development on the full scale RLV had come to the difficult to swallow conclusion that the X-33 could not be scaled up to be a full single stage to orbit vehicle, the result of a frustrating perfect storm of underperformance from RS-2100 engines on test stands and discovery that the planned full size reusable launch vehicle would be heavier than original studies indicated. However, it could be a part of the two-stage system, either the first or (more ambitiously) second stage. It was around this time that a team from Rockwell approached Orbital Sciences Corporation, maker of the Pegasus launch vehicle.

Pegasus is a three or four stage rocket that is dropped from an aircraft (either a NASA NB-52 or Orbital’s own L-1011), and is capable of delivering small payloads to a myriad of orbits. Launching a Pegasus rocket from the back of an X-33 would result in a higher payload as the X-33 could lift the Pegasus nearly 6 times higher than the ~40 thousand feet (12 kilometers), and nearly 20 times faster than the mach 0.8 that the L-1011 Stargazer can achieve. This improvement would allow a Rockwell-Orbital venture to start impacting the launch market in ways that the existing Pegasus rocket could not. The X-33/Pegasus system, capable of 2,500 kg to a low Earth orbit, offers comparable launch performance to the Athena II, Dnepr, and low-end versions of the Delta II rocket. For Rockwell this represented the best way to make use of a vehicle that would otherwise likely end up in a museum gathering dust. Discussions about this proposal between the two companies continued until October, when on Wednesday the 9th, Rockwell and Orbital made a joint press conference announcing their plans to launch a demonstration payload atop a modified wingless Pegasus XL rocket, lifted to near-space via an X-33.

This bold announcement caused another organization with interest in space launch to take interest. The US Air Force, dealing with the emerging Evolved Expendable Launch Vehicle (EELV) Scandal, in which Boeing had been caught with proprietary Lockheed pricing information, began seeking out alternative means of achieving a rapid response launch capability. The most visible support the USAF gave was the continued lease to Rockwell of the Haystack Butte Complex, and preferential commercial use landing rights at a pair of USAF bases in the western United States - Malmstrom in Montana, and Cannon in New Mexico. Cannon was selected for its due-east location relative to Edwards, which would be critical for Rockwell and Orbital if they were to launch payloads into relatively low inclination orbits. The Air Force, at least initially, got a major reason to keep the fixed-wing aviation capability at Cannon open in the face of possible closing under the next round of the Base Realignment And Closing Commission.

While receiving ground support, and considerable interest from USAF engineers, a much more visible vote of confidence came six months into the program to modify Kitty Hawk from an X-plane to a reusable lifter when in early February 2004, the Defence Advanced Research Projects Agency (DARPA) and the Space Test Program (STP) contracted Rockwell and Orbital to launch a demo-class payload into space following a successful demonstration launch of the X-33/Pegasus. It was at this point that the commercial X-33 program received a new name - Sprite. The name was considered an allusion to both the ethereal flying beings of myth, and the high altitude lightning events first photographed in 1989. The name was especially apropo as the sprite lightning occurs between 50 and 90 kilometers above the surface, which was the maximum flight altitude of the vehicle. The FalconSAT payload manifested was built by cadets at the USAF Academy at Colorado Springs, and represented a scientific payload that while useful, was not critical.

One minor note is that prior to the program being named ‘Sprite’, internal Rockwell documents shared exclusively with American Rocket News’ Stage 2 referred to the program during conceptual development as ‘Turbo-Encabulator’, an obvious reference to the late 1970s parody of technical presentations done by acclaimed voice over artist Bud Haggert.

While Rockwell and Orbital were working on Sprite, NASA and the White House were finalizing the “Vision for Space Exploration” (VSE) report. This report, released in January 2004, called for NASA to end Space Shuttle flights following a rapid completion of the International Space Station, fund commercial resupply and crew rotation flights to the station, and for NASA to focus on the “Moon, Mars, and Beyond.” For Rockwell, the official VSE position of ‘Replacing’ the Space Shuttle seemed promising.

Through the summer of 2004 as NASA and Rockwell moved forward with the changes to bring the shuttle back to flight status and NASA began working on what would eventually become the Exploration Systems Architecture Study, Rockwell and Orbital proceeded with the structural work needed to make the Kitty Hawk into a mothership for the modified Pegasus rockets. The modifications to the Pegasus rocket involved removing the wing and moving the structural supports from the top of the first stage to the bottom. The X-33 saw a large structural member inserted along the top of the vehicle’s payload bay that would provide support for the Pegasus in flight, as well as an aerodynamic blister to fair-in the rocket while the booster launched and rose to near-space altitudes.

While Rockwell and Orbital moved forward with both the first (a set of five small ‘cubesat’s, each about 10 centimeters on a side) and second (Taiwan’s ROCSAT-2 [1], which had moved from Orbital’s Taurus vehicle for a substantive reduction in launch costs), NASA was busy conducting early work into what would eventually become the Exploration Systems Architecture Study or ESAS. The impacts of this work would not be felt for almost a year, and on Tuesday, September 21st, the first launch of the Sprite system occurred. The launch proceeded without issues and the vehicle released the modified Pegasus rocket over the desert of eastern Arizona, at an altitude of 65 kilometers and a speed of mach 14 (slightly over four kilometers per second) before gliding the remaining thousand kilometers across Arizona and New Mexico to land at Cannon Air Force Base. While Rockwell had been very careful in press releases to describe Kitty Hawk as a ‘Near Space Vehicle’, the winged shape’s similarities to the Space Shuttle caused the headline the next day in the Clovis City Journal to read “Commercial Spaceplane Visits Cannon.” This headline, and not Rockwell’s insistently correct terminology, was picked up by media around the world as the flight was hailed as the world’s first commercial spaceplane launch. Teams at Rockwell were loath to rest on their laurels however, as there were plans for two more launches in 2004, the first of which would require the Sprite program to launch on a second, untested azimuth. The payload, ROCSAT-2, was being delivered to a polar, sun-synchronous orbit. While most polar launches out of California are directed south from Vandenberg Air Force Base so that they fly over the Pacific ocean, the X-33 needs a runway to land on after the Pegasus separation, and thus would be launched to the North-Northwest from Haystack Butte. For this, the Sprite program team revisited the list of possible landing sites from the early days of the X-33 program, and found a site in roughly the correct location, and at the correct distance downrange - Grant Lake International Airport, located 10 kilometers northwest of Lake Moses in central Washington state. While Grant Lake is classified by the Federal Aviation Administration as a General Aviation Airport, it does maintain a pair of runways not less than 3000 meters long, allowing for ease of operations when landing the Sprite vehicle.

That launch however, was still weeks in the future, and on Monday, October 4th, 2004 Rockwell having started to secure further launch contracts thanks to both apparent USAF confidence, and a first successful launch announced at a major press conference that they would begin the construction of a second vehicle, the X-33B. This craft would be slightly different from her earlier sister in that she would be structurally optimized from the beginning to operate as a launch platform for Pegasus. The second vehicle would be built using structural spares from Kitty Hawk where they existed (such as the main propellant tanks). Rockwell Space Systems Senior Vice President Robert Davidoff gave a speech outlining the plans for the Sprite Program, including a decision to begin the procurement of long-lead items for the construction of a third vehicle if the market demand materialized. As the press conference came to a close, it ended thusly:

“As the world’s only manufacturer of reusable spaceplanes, Rockwell remains committed to developing the systems needed to improve space access.”

"After they first took to the skies at Kitty Hawk, the Wright Brothers returned home to Dayton, where they worked for two more years at Huffman Prairie. There, they took the demonstration of short, simple flights flights and worked to make it a practical, repeatable, sustainable reality. Even today, a century later, engineers and pilots follow in their footsteps. With the second X-33, we at Rockwell strive to set a similar path to reliable, repeatable, sustainable exploration of space, and we dub this vehicle in honor of the place they made their advancements."

The name selected (Huffman Prairie) was an obvious nod to the US Air Force, whose Wright-Patterson Air Force Base surrounds the Huffman Prairie Flying Field Historic Site (which is run by the National Parks Service, and is open to visitors year-round). Other names brought up for consideration included Langley, Enterprise, JFK, and Doolittle.

As Rockwell began to achieve a steady launch cadence with three launches in 2004, and more planned for 2005, the first signs of issues at NASA headquarters were coming forward. With the announcement in late 2004 that Administrator Sean O’Keefe would be stepping down, hopes for the development of a two stage completely reusable launch vehicle waned. Administrator O’Keefe’s replacement, Michael Griffin was an academic who had previously been the Associate Administrator for Exploration and represented a radically different background than O’Keefe, who had been a Director of the Office of Management and Budget. While not visible on the surface, the effects of this change in leadership would come to the fore later.

As 2005 rolled on, NASA and it’s partners Rockwell and Lockheed (by way of their joint venture, United Space Alliance, which had taken over as the major program contractor for the Space Shuttle in the fall of 1996) were preparing for the Space Shuttle’s ‘Return To Flight’ mission following the Columbia disaster in February of 2003. The launch, on July 26th had been delayed by nearly two weeks from the scheduled date of July 13th due to unresolved fuel sensor issues with the external tank. These issues were left unresolved at the time of launch. It turned out that the fuel sensors issues, and the resulting fill and drain cycles of the external tank would help solve the foam shedding issue that was still apparent on STS-114. X-Ray tests done at the Michaud facility where the External Tanks are built discovered that the repeated fill and drain cycles put the sprayed-on foam under structural stress that caused certain layers to delaminate from each other, and to be shed under aerodynamic loads at launch. These facts would not be known until Late December however, and first the long awaited Exploration Systems Architecture Study Report Summary was released on the 17th of September. The findings of the report were a surprise to many in the aerospace industry. It called for the development of two new shuttle-derived vehicles to support both the International Space Station and future missions to the moon and Mars. The first vehicle, the ‘Ares 1’ will consist of a space shuttle solid rocket booster first stage and a space shuttle main engine powered upper stage that will be ignited after first stage burn-out. This rocket will place a crew capsule, derived from Apollo aero-ballistic research, into orbit to satisfy NASA crew flights to the ISS, and to place crew into orbit for missions to the Moon and Mars. The second vehicle, called ‘Ares V’ will use a core stage derived from the Space Shuttle’s External Tank, and be boosted by two newly developed five segment solid rocket boosters while the core is powered by five expendable RS-25 Space Shuttle Main Engines.

This architecture announcement stunned the aerospace industry who expected that either a derivation of the existing Evolved Expendable Launch Vehicles or a new reusable launch vehicle based on the work of the X-33 and Sprite programs would be selected. Rockwell and the newly forming United Launch Alliance figuratively retreated to their corners to evaluate the next steps they would take when, after a week of final vehicle check-out, when on Thursday, October 27th, 2005, at noon local time, under clear skies and light winds, the Kistler K-1 vehicle launched for the first time. The mood both at Woomera, and at Kistler headquarters in Seattle Washington (where it was 7 pm Wednesday) was ecstatic, as the work of over a decade finally came to fruition. The first stage steered itself down under both aerodynamic fins, a late addition to the vehicle’s design to improve landing accuracy, and parachutes before final touchdown on airbags, just a few hundred meters from the center of the kilometer wide landing zone. The second stage continued on to orbit where it opened the forward hatch exposing the demonstration payload for release. Once the satellite was separated from the K-1, the hatch closed, and the vehicle waited until it’s orbit brought it back over the recovery area at Woomera. The vehicle’s orbital maneuvering engines made a brief burn, and the craft re-oriented itself for a nose-first entry. Tension was high as the reentering craft was tracked on radar, right up to the point of airbag deployment and touchdown. As the upper stage was being recovered in Australia, at Kistler Headquarters, the company held a press conference announcing the recovery of the world’s first completely reusable orbital class rocket. When asked about the mood in the company by American Rocket News’ reporter Arnold Holmes, K-1 project director Jean-Pierre Boisvert [2] replied "I'd just like to say that this is a week we've been waiting for for a long time. I’m incredibly proud of my team, the vehicle performed better than we could have expected. For my part, I'd like to say it's a Friday night, and it's happy hour at the bar up the street in 15 minutes. For my team, I’d like to say that the first round is on me."

With two companies now demonstrating at least partial vehicle reuse, it should have come as no surprise that there were factions inside NASA that were opposed to the ESAS summary finding that the proper course of action for the agency was to fund the development of not one but two shuttle derived expendable launch vehicles over the next decade. It was this tension that caused unknown sources in NASA to leak the unreleased full text of the report to Rockwell. The full text of the report, subsequently leaked by Rockwell to both members of congress and the news media including American Rocket News, outlined that the ESAS committee considered reusability in rockets to be both speculative and lacking firm cost estimates, and thus reusable designs were penalized from the start. In statements to American Rocket News both Rockwell and Kistler confirmed that NASA made little to no inquiries as to operational parameters and cost structures of their vehicles. While Rockwell and Kistler’s leaks and statements were initially discounted by elements of NASA as complaints from contractors who didn’t win, they did resonate with people outside of NASA. With demonstrations of reusability being able to help drive costs down (a flight on a K-1 vehicle, which has a similar payload to a Delta II, comes at just a third of the price), it seems to the public, members of Congress, and the media that NASA imay be failing to take advantage of reusable systems to reduce the costs of missions to space, even ignoring lessons learned on NASA’s own development budget. A more in-depth review of the full ESAS report, which is available in its entirety on American Rocket News Stage 2, shows that on measures of safety and costs certain presumptions were made that do not stand up under closer scrutiny, such as presuming that the failure rates of multi-core heavy EELVs would match those of Titan IVs. After the report’s release, representatives from the EELV’s manufacturers, Boeing and Lockheed-Martin, protested that the EELVs, which lack large solids, are not likely to suffer the same failure modes as Titans, while the more directly Shuttle-heritage designs studied in the report were not similarly penalized for issues that occured in the STS program.[3]

The issues with the ESAS report, combined with the success of the X-33/Sprite and Kistler K-1 programs were enough to cause Congress to take notice, and announce the hearings that will start in the coming weeks. American Rocket News will be there to cover both House and Senate committee meetings in their entirety. While the content of the hearings is as of yet unknown, what is known is that the X-33 program has changed the dynamics of the future of space launch.


[0] That temperature is accurate, as are all of weather conditions for launch dates.

[1] Orbital Sciences doesn’t know how lucky they are here. ROCSAT-2 came in the only successful launch of a Taurus (now Minotaur-C) in the 2001-2011 timeframe.

[2] Last seen in Morning of the Maple Leaf as the backup commander who flew on the Arrow-5 mission to Save Skylab.

[3] This really happened.


I would like to thank @e of pi for both the idea and all of his support in helping me get past my occasional roadblocks, and general support.
So basically an earlier SpaceX, if I understand correctly?
No, not really. What happens is that NASA selects the low-risk Rockwell design for the X-33, which is then able to successfully complete testing and eventually fly, whereas in the real world Lockheed's design was selected and had testing failures that led to its being canceled before flight. Rockwell then leverages this to create a partially reusable LV based on X-33 serving as a carrier aircraft for Pegasus, while butterflies lead to Kistler, which was an actual launch vehicle company in the 1990s and 2000s, getting enough investment to put a vehicle together and launch. Having two reusable or partially reusable LVs mucking around then messes with ESAS when that comes out in the mid-2000s after Columbia.

EDIT: The thing is that Rockwell is Old Aerospace, of course, and Kistler had a rather different corporate culture than SpaceX, being much more Old Aerospace as well despite actually being a startup (they brought in a ton of old NASA hands to run things). So while they're demonstrating the reusability is really possible, better than Shuttle, they're not like SpaceX in how they do things at all. Much more staid, buttoned-down. No Elon, no grand Mars plans, nothing like that. So the effects are likely to be rather different in the long run.
Last edited:
First of all, thank you everyone for your interest, likes, and posts.

So basically an earlier SpaceX, if I understand correctly?

@Workable Goblin covers it well, but I would like to add that in TTL, it is actually rather likely that SpaceX doesn't make it. The Falcon 1 has a payload capacity less than Sprite/Pegasus, and while they do get to look at what Kistler is doing for reuse as signposts for what works, I don't see any reason why the first three flights don't continue to fail as they did OTL. Furthermore with changes at the start of 2006, the entire COTS program is going to be substantively butterflied, and it was the COTS investments and support from NASA that kept SpaceX alive until they got the Falcon 9 up.

To further expound upon the second part of Workable Goblin's post, the CEO and chief vehicle designer of Kistler was George Mueller, who was the man behind the all-up testing of the Saturn V rocket. The first version of Kistler (before they historically went bankrupt) was very much a newer Old Aerospace company.
@Workable Goblin covers it well, but I would like to add that in TTL, it is actually rather likely that SpaceX doesn't make it. The Falcon 1 has a payload capacity less than Sprite/Pegasus, and while they do get to look at what Kistler is doing for reuse as signposts for what works, I don't see any reason why the first three flights don't continue to fail as they did OTL. Furthermore with changes at the start of 2006, the entire COTS program is going to be substantively butterflied, and it was the COTS investments and support from NASA that kept SpaceX alive until they got the Falcon 9 up.
SpaceX's chances do look grim, but they might have one saving grace. An interesting observation is that Pegasus is rather low-performance and high-cost as a drop-stage for X-33, and its fairing is small for the capacity of the Sprite system, being originally designed for ~200 kg to LEO.

SpaceX might have a shot if they offer Falcon 1 as an alternative upper stage set for Sprite--it'd be lower cost (about half the cost of Pegasus solids, though some of the expense there is the low build rate which might be mitigated by higher production to support Sprite) and higher performance (about 3.3 to 4.2 tons, depending on how much of Falcon 1 you have to cut off to stay within acceptable liftoff T/W for the Sprite system). The larger core also comes with a larger fairing: though it's probably possible to put a 2 or maybe even 2.5m fairing on Pegasus if you're incorporating it to an aerodynamic fairing on the back of Sprite, the default fairing is only ~1m, while Falcon 1 is already in the 1.5m range range. Rockwell would just need to take a chance on changing subcontractors and "selling out" Orbital. Whether the first of the drop stages is made reusable is somewhat more questionable--though flying overland with X-33 means there's more ability to land downrange, burnout delta-v for the second stage means it may land >2000 km from the launch site. It might take something like an RTLS-level retroburn just to land at the sites the X-33 glides to, and parachutes like SpaceX was initially proposing for stage recovery likely don't cut it. The cost savings would have to be carefully evaluated, and they'd have to sell Rockwell on the risks--as a subcontractor, you don't have the freedom they've had historically with Falcon 9 being all their own vehicle.

Even if Rockwell doesn't go with SpaceX, some other kind of moderate-performance 25 metric ton upper stage stack is a likely performance improvement/cost reduction option for Sprite if Rockwell has a serious commitment to that size of payload in the future.
Last edited:
I am vindicated!

As more information comes out on the DARPA XS-1 / Boeing "Phantom Express", and the use of the AR-22 engine, it becomes apparent that the vehicle, flight, and cost models used for this TLIAP were more or less accurate to what is planned, 20 years later, by the successor to Rockwell.

To that end, I'd like to point people to This Nasa Space Flight article on the AR-22 tests. Rocketdyne, using flight-proven SSME hardware was able to run 10 tests in rapid succession, with an average turn around time on the engine of 18 and a half hours.