A NEW SPACECRAFT

The vehicle which Grissom and Young would fly, and which would demonstrate many of the techniques needed for missions to the Moon, represented a stopgap effort to bridge the gulf between Projects Mercury and Apollo. On 7 December 1961, Bob Gilruth announced in Houston the approval of a $530 million project to use a large Mercury capsule for a series of two-man flights, launched atop the Air Force’s Titan II rocket, to practice rendezvous and spacewalking. It was originally dubbed ‘Mercury Mark II’ or ‘Advanced Mercury’, but the project name ‘Gemini’, with its nod toward ‘the twins’ of classical Greek lore, was suggested by Alex Nagy of NASA Headquarters. Nagy won a bottle of scotch for his trouble and the name was officially announced on 3 January 1962.

Until that time, Project Mercury remained the United States’ only approved manned spacecraft, although plans were afoot to develop it further, and two concepts in particular emerged: one for a temporary orbital station, housing two or more men for several weeks or months, and another for a manoeuvrable vehicle with sufficient aerodynamic ‘lift’ to adjust its flight in the atmosphere. In its 1960 budget request to Congress, NASA asked for $300,000 to study ways of transforming Mercury into a long-duration laboratory, a million dollars to explore methods of making it manoeuvrable and a further three million to investigate rendezvous techniques. ‘‘These modest sums,’’ wrote Barton Hacker and James Grimwood in their 1977 tome ‘On the Shoulders of Titans’, ‘‘signalled no great commitment. When NASA ran into budget problems, this effort was simply shelved and the money diverted to more pressing needs.’’

Still, the Space Task Group was interested in novel ways of controlling the landing of manned capsules and their attention was drawn to a technique devised by NACA engineer Francis Rogallo more than a decade earlier. He had worked on a flexible ‘kite’, with a lifting surface draped from an inflated fabric frame, which had the effect of producing more lift than drag; not as much, admittedly, as a conventional rigid wing, but it had the benefit of being foldable and lightweight. Early in 1959, Rogallo explained his concept to Gilruth, who was sufficiently impressed to implement further study of a follow-on, manoeuvrable Mercury spacecraft which could touch down precisely on land, thereby saving the cost of a naval recovery force. Other suggestions included a two-man Mercury capable of remaining aloft for three days, the addition of a 3 m cylinder at the back of the capsule to support two-week missions or even installing cabling to link the spacecraft to the booster, rotating them and providing experimental artificial gravity.

Unfortunately, with initial steps to develop Project Apollo, plans for advanced Mercury capsules were turned down by NASA Administrator Keith Glennan’s budget analysis team in May I960. Plans at the time amounted to achieving manned suborbital flight before the year’s end and an orbital mission thereafter. These would be followed by an unmanned flight to Venus or Mars by 1962, a controlled robotic landing on the Moon, an unmanned circumlunar mission around 1964 and eventually crewed space stations and circumlunar flights by 1967. Manned landings on the Moon, it was expected, would be a longer-term goal for the Seventies. Of course, this plan would change dramatically with John Kennedy’s speech to Congress in May 1961, but was limited at the time by the weight-lifting capacity of existing rockets and the widely-held assumption that lunar missions would be launched directly from Earth atop very large boosters.

The shift in climate from flying circumlunar missions to actual Moon landings began early in 1961, when George Low, head of manned spaceflight in NASA’s Office of Space Flight Programs, advocated Earth-orbit and lunar-orbit rendezvous techniques at the quarterly meeting of the Space Exploration Program Council. In February, he submitted a report to Bob Seamans, NASA’s newly-appointed associate administrator, stressing that orbital operations and large boosters would be needed, but that developing rendezvous techniques “could allow us to develop a capability for the manned lunar mission in less time than by any other means’’. By the end of that month, NASA Headquarters had taken greater notice and the possibilities of orbital rendezvous assumed centre stage in congressional hearings for the agency’s budget. The House Committee on Science and Astronautics also expressed an interest, scheduling a special hearing on the subject for May and recommending that NASA be awarded the full $8 million it had requested for rendezvous research. The Bureau of the Budget had previously cut this rendezvous spending down to just $2 million, but the committee’s inputs eventually led to NASA getting the funding it needed.

Elsewhere, Jim Chamberlin, an aeronautical engineer working for Toronto-based AVRO Aircraft Inc., joined the Space Task Group and was assigned by Bob Gilruth to oversee the development of an advanced Mercury capsule. Chamberlin seized the opportunity to effectively design a completely new spacecraft, retaining only the proven aerodynamic bell-like shape. In March 1961, at a weekend retreat in Wallops Island, Virginia, he described his plans to Gilruth and NASA’s head of spaceflight programs, Abe Silverstein, sketching out an ambitious machine with its equipment located outside the crew compartment in a self-contained module that would be far easier to install and test. One of Chamberlin’s suggestions was that the advanced

Mercury could be enlisted for circumlunar missions. Although Silverstein dismissed this lunar possibility, he and Gilruth expressed interest in the design itself and on 14 April the Space Task Group and McDonnell signed an amendment to the original Mercury contract, which provided for the procurement of long-lead-time items for six additional capsules. These items would then be used in support of what was now being dubbed ‘Mercury Mark II’.

McDonnell’s early efforts involved making no alterations to the shape of the spacecraft or its thermal protection system, but simply moving retrorockets and recovery equipment into modular subassemblies and, in Chamberlin’s words, creating ‘‘a more reliable, more workable, more practical capsule’’. It would transform, effectively, from an experimental machine into an operational one. By June 1961, when Chamberlin revealed his Mercury Mark II design, some members of the Space Task Group were taken aback: not only did it fulfil the key requirements of extending the spacecraft’s orbital lifetime and making it easier to test, but it essentially involved the repackaging and relocation of virtually every subsystem. This was needed, Chamberlin reasoned, because most of Mercury’s components were inside the cabin, meaning that equipment had to be disturbed in order to reach and fix one malfunctioning device. As it stood, Mercury could do its job, but was far from being a convenient and serviceable spacecraft. Chamberlin’s design allowed for any malfunctioning unit to be removed and tended, without the need to tamper with anything else. ‘‘If one system goes haywire,’’ said Gus Grissom, ‘‘you take it out and plug in a new one.’’

It also tackled the problem of Mercury’s sequencing system, in which many of its operations were automated for safety, by relying for the first time on pilot control; this, too, contributed to a far simpler machine. Chamberlin also advocated the inclusion of an ejection seat and eliminated the need for a Mercury-type escape tower, which he felt contributed hundreds of kilograms of weight to the capsule and argued that its extreme complexity made it inherently dangerous. Moreover, Mercury abort modes were automated, which could terminate a mission in some circumstances where such action may not be necessary. Flying an advanced Mercury with an ejection seat eliminated the option of using the Atlas – the seat could not push the pilots to safety quickly enough in the event that the rocket’s volatile liquid oxygen and RP-1 hydrocarbon mixture exploded. In its stead, Chamberlin suggested the Titan II, which the Martin Company had been developing for the Air Force as an intercontinental ballistic missile.

Martin had already proposed the Titan II as a candidate for lunar missions and, although both Seamans and Silverstein doubted its usefulness, they were sufficiently interested to ask Gilruth to explore ways in which it could be used for other manned projects. Two and a half times more powerful than the Atlas, the Titan seemed, to Chamberlin, perfect for lofting a correspondingly heavier Mercury capsule. The rocket was fed by hydrazine and unsymmetrical dimethyl hydrazine, together with an oxidiser of nitrogen tetroxide. In a catastrophic failure, an ejection seat would be able to outrun the fireball of these less-explosive chemicals. This combination of ‘hypergolics’, capable of spontaneously igniting upon contact, meant that the Titan needed no ignition system and, since they could be held at normal temperatures,

A Gemini-Titan launch. Note the absence of an escape tower; Gemini crews, aboard their conspicuous black-and-white spacecraft, relied instead upon an ejection system. Privately, many astronauts doubted its usefulness.

required no cryogenic storage or special handling facilities. Self-igniting propellants were intrinsically safer and easier to control than the violently-reactive cryogen used by the Atlas.

In any case, Chamberlin reasoned that because the Titan II was two and a half times more powerful than the Atlas, it would be possible to relax the constraints on the spacecraft’s weight. His decision to incorporate an enlarged overhead mechanical hatch in his modified Mercury, primarily as a means of emergency escape, soon expanded to fill another important requirement for lunar missions: the ability to conduct extravehicular activity (EVA), or spacewalking. Meanwhile, efforts to develop Francis Rogallo’s paraglider as a recovery system were gathering pace. The Space Task Group, however, which met with Rogallo and his team in the early months of 1961, felt that too much work had still to be completed before such an experimental device could be committed to a manned spacecraft. Questions were posed over its deployment characteristics, how it was to be packaged and whether the pilot’s view would be good enough to fly and land with it. They advised gathering at least six months’ worth of data before making a decision on whether or not to award actual development contracts. In May 1961, three $100,000 studies were authorised to design an effective paraglider and identify its problems.

Despite the changes to the launch vehicle, the escape system, the hatch, the packaging of components and the recovery operations, Chamberlin’s new spacecraft still resembled the bell-like Mercury capsule and, in its earliest form, was not expected to remain aloft for much longer than a day. Little interest was shown towards developing it further. Then, in July, the Space Task Group began looking at a so-called ‘Hermes Plan’, which envisaged a greatly expanded Mercury Mark II along the lines of that proposed by Chamberlin and, that same month, McDonnell’s Walter Burke outlined three possible forms of an advanced spacecraft. The first simply cut hatches in the side of the capsule to improve access to components, the second – valued at $91.3 million – adhered closely to Chamberlin’s proposal, whilst a third, $103.5 million suggestion envisaged a Mercury carrying not one pilot, but two.

This was not an entirely novel idea, having been brought to the table and quickly rejected in January 1959, but returned to the fore now that the capsule seemed likely to be extensively redesigned. ‘‘If we’re going to go to all this trouble to redesign Mercury,’’ said Max Faget, father of the spacecraft, ‘‘why not make it a multi-place spacecraft in the process?’’ In truth, Faget had already approached McDonnell several months earlier with a similar suggestion. Late in July 1961, Silverstein, Gilruth and several astronauts visited St Louis to view quarter-scale models of four basic spacecraft configurations, together with a full-size, wood-and-plastic mockup of a two-man Mercury, which Wally Schirra clambered inside. His first comment: ‘‘You finally found a place for a left-handed astronaut!’’

Humour aside, the visit proved pivotal, convincing Silverstein that Mercury should be extensively upgraded into a two-man machine. This decision was accompanied by President Kennedy’s commitment to a lunar landing before 1970, which prompted significant changes: the Space Task Group, based at the Langley Research Center in Virginia and originally devoted exclusively to Project Mercury, would be superseded by a Manned Spacecraft Center (MSC), to be situated near

Houston, Texas, as part of a much wider, far more complex and infinitely more expensive effort to land a man on the Moon. Rendezvous provided a means of achieving this goal far sooner than a direct-ascent method and the growing conviction throughout the summer and autumn of 1961 that rendezvous needed to be utilised in some form would provide a framework for what would become Project Gemini.

With the approval of the new project came more emphasis on the Titan II as its launch vehicle. Even though the contract with the Air Force to build the rocket had been signed scarcely a year earlier, Martin’s James Decker proposed that NASA purchase nine Titans for a bargain price of $48 million, the first of which could be ready to fly by early 1963. Among the modifications needed to make it suitable for the Mercury Mark II were lengthened second-stage propellant tanks to increase its payload by 300 kg. Also, the risk of first-stage ‘hardover’ – a malfunction in its guidance system which could drive the gimballed engines to their extreme positions, thereby subjecting the Titan to massive dynamic overloads – could lead to the rocket breaking up before the astronauts could react. A second first-stage guidance system was added to erase this risk.

At the same time, if rendezvous was on the agenda, a rendezvous target was needed and, in August 1961, Chamberlin made his first contact with the Lockheed Missile and Space Company in Sunnyvale, California, with a view to using its highly – successful Agena-B rocket stage. Like the Titan, the Agena ran on storable hypergolics – unsymmetrical dimethyl hydrazine with an oxidiser of inhibiting red fuming nitric acid – and had a ‘dual-burn’ capability; in effect, it could be fired, shut down, then fired again. It also had the potential for the Mercury Mark II, after docking with it, to demonstrate advanced manoeuvres.

The growing importance of rendezvous, docking and manoeuvring was such that the new spacecraft was beginning to change into a new project in its own right and another question that it would be pressed to answer would be the effect of long-term missions on the human body. A journey across the 400,000 km gulf to the Moon and back was expected to require a flight lasting almost two weeks and Mercury Mark II might not only be able to demonstrate that such missions were survivable, but also could evaluate advanced technologies such as electricity-generating fuel cells and more stable attitude-control propellants than hydrogen peroxide. The astronauts, too, would need their own ‘modifications’, in the form of improved space suits to support longer missions.

Ten Mercury Mark II flights, the first in March 1963 and the last in September 1964, would be launched every two months to fly men for up to seven days and animal passengers for as long as two weeks. Investigation of the Van Allen radiation belts around Earth was a second major objective and, indeed, the first flight would be an unmanned test to ensure that the Titan and Mark II were compatible and to boost the capsule to an apogee of 1,400 km. Controlled landings would be the third goal, to be accomplished on each manned mission, most likely with the aid of a paraglider, and rendezvous and docking stood fourth. Later flights would require dual launches of the Titan II and the Agena-B, such that the Mark II could rendezvous and dock. The hope was also there that, if the spacecraft achieved all of its objectives without problems, particularly the long-duration aspect, a Mark II could be launched to dock with a liquid hydrogen-fuelled Centaur stage and boosted onto a circumlunar trajectory. Some short-lived plans even envisaged a manned around-the-Moon mission as early as May 1964. Although they did not get far beyond the drawing boards, one of Chamberlin’s ideas included launching a Mark II atop a Saturn C-3 rocket and placing a manned craft on the lunar surface. Such a landing could, he suggested, be achieved late in 1966.

By the end of October 1961, however, greater emphasis had been placed on developing rendezvous techniques and flying long-duration sorties; Van Allen studies, animal flights and lunar missions were gone. There would be a dozen Mark IIs: an unmanned precursor, followed by an 18-orbit manned mission, a series of extended-duration flights of up to 14 days and, later, rendezvous and docking exercises. Two weeks after NASA formally announced its intention to proceed with Mercury Mark II, on 22 December 1961, James McDonnell signed the contract for its development, agreeing that his company would provide full-scale spacecraft mockups within six months and a mockup of the Agena-B target adaptor by October 1962. By the following March, read the provisions of the contract, McDonnell would supply the first flightworthy spacecraft, with others to follow at 60-day intervals until 12 had been delivered. In effect, this contract replaced the earlier one to procure long-lead-time items for extra Mercury capsules.

It was shortly after the awarding of the main contract that McDonnell began to subcontract out several systems which would prove instrumental in demonstrating the capabilities of the new spacecraft, by now known as ‘Gemini’. One of these was the Orbit Attitude and Manoeuvring System (OAMS), which not only allowed the astronauts to ‘steer’ their spacecraft, but also helped them to station-keep and push away from the second stage of the Titan II. It comprised 16 small engines, fed by hypergolic mixtures of monomethyl hydrazine and nitrogen tetroxide. Each engine was mounted in a fixed position and ran at a fixed thrust level. Eight of them were rated at 11 kg of thrust and provided attitude control. These fired in pairs, permitting the Gemini to roll, pitch and yaw. The remaining eight were ‘translational’ thrusters, each capable of 45 kg of thrust, and were oriented in pairs to fire forward, backward, up/down and left/right. This would form the ‘manoeuvring’ component of the system, although the thrust level of the two forward-firing engines was reduced to 38 kg in July 1962. The re-entry controls, developed by the same subcontractor, North American Aviation’s Rocketdyne Division of Canoga Park, California, consisted of two independent rings of eight 11 kg thrusters located in the nose of the Gemini, forward of the crew cabin. After the main manoeuvring system had effected retrofire, either ring could control the attitude of the spacecraft during re-entry. By the end of May, all major subcontractors had been selected to begin work.

Meanwhile, efforts to secure the Titan II and Agena required Bob Seamans and Assistant Secretary of Defense John Rubel to agree that the Air Force would act in the capacity of a contractor to NASA, whilst at the same time allowing the former ‘‘to acquire useful design, development and operational experience”. The Los Angeles-based Space Systems Division of the Air Force Systems Command would handle the development of the Titan and Agena for Mark II operations and, early in January 1962, NASA issued formal instructions for work on the rockets to begin. Adapting the Titan to handle the spacecraft proved more complex than originally expected, since it required new or modified systems to ensure the astronauts’ safety during the countdown and ascent. By the beginning of March, Lockheed had also started work on the development of a more advanced Agena-D, with an engine capable of being fired more than twice, which would be boosted into orbit atop an Atlas. Unlike the Agena-B, this new version boasted a radar transponder, a forward docking adaptor and an improved attitude-control system. At this time, the first – unmanned – launch of Gemini had slipped from May to August 1963, although the three inaugural flights would be spaced out at just six-week intervals.

Elsewhere, the contract for the paraglider, intended to guide Gemini to a touchdown on land, was awarded to North American on 20 November 1961, but its real future seemed less assured. Chamberlin had defended it vigorously, but Max Faget’s engineering directorate within the newly-established MSC in Houston was cool to the idea, considering its reliability as lower than having main and backup parachutes. Faget instead advocated a steerable parachute, together with landing rockets to cushion the touchdown. Others, including Chris Kraft, felt that neither the paraglider nor the ejection seats were reliable and posed enormous practical obstacles to safety. The paraglider was not aided by North American’s slow progress on its development, which had been unavoidable as Gemini grew from little more than a modified Mercury into an entirely new – and bigger – spacecraft. North American planned free flights of a half-size paraglider for May 1962, although this was delayed because backup parachutes were needed to avoid losing the costly test vehicle. Initial plans called for the first unmanned Gemini to land with ordinary parachutes and the second (manned) flight to utilise the paraglider, although by mid – June it became clear that it would not be available until the third mission. Still, a maiden launch in August 1963 did not seem unreasonable.

However, as 1962 wore on, it was apparent that project costs would be far higher than anticipated. Modification of the Titan II, for example, had climbed from $113 million to $164 million, owing to a multitude of changes to ‘man-rate’ it. These included a fully redundant malfunction-detection system, backup flight controls, an electrical network with backup circuits for guidance, engine shutdown and staging and new launch tracking hardware. Costs of developing the Agena-D similarly increased from $88 million to $106 million and the Gemini spacecraft itself ballooned from $240.5 million to $391.6 million. This cost hike came as a huge surprise, yet it encompassed McDonnell’s enlarged view of what should be included in the project: from ‘realistic’ flight simulators and trainers in Houston and at Cape Canaveral to structural mockups for static and dynamic tests and even the development of an extra spacecraft and docking adaptor for an extended series of unmanned orbital missions (dubbed ‘Project Orbit’) ‘‘to investigate potential problems and to evaluate engineering changes’’. When the Office of Space Flight submitted its Project Gemini review to Administrator Jim Webb in May 1962, the cost of the overall programme had climbed markedly from $529 million to $744 million – and continued to grow.

The half-size emergency parachute experienced difficulties of its own. In a series

A model of the Gemini paraglider under test.

of drop tests at the Naval Parachute Facility in El Centro, California, between May and July, four failures led to an extensive redesign and placed it and the paraglider two months behind schedule. Its full-size counterpart also suffered problems and, despite some successes, all three parachutes failed during a November 1962 test and the test vehicle was destroyed when it hit the ground. Although these woes did not directly affect its potential performance, they did introduce worries. Early tests of a half-sized paraglider at Edwards Air Force Base in mid-August had failed to deploy properly after being towed to altitude by helicopter and in two subsequent attempts it was released too soon and landed too hard. A fourth try failed to deploy and a short circuit cancelled a fifth. By the end of September, even Jim Chamberlin was losing patience and ordered North American to halt all tests until it could spell out how it intended to correct the problematic electronics and pyrotechnics. After rework, the half-scale paraglider was shipped back to Edwards and, on 23 October, sailed through a perfect flight, finally demonstrating its stability.

Even the development of the Titan II rocket presented problems. In its first flight on 16 March 1962, it began to experience longitudinal vibrations, occurring 11 times per second for about half a minute. Although these did not pose a problem for the Titan, they would pose a risk for the astronauts, who would be exposed to two and a half times normal gravity and might not be able to react properly to an emergency. The vibrations, nicknamed ‘pogo’, partially disappeared thanks to higher pressure in the rocket’s first-stage fuel tank – perhaps, engineers speculated, it was caused by oscillating pressure in propellant lines – and Martin suggested installing a surge – suppression standpipe in the oxidiser line of later Titan IIs.

Escalating costs and a spending cap limited to $660 million for 1963 eventually led to the realisation that Gemini could only go ahead with the cancellation of the paraglider, Agena and perhaps all rendezvous hardware. Surprisingly, in the subsequent rescoping of the project to take account of budget limitations, the paraglider survived almost untouched and the spacecraft itself retained most of its original features, although the Titan II testing programme was drastically reduced and the decision as to whether the Agena had a role in the project remained fluid for some months. These difficulties conspired to delay the first unmanned mission to December 1963, followed by a piloted flight three months later.

Meanwhile, development of the paraglider continued to be mired with problems. After its October 1962 success, North American refitted the half-scale test vehicle and shipped it back to Edwards for a late November flight. Minor electrical problems postponed the attempt and, when it eventually flew on 10 December, its performance was disappointing: the capsule tumbled from the helicopter, fouled the stabilising drogue parachute and the inflation of the paraglider wing only made matters worse. When the capsule spun down past 1,600 m – the minimum recovery altitude – a radio command detached the wing and allowed it to descend on its emergency parachute. Another try a few weeks later was worse still: it did not tumble this time, but the paraglider storage can was late in separating. The capsule was falling too fast when the wing started to inflate and its membrane tore. Moreover, a faulty squib switch meant the main parachute failed to deploy and the capsule crashed. Despite reporting that five distinct failures had been identified and repaired,

North American’s paraglider was about to breathe its last. Chamberlin gave the project a final chance, but another attempt to deploy a half-scale wing on 11 March

1963 concluded dismally when the storage can failed to separate. The emergency parachute then failed and the paraglider, figuratively and literally, ended its days as a heap of smouldering wreckage.

For the Titan II, the key problem was overcoming its ‘pogo’ oscillations and Martin duly installed a surge-suppression standpipe. However, a test flight on 6 December 1962 actually worsened the pogo effect and, indeed, induced such violent shaking that the first-stage engines shut down too early. Two weeks later, another rocket with no standpipes and increased fuel-tank pressures launched successfully and exhibited lessened pogo levels. A third launch on 10 January 1963 produced similarly encouraging results and the G forces to which it would have exposed a human crew were only slightly higher than those tolerated by Mercury pilots. On the other hand, in this case, the Titan’s second-stage thrust was half of what it should have been, suggesting that its engines had difficulty reaching a steady burn after the shock of ignition. This ‘combustion instability’ proved somewhat more complex than the pogo obstacle and led to a decision in March 1963 to increase the number of unmanned tests and reduce the total of manned flights to ten. To make matters worse, on 8 March, Gemini project cost estimates topped a billion dollars. Days later, Bob Gilruth relieved Chamberlin of his duties and replaced him with Charles Mathews.

Among Mathews’ earliest moves was to insert an unmanned mission in place of one of the manned flights, largely in response to the ongoing Titan II problems. Scheduled for December 1963, the new mission would demonstrate the Titan as Gemini’s launch vehicle. After the upper stage had achieved orbit, a ‘boilerplate’ spacecraft would remain attached and the entire assembly would be left to fall back into the atmosphere. An unmanned suborbital flight with a real spacecraft in July

1964 would then show the ability of the spacecraft to support manned missions. On this plan, the first manned mission, Gemini 3, would come in October 1964. Originally intended as a day-long, 18-orbit mission, Gemini 3 would be reduced by nervous managers to around three orbits – or five hours – and would test the spacecraft’s systems. Earlier plans to fly a Rendezvous Evaluation Pod (REP) on the first manned mission were rescheduled for Gemini IV, which would run for seven days in January 1965. The new schedule implemented three-month gaps between each manned mission, in response to concerns that equipment checkouts and astronaut training required more time. Gemini V would then conduct the first rendezvous and Gemini VI would attempt a 14-day mission, to mimic the length of a full-duration lunar landing expedition. Subsequent flights would last three days apiece, each consolidating and extending rendezvous and spacewalking expertise.

Interestingly, Mathews’ plan did not entirely omit the paraglider, but pushed its maiden flight back to Gemini VII, with parachutes supporting earlier missions. The plan was approved at the end of April 1963. Although paraglider tests in May and June satisfactorily proved its stabilising parachutes, a final drop on 30 July suffered a total failure and crashed. By the year’s end, the paraglider was itself being challenged by the concept of the ‘parasail’, a flexible gliding parachute which offered a quick

and relatively cheap device to achieve a touchdown on land. It could be ready, McDonnell told NASA Headquarters in September 1963, in time for Gemini VII and at a cost of just $15.7 million. It was ruled out, partly due to lack of funds, but chiefly because the paraglider’s vocal supporters objected to giving up on an effort that had already consumed much time and work and was almost ready for flight testing. However, the paraglider itself was on the wane: its landing system programme was stripped of all objectives, save that it would prove the ‘technical feasibility’ of the concept. Parachutes would support most of the Gemini missions, with the paraglider possibly used for the last three. Although Bob Gilruth insisted that it might still fly on Gemini if its tests were successful, the first mutterings of cancellation had reared their heads.

Meanwhile, the Titan Il’s woes continued. A launch on 29 May 1963, carrying pogo-suppression devices for both oxidiser and fuel, burst into flames during liftoff, pitched over and broke up. Although the pogo devices themselves were absolved from blame in the incident, the flight was too brief for their effectiveness to be judged. Three weeks later, a military test of the missile from a silo at Vandenberg Air Force Base in California, although trouble-free during first-stage ascent, with pogo levels within acceptable limits, exhibited faltering second-stage thrust. Had a Gemini crew been aboard, it would have been an abortive mission. The Air Force now shifted its focus to ensuring that the Titan worked as a missile, first, before committing it as a Gemini launch vehicle, and some within NASA even came to doubt that it was the right rocket for the job.

In fact, concerns were so high that the space agency even considered adding yet another unmanned Gemini flight to test the Titan, pushing the total number of missions to 13. Designated ‘Gemini 1A’, it would be slotted in, sometime around April 1964, between the first and second unmanned launches and, like Gemini 1, would consist of a ‘boilerplate’ capsule equipped with instrumentation to examine the performance of the rocket. It would only fly, however, if the unmanned Gemini 1 failed to meet all of its objectives. Although the Gemini 1A hardware was delivered in September 1963, the extra mission had been cancelled by February of the following year, thanks to improved prospects for the Titan II. Despite the fact that two launches in August and September 1963 had gone wrong due to short circuits and guidance malfunctions and the effects of pogo were higher than expected, circumstances improved as the year drew to a close. A Titan launch on 1 November, equipped with standpipes for its oxidiser lines and mechanical accumulators on its fuel lines, reduced pogo effects well below the limit demanded by NASA. Moreover, in the next five months, the rocket would score an impressive and unbroken chain of successes, enough to ‘man-rate’ it in time for Gemini 1.

For all of these problems, the development of the spacecraft itself was going relatively well. Key problem areas remained the fuel cells, propulsion and the ejection seats. McDonnell had already subcontracted to General Electric to build the cells and the first serious development problem of preventing oxygen leakage through its ion-exchange membrane was soon resolved. However, resolution of this problem produced another: test units working over long periods showed degraded performance, apparently due to contamination of the membrane by metal ions from the fibreglass wicks responsible for removing water from the cell. Leakages in the tubes which fed hydrogen to the cells created another obstacle. General Electric replaced the fibreglass wicks with Dacron cloth and an alloy of titanium-palladium replaced the pure titanium tubing, although these developmental headaches pushed the project further behind schedule. NASA was sufficiently concerned to request McDonnell to conduct an evaluation of batteries to be used on Gemini 3, the first manned mission in October 1964, with fuel cells aboard, but only used for flight qualification purposes.

Ongoing problems with the fuel cells, it was realised, could restrict Gemini missions to only a few days under battery power alone. In November 1963, Charles Mathews issued instructions to adapt the electrical system to house batteries or fuel cells and, within weeks, the decision to fly Gemini 3 on batteries was official. Nonetheless, the unmanned Gemini 2 would be equipped with both systems to qualify them. A month later, Mathews decided that Gemini IV – the proposed seven – day mission – should also utilise batteries, which would have a corresponding impact on its duration, shortening it by almost half. Eventually, it would be Gemini V that would first demonstrate the use of fuel cells in space on a week-long flight, with mixed success.

The ejection seats were another concern; so much so, in fact, that the MSC even considered replacing them with an escape tower, akin to that used during Project Mercury. Simulated ‘off-the-pad ejection’, or ‘sope’, tests had been suspended late in 1962 until all components were ready. During one test, the overhead hatch failed to open and the seat shot straight through the 5 cm-thick hull, prompting John Young to remark: ‘‘That’s one hell of a headache… but a short one!’’

Ultimately, the escape system involved a balloon-parachute hybrid, known as a ‘ballute’, which would prevent the astronaut from spinning during freefall if he had to eject from an altitude higher than the 2,000 m at which his personal parachute was supposed to open. The first tests in February 1963 were disappointing: the ballutes failed to inflate and the personal parachute did not deploy correctly, although managers felt that the problem was a dynamic one, caused by the relationship between the rocket motor’s thrust vector and the shifting centre of gravity of the man-seat combination. By May, sope testing resumed and met with greater success and, the following month, dual-seat ejections were demonstrated. Nonetheless, rising costs and unending technical problems, involving the spacecraft, the Titan, the paraglider and even the Agena-D, had conspired to delay the first unmanned Gemini launch until the spring of 1964.

For the paraglider, which experienced yet more failures in December 1963 and February 1964, the end was in sight. Indeed, NASA’s public stance was now that a land touchdown was riskier than an already-proven ocean splashdown, although the Air Force, which was planning its own version of Gemini in conjunction with an orbital laboratory, kept the concept alive for a time. However, the military was not keen to commit funding to the paraglider until it had first been satisfactorily demonstrated by NASA. After yet another test failure on 22 April 1964, William Schneider, Gemini’s project chief at NASA Headquarters, officially announced that no more money would be spent on the paraglider. Ironically, its cancellation was actually followed by several wing-deployment successes. North American even invested its own cash in further development work and, in December, a pilot flew with the helicopter-towed vehicle and guided it to a safe landing. It was, however, too late for it to be reinstated into Gemini.

Of course, the design of the capsule itself had long since been finalised. In fact, the re-entry capsule was comparable in size to the Mercury spacecraft. A broad conical adaptor at its base, which would be shed following retrofire, held the propulsion and long-term power and life-support systems. The capsule, wrote Neal Thompson in his biography of Al Shepard, was “like a snug little sports car”. The white-coloured adaptor was 2.28 m high and 3 m wide at its base and was itself split in two: an equipment section for fuel and propulsion and a retrorocket to support re-entry. Both segments were isolated from each other by means of a fibreglass honeycomb blast shield. The crew cabin, meanwhile, was a truncated, titanium – and-nickel-alloy cone, measuring 2.28 m wide at its base and 98.2 cm at its apex and topped by a short cylinder for re-entry controls and parachutes. In total, the cabin was 3.45 m high and, when mounted on the adaptor, the full Gemini stood 5.73 m tall. Inside, conditions were cramped. The Gemini’s total pressurised volume was little more than 2.25 m3 – “like sitting sideways in a phone booth,’’ John Young said years later – and its outward-opening hatches were positioned directly above each astronaut’s head. Each man was provided with a small, forward-facing window. The spacecraft was equipped with horizon sensors and an inertial reference system and the command pilot flew using a pair of hand – controllers, one for translation, the other for orientation, whilst referring to an ‘artificial horizon’ display.

Key to its manoeuvrability were the 16 OAMS thrusters, spaced around the capsule, and power, at least before the first demonstration of fuel cells, came from three silver-zinc batteries. During preparations for re-entry, the equipment section of the adaptor would be jettisoned, exposing the retrorocket package, whose four motors would initiate the Gemini’s return to Earth. After retrofire, the package would itself be released, leaving the crew cabin, protected by an ablative heatshield and radiative shingles, to withstand the intense heat of re-entry. The centre of mass was deliberately offset to generate aerodynamic lift during re-entry and rolling the capsule using the thrusters on the nose enabled the trajectory to be controlled to aim for a specific geographic position. A parachute-aided descent would finally bring the capsule gently into the Atlantic or Pacific.