On 14 March, a lone Lockheed Martin F-22 Raptor took to the skies over Georgia on a test flight. It was the last of 187 aircraft ordered by the US Air Force.
The USAF had originally wanted 750 of the stealthy fifth-generation fighters, but a political debate over the very nature of aerial warfare in the 21st century ultimately sealed its fate. The debate started in the halls of the Pentagon in the wake of the US invasion of Iraq and Afghanistan around 2004 and came to a head in 2008 when then-Defense Secretary Robert Gates fired the USAF's top leaders - secretary Michael Wynne and chief of staff Gen Michael Moseley.
Marvin Sambur, who was the USAF's acquisitions chief during much of Donald Rumsfeld's tenure as secretary of defense, says the Department of Defense leadership was convinced that air combat was a relic of the past. The F-22 was "overkill" in an era when most wars would be fought against small groups of guerrillas, and unmanned aircraft would be the wave of the future.
© Lockheed Martin
The last of 187 Lockheed Martin F-22s ordered by the US Air Force took to the skies over Georgia on 14 March
The other aspect that influenced Rumsfeld's thinking was a desire for an aircraft that could fulfil multiple missions for more than one service. The USAF had learned the hard way - through the 1960s and 1970s, over the skies of Vietnam - that single-mission aircraft perform superbly. Rumsfeld's bias helped push the DoD to favour Lockheed's tri-service F-35 Joint Strike Fighter (JSF) over the Raptor.
The USAF position was that the Raptor, with its stealth, sensors and blistering kinematic performance, could be used to defeat integrated air defence systems and fly reconnaissance missions in addition to clearing the skies.
Indeed, the Raptor can cruise without afterburners at speeds of about Mach 1.8 and has a top speed of around M2.2.
"It is the world's greatest air dominance fighter," says Col Kevin Huyck, who was one of the eight original initial operational test and evaluation pilots to put the Raptor through its paces from March 2003.
"The first thing I noticed was the power and acceleration on my first take-off at Tyndall air force base," says Col Dirk Smith, commander of the 3rd Wing in Alaska and one of the first operational commanders of a Raptor squadron. "It seemed like the jet would not stop accelerating and it wanted to climb about three times as fast as the [Boeing] F-15C. Later, on subsequent training flights, the ability to sustain supersonic flight and manoeuvre above 50,000ft [15,200m] was a very unnatural feeling. I'd never seen that manoeuvrability at such a high altitude."
Nonetheless, Rumsfeld believed that the F-22 was overkill because Russia and China were far behind in developing similar technology.
"The feeling was: why do we need to spend more money?" Sambur says. "The view from the third floor became: we just need to be more advanced than them; we don't need to be so far advanced."
The air force disagreed, which eventually led Gates, who by that time had succeeded Rumsfeld, to fire the service's two top leaders in 2008.
Although a fleet of 187 stealthy fifth-generation air dominance fighters is a formidable force, the current arsenal is not what the USAF originally envisioned when it first conceived of the Advanced Tactical Fighter (ATF) programme in the early 1980s.
Over the course of two-and-a-half decades, the programme was perpetually stretched out and truncated numerous times in an all-too-familiar pattern that is now repeating itself for the JSF.
The USAF first recognised that it would eventually have to replace its McDonnell Douglas - now Boeing - F-15 Eagle air superiority fighter in November 1981. While Congress denied the service funding to start developing a new jet, it did direct the USAF to continue to study the problem.
In 1983, Congress granted the air force $23 million to award concept definition study contracts to seven airframers and two engine manufacturers.
© Lockheed Martin
Lockheed Martin's original designs for the US Air Force's "ATF"competition
Boeing, General Dynamics, Grumman, Lockheed, McDonnell Douglas, Northrop and Rockwell each received contracts worth around $900,000 in September 1983, according to a Teal Group report. Meanwhile, engine makers General Electric and Pratt & Whitney were well on their way to developing powerplants for the nascent fighter programme.
Under the commission's guidelines, instead of evaluating the companies' proposals over six months and selecting up to four contractors for a 36-month demonstration/validation phase, the USAF would select only two contractor teams.
The restructuring forced the would-be bidders to pool their resources. Two rival consortia were formed, with one team consisting of General Dynamics, Lockheed and Boeing, and the other being formed by McDonnell Douglas and Northrop. The remaining contractors, Grumman and Rockwell, chose to go it alone.
By 31 October the USAF had chosen the two consortia to design and build prototypes under 50-month, $691 million contracts. The two designs would eventually evolve into Lockheed's YF-22 and Northrop's YF-23. In addition to designing and building prototypes, the teams had to design and test subsystems. A pair of General Electric YF120 or Pratt & Whitney YF119 engines would power one of two prototypes built by each team.
As if the ATF programme was not complicated enough, the US Navy also wanted a variant of each design to fill its need to replace the Grumman F-14 Tomcat carrier-based interceptor, with a projected buy of about 618 aircraft.
But the teams persevered, and in June 1990 Northrop unveiled the YF-23, with Lockheed's YF-22 following suit in August. Both aircraft flew later that year during a competitive fly-off.
The YF-23 was faster and could fly an entire mission at M1.4 without fuel-guzzling afterburners, says Barry Watts, who was on the Northrop team before becoming an analyst at the Center for Strategic and Budgetary Assessments. It was also stealthier than Lockheed's entrant.
However, the YF-22 was far more agile with its thrust-vectoring nozzles and could more readily be adapted for the navy's needs, according to other sources. Watts compared the Lockheed jet to a "super F-15".
Both designs met the service's requirements for a stealthy, agile, supersonically cruising fighter with next-generation integrated avionics, but the YF-23 was the better performer.
"Aerodynamically, you could argue it was the better-performing airplane," says Watts. "You got a lot more super-cruise out of it than the -22."
Picking either jet would have resulted in a revolutionary leap in air combat capability, however.
Ultimately, on 23 April 1991, the USAF announced its selection of Lockheed's YF-22 to be its next-generation fighter. The design had "better capability and lower cost", the service said as the aircraft entered into engineering and manufacturing development.
© Lockheed Martin
Lockheed Martin's YF-22 (front) & Northop's YF-23 (back) went head to head in the ATF contest
Northrop was very disappointed. Watts says it is most likely the company's abysmal performance on the B-2 bomber programme sealed its fate. The USAF was extremely unhappy with Northrop's performance, he adds.
Matters started to turn sour within a year of the contract award. Despite being the driving force behind selecting the YF-22, the USN dropped out of the programme.
The early years of the F-22 development effort came right during the middle of the post-Cold War draw-down. Being the USAF's highest priority programme, the Raptor survived, but it did not emerge unscathed.
But the F-22 was not the first - nor will it be the last - aircraft to fall into the so-called procurement "death spiral". The Raptor was postponed, then suffered from weight problems and even faced "shortfalls" in its radar cross-section (RCS), even as costs increased. Its corporate sibling - the F-35 - is now trotting down the Raptor's well-worn path.
But unlike the F-35 programme thus far, the Raptor programme lost a variant - the two-seat F-22B combat-capable trainer version was eliminated in 1996 to help reduce costs. Four pre-production aircraft were also cut the same year, which, combined with the cancellation of the F-22B, reduced costs by about $1 billion.
As budgets fell, the USAF was also compelled to reduce overall procurement numbers from 750 to 648, and eventually down to 438 by 1996. The programme also suffered from the loss of a YF-22 prototype in April 1992, which left the air force without a flyable airframe until 1997.
More bad news emerged when the 1997 Quadrennial Defense Review further reduced the Raptor programme to 339 aircraft. But 1997 also marked the beginning of flight-testing for the stealthy fighter.
On 7 September, Lockheed's Paul Metz, a veteran test pilot who, as it happened, had flown Northrop's YF-23 during the ATF contest, took to the air in the first Raptor - tail number 4001.
Flying over the California desert, the Raptor quickly showed that it was in a league of its own. Even early on, it was apparent that the jet was faster and far more manoeuvrable than anything that had come before. Raptor test pilots described super-cruise flights where they would be "walking away" from their mostly clean-configuration fourth-generation chase aircraft - usually Lockheed F-16s and Boeing F-15s, which struggled in vain to keep up.
But even as the jet started to prove itself in testing, California Republican Congressman Jerry Lewis - who was then the chairman of the powerful House defence appropriations subcommittee - mounted a nearly fatal attack on the USAF's prized F-22 in late 1999. Lewis was ultimately unsuccessful, but six low-rate production aircraft were rebadged as test assets in an effort to enable the congressman to save face.
Low-rate production was finally approved in August 2001 and the production cap was raised to $45 billion, but with that amount the USAF could only buy 277 jets. The entire programme was capped at $72 billion, with cost overruns being paid for by sacrificing production numbers.
In an example of history repeating itself, on 20 March 2012, in his testimony before the Senate Armed Services Committee, current USAF secretary Michael Donley raised the prospect that further cost overruns on the F-35 would be made good by cutting production numbers.
But what ultimately sealed the Raptor's fate was Programme Budget Decision 753, which was issued by Rumsfeld on December 23, 2004. The memorandum axed over $10 billion from the programme and reduced the total buy to 179 aircraft.
The USAF did eventually manage to secure a multi-year deal for 60 aircraft in fiscal year 2007 that would stretch procurement past the George W Bush administration and raise the total number of aircraft to 183, but it could not reverse the decision.
Gates, who succeeded Rumsfeld, reaffirmed the decision, but allowed the USAF to buy four more aircraft. That raised the total to the final figure of 187 jets.
Meanwhile, the F-22 flight test programme was proceeding, but not as smoothly as had been predicted initially - as has also been the case on the F-35. And in another parallel with the JSF, software was the single biggest headache.
On the Raptor, where much of the jet's functionality is dependent on software, the computers would constantly shut down, requiring test pilots to reboot the system in mid-flight.
But there were signs of immense promise as well, says Lockheed's current F-22 programme manager Jeff Babione. He recounts a story from around 2002 about how Metz, during a test sortie, told a flight of F-16s where to point their radars but they still could not see him until he flew his F-22 right in front of the incoming fighters.
"That was the first time I realised just how good this airplane is," Babione says. "We really didn't have a good understanding of just how good it was going to be."
While the Lockheed team had the engineering numbers that gave them an understanding of how the aircraft might perform, the Raptor was revolutionary in concept.
"We were working on a revolution in the way airplanes were designed and the capabilities," Babione says. "We were always a bit suspicious as to whether or not those capabilities were actually going to be realised."
The F-22 was essentially an entirely new class of fighter, which the USAF has now termed "fifth generation". It was only with the start of operational testing in 2003 that Lockheed and the air force started to understand what they had developed.
Among the first group of eight hand-picked pilots who would start off direct initial operational test and evaluation with the 31st Test and Evaluation Squadron (TES) at Edwards AFB in California was then-Major Kevin Huyck.
When initial operational testing started, developmental testing was still concurrently under way at the base, Huyck says. But the orders were to find out what the jet could really do.
"One of the great vectors was from the chief of staff, Gen [James] Jumper," Huyck says. "He said: 'Don't fly it like an F-15,' and we took that to heart. It wasn't just a 'super F-15'... We were going to fly it and employ it like an F-22."
There were challenges, such as incomplete documentation. Nor did the air force fully understand how the aircraft worked and the maintenance system was in its infancy.
"When you're in developmental testing, you don't have all the answers," Huyck says.
But with the help of engineers, the pilots developed new documentation and some basic tactics for the new jet. The focus was on "validating" the Raptor's capabilities and obtaining a fielding decision, Huyck says.
"We took that airplane and we went to the edges of the envelope," he says. "We would push the edge of the tactics to see just how effective we could be." The aircraft met or exceeded every parameter set for it in the air, Huyck says. "It exceeded our awesome expectations."
Even Babione was awestruck by the jet's performance. "We were shooting for twice as good as an F-15," he says. "But the survivability and lethality were four or five times better than the F-15."
In some cases, as pilots point out, the Raptor is infinitely better because it can survive inside airspace defended by advanced integrated air defence systems. One pilot describes effortlessly "killing" more than half a dozen aggressor aircraft in a single engagement.
But operational testing also exposed how immature the F-22 was, particularly regarding its avionics.
"[We] had no shortage of emergency procedures and anomalies, as is to be expected," Huyck says. "Stability was definitely a challenge in the early days."
© Lockheed Martin
Even during the subsequent follow-on operational testing with the 422nd TES at Nellis AFB in Nevada and later at the formal training unit at Tyndall AFB in Florida, the F-22 experienced problems with its avionics. At the 422nd, F-15 pilots flying with the Raptor would carry checklists to help F-22 pilots in case they ran into problems.
Col Max Marosko, until recently the commander of the F-22 Weapons School at Nellis, was one of eight pilots who were part of the initial training cadre at the 43rd Fighter Squadron at Tyndall.
Like Huyck, Marosko cites a lack of available aircraft, caused by maintenance issues.
"Some of that I would say could be attributed to the growing pains of a new aircraft," he says. "Other portions were the result of not having enough aircraft."
That was because the USAF did not buy enough jets to do developmental testing, operational testing and training all at the same time - a problem that is also emerging on the F-35 programme as it begins to follow the same pattern of concurrency.
Later, as the aircraft matured and problems with the avionics and maintenance were solved, particularly at Nellis, there were simply not enough aircraft. The 422nd TES and Weapons School have to share their aircraft, but the air force has no other alternative, given the small fleet of F-22s.
But the Raptor has markedly improved since the early days of the programme.
"It definitely has," Marosko says. "From my three years at Nellis, I think I ground aborted maybe four or five times in three years."
Marosko attributes at least some of the improved performance to the newer-build Raptors, which are far more reliable than their older counterparts. But the dry Nevada weather helps, too. Raptors in the wet environment of Guam initially fared poorly, until they had some minor modifications to drain moisture.
"Some of the tail numbers that we had were in the 120s," Marosko says. "Those are much later produced jets: they're going to have a lot more robust parts that have been swapped out."
Babione does not disagree - Lockheed has worked diligently to improve each successive Raptor that rolled off its production line in Marietta. In the most recent jets, such as final aircraft 4195, reliability improvements and more robust stealth materials are incorporated. "That will be the most reliable Raptor we have ever built," Babione says.
In the early days of the programme, aircraft on the production line looked almost hand-built as workers learned how to build aircraft with new materials and to extremely low tolerances. New manufacturing techniques had to be developed, Babione says.
"The number of hours to manufacture and deliver the aircraft are one third of what they were on ship 10," he says. "That's a 78% learning curve."
These days the Raptor's avionics work almost perfectly, Huyck says. Malfunctions and software hiccups are rare and it is normal to fly an entire sortie without any anomalies. Of course, an occasional random error can occur, but it is exceedingly rare, he adds.
"That's natural when you're flying a fighter," Huyck says. "I don't think night and day are enough of a contrast to explain how much more stable and how much more confident we are in the systems."
Pilots say that the much publicised oxygen system problem that grounded the F-22 last year was vexing, but will eventually be solved. The Raptor is not perfect, but then no aircraft is.
© Lockheed Martin
The last of 187 Lockheed Martin F-22s ordered by the US Air Force took to the skies over Georgia on 14 March
The USAF is continuing to improve the Raptor with the Increment 3.1 upgrade that is now being fielded, and with the forthcoming Increment 3.2. The modifications will greatly increase the Raptor's capabilities in both air-to-air and air-to-ground realms. The service is also hoping to add open-architecture avionics to the F-22 in the future.
On 2 May, USAF leaders and executives from Lockheed - and presumably subcontractors Boeing, Pratt & Whitney and others - will gather in Marietta, Georgia, to mark the end of the production line and formally turn over 4195 to the air force.
Lt Col Paul "Max" Moga says he feels honoured to take delivery of the last jet.
"It's going to leave Marietta with two names on it: mine and my crew chief's, and the tails will proudly display 525 FS on them," he says. "I look forward to saying thanks to the entire F-22 team."
But despite the end of production, the Raptor will remain in the USAF inventory for decades to come and will need further upgrades. The air force has decided to retain the tooling for the aircraft so that it can perform major overhauls or extensive upgrades later. The tooling is being stored at the Sierra Army Depot in California along with extensive video archives to retain production knowledge, Babione says. But it also raises the possibility of restarting production or building F-22 derivatives one day.
"That would be an affordable way of getting totally new enhanced capability," Babione says. And, given the current fiscal climate, it would also be far cheaper than an all-new design. Larger delta wings, for example, would greatly add to the Raptor's range and offer the ability to carry far greater payloads.
But the true legacy of the Raptor lies in the F-35 - the radar, the engine and electronic warfare system are all based on the F-22's systems. Many of the same engineers also work on the project.
Richard Aboulafia, vice president of analysis at the Teal Group, sums it up: "Some of its legacy will live on, if the F-35 programme is truly successful."