The Space Reviewin association with SpaceNews

capsule illustration
A 1958 Lockheed illustration of the company’s proposed manned spacecraft. The vehicle on the left evolved into the pressurized Samos E-5 reconnaissance spacecraft. This was the US Air Force’s way of keeping its foot in the door of manned spaceflight at a time when that mission had been taken away from the military. (credit: NRO)

A sheep in wolf’s clothing: the Samos E-5 recoverable satellite (part 1)

<< page 1: Sentry Man In Space

The need for higher resolution

While all of these changes were occurring in the manned space program, there were new developments in the reconnaissance space program as well. In September 1958 the Air Force intelligence community issued a revised version of General Operational Requirement 80 (GOR 80), a document that formally stated the service’s requirements for satellite reconnaissance. The first version of GOR 80 had been issued in 1955, before the WS-117L satellite program even officially existed. The revised GOR 80 contained several addenda, including one for the “Visual Reconnaissance System.” The addenda stated that: “Development of the visual satellite will involve the progression from lesser to greater resolution as the state-of-the-art improves in satellite reconnaissance in order to realize an operational capability at the earliest date.”

The addenda also established some operational characteristics for the satellite, which it stated were not mandatory requirements but “ultimates” and could be sacrificed in favor of early availability. GOR 80’s authors explained the target goals for how much the satellite could see on the ground and the reasons behind them: “Resolution of photographic/visual images of low contrast objects from 20 feet to 5 feet [6.1 to 1.5 meters] in length on a side is required for production of most intelligence information, air navigation and target materials. Resolution of photographic images of low contrast objects one foot [0.3 meters] on a side on the ground is required for the production of technical intelligence.” GOR 80’s authors did not claim that this kind of resolution was achievable with a satellite system, only that it was what was required to provide quality intelligence.

Whereas this new reconnaissance document had indicated that higher resolution satellites were useful for intelligence purposes, it omitted a key piece of information. GOR 80’s authors failed to state a requirement for the amount of territory that needed to be photographed by such a satellite, both in terms of how much it could image in each photograph, and how much total territory it could photograph in a day or week or year. But they did declare a need to image areas off the ground path of the satellite, something that the existing Sentry film-readout cameras, which essentially pointed straight down, could not accomplish.

Exactly why the members of the Ballistic Missile Division decided in the fall of 1958 that they needed a large recoverable capsule is unknown and remains one of the important mysteries in this story.

GOR 80 did not represent a clear and unambiguous Air Force statement of need for a high-resolution reconnaissance satellite, but instead clarified the kinds of intelligence that could be derived from imagery of varying resolution. If the reconnaissance satellite was to provide useful intelligence, it would have to achieve relatively high resolution—at least 5–20 feet (1.5–6.1 meters). But none of the camera systems then in development was capable of achieving this kind of resolution. CORONA was expected to produce images of approximately 25 feet resolution (7.6 meters). Sentry E-1 was much worse. E-2 and E-3 were at the bare acceptability level for intelligence needs. But while they would not be much better than CORONA, they would also photograph a much smaller patch of ground. In effect, GOR 80 told the Air Force leadership that if they wanted to conduct higher-quality intelligence collection from space, they needed a new satellite to do it.

Around the same time that GOR 80 was revised, members of the Air Force Ballistic Missile Division (BMD) office that managed the Sentry satellite program in Los Angeles became convinced that they also needed to develop a recoverable satellite, like CORONA. Unlike CORONA, however, whose recovery vehicle was so small that it could fit inside an oil drum, they wanted a larger recovery vehicle.

Exactly why the members of BMD decided in the fall of 1958 that they needed a large recoverable capsule is unknown and remains one of the important mysteries in this story. But the decision to build a large recoverable satellite was apparently not explicitly driven by the new resolution requirements established by GOR 80. It may have been a coincidence that this new requirement for a large recoverable capsule emerged only a few weeks after Lockheed proposed a manned recoverable capsule and at the same time that NASA was undertaking Mercury. But the leaders of the Air Force space program, particularly Lieutenant General Bernard Schriever, clearly coveted the manned spaceflight role, and it is hard to avoid the conclusion that they were covertly trying to develop their own manned spaceflight capability.

In late September 1958, Air Force leaders in Washington, DC, undoubtedly acting upon the advice of members of BMD in Los Angeles, issued a directive that “consideration… be given to the use of a recoverable satellite in order to achieve maximum accuracy, information content, reliability of receipt of collected data, and reuse where economically feasible.” This was a rather vague set of requirements for a new satellite, but it provided high-level Air Force approval of an entirely new satellite system in addition to the ones already underway.

Despite this declaration by senior Air Force leaders that a new recoverable satellite program was necessary, the Air Force could not simply start building it. ARPA still had overall authority for military space. Air Force officers managed Sentry on a daily basis, but ARPA could ultimately approve or reject any decisions, and ARPA officials controlled the budgets—money allocated to military space projects did not actually belong to the military services that managed these projects. In fall of 1958 ARPA officials in Washington were apparently less enthusiastic about recovery techniques for Sentry than were officials at either Air Force Ballistic Missile Division or Air Force Headquarters.

By mid-December 1958, ARPA Director Roy Johnson approved a three-phase approach for Sentry that included film recovery, electronic intelligence (or “ferret”), and readout systems. He also tentatively approved an Air Force recoverable reconnaissance satellite. But Johnson apparently still saw the highly advanced E-3 system, and not a recoverable satellite, as a solution to the shortcomings of the E-1 and E-2.

This new recoverable satellite therefore posed a bit of a dilemma to the Air Force leadership. Because it represented an alternative technology to film-readout, and because film-readout had some glaring limitations, Air Force officials had to be careful that by arguing for the recoverable capsule they did not undercut the justification for the existing readout system. They had to sell the recoverable satellite to Washington, but be careful that they not justify it on the shortcomings of the E-1 and E-2.

By January 1959, with ARPA approval, the Ballistic Missile Division issued a revised “development plan” for Sentry that formally established the goal of developing a large recovery capsule, 60 inches (1.5 meters) in diameter, weighing 1,200 pounds (540 kg), and carrying a 600-pound (270-kilogram) payload. The spacecraft could use an ablative heat shield and had to be capable of being snatched out of the air by an aircraft, just like CORONA. Surprisingly, the BMD plan mentioned almost nothing about actual reconnaissance requirements, a significant detail that should have driven the spacecraft design. In many ways the Air Force was designing this new reconnaissance system backwards—defining the capsule first and leaving out the details of what would fit inside of it and, most importantly, what it would do. It was like designing a product to fit the box rather than a product that would actually sell.

Over these few weeks officials at BMD apparently decided that the new recoverable capsule would actually support two missions: mapping and high resolution photography. Their goal for the latter system was five-foot (1.5-meter) ground resolution, considerably better than any of the other reconnaissance satellites then in development and capable of meeting the requirements established by GOR 80.

The Air Force’s approach to the new recoverable satellite was totally unlike the CIA’s approach to designing the CORONA spacecraft. In March 1958 Richard Bissell, at the strong urging of his advisors, had abandoned CORONA’s initial spin-stabilized design that would have returned the entire camera to the ground in favor of a three-axis stabilized spacecraft that returned only the film to the ground inside a small recoverable capsule. Bissell did this in order to adopt a bigger and better camera, designed by the Itek Corporation. He told his people to design the spacecraft to meet the reconnaissance requirements and not the other way around. Bissell decided that the camera was the most important part of a reconnaissance satellite and everything else must support it. But the Ballistic Missile Division’s Sentry Program Office was approaching the task in the opposite direction, subordinating the camera considerations to the recoverable capsule’s design.

The Air Force had approved the recoverable satellite without even picking a camera to go inside of it. Without requesting competitive bids, the Air Force Sentry program office in Los Angeles awarded the recoverable satellite development contract to Lockheed. Lockheed was the “systems contractor,” which meant that instead of the Air Force negotiating separately with a launch vehicle provider, a spacecraft provider, a camera contractor, and so on, it signed a single contract with Lockheed which then was responsible for pulling all of these other components together. Lieutenant General Bernard Schriever, the former head of Ballistic Missile Division and now head of the Air Research and Development Command with overall responsibility for all advanced Air Force development, had advocated the systems contractor approach and it was now common for most Air Force space and missile contracts. Lockheed officials had to find a company capable of building a camera to put inside its pressurized capsule, and in early 1959 they set out to find one.

The Air Force’s approach to the new recoverable satellite was totally unlike the CIA’s approach to designing the CORONA spacecraft.

There were roughly half a dozen manufacturers of aerial photography cameras in the United States, but only three of them had any experience working on satellite cameras. These were Eastman Kodak, Fairchild Camera and Instrument Company, and the Itek Corporation. Lockheed approached Itek, which was then overseeing the development of the CORONA reconnaissance camera. In March 1959 Itek agreed to develop a camera for the new recoverable reconnaissance satellite. For Itek it was a big win, solidifying the small company’s position as a major satellite reconnaissance camera provider and expanding its customer base from the CIA to include the Air Force.

This camera was soon designated the Sentry E-5. But work on the project could not proceed until after April 3, when Ballistic Missile Division provided ARPA with a formal development plan for the overall Sentry program.

Samos E-5 on launch pad
A Samos E-5 spacecraft on the launch pad. (credit: USAF)

Selecting a recoverable capsule design

In April 1959 Lockheed outlined for the Air Force the flight objectives of the Sentry recovery capsule for both the high resolution camera and the mapping camera. The capsule had to have a diameter of five feet (1.5 meters), a payload capacity of approximately 500 pounds (225 kilograms) of film for the high-resolution mission, and a reentry accuracy of 30 miles (48 kilometers). It had to be recovered by air, but capable of surface ship recovery as a backup. The 30-mile reentry accuracy was a demanding requirement. By contrast, the reentry field for the CORONA was 60 nautical miles wide by 200 nautical miles long (111 by 370 kilometers).

Assuming that both the mapping and high resolution cameras utilized the same spacecraft, Lockheed engineers envisioned that the mapping version would weigh 4,390 pounds (1,990 kilograms) and the high resolution version would weigh 4,630 pounds (2,100 kilograms), with the primary difference being the weight of film carried. The actual payload, however, would weigh 1,100 pounds (500 kilograms) for the mapping version and 1,809 pounds (820 kilograms) for the high resolution version. Lockheed stated that the camera objectives were for 30 by 30 nautical mile (56 by 56 kilometer) photographs of specific targets at 5-foot (1.5-meter) resolution and 60 by 60 nautical mile (111 by 111 kilometers) photographs at 10-foot (3-meter) resolution. The camera also needed stereo capability and should be able to determine the location of its targets on the earth within one nautical mile (1.85 kilometers). First flight would be in January 1961, only 20 months away.

Before Itek could perform any work on designing a high-resolution camera for the recoverable capsule, Lockheed engineers sought to flesh out the capsule design. They proposed three different approaches. The first design would have used a rounded cone-shaped reentry vehicle, like a small Apollo capsule, with a highly unusual retrorocket system. The retrorockets, and small spin rockets intended to spin up the vehicle for stabilization prior to retrofire, would have been mounted atop a telescoping boom at the front of the vehicle. After the reentry vehicle had been ejected from the Agena, this boom would have extended out and the vehicle spun up. When the retrorockets fired, their exhaust would have traveled down past the sides of the vehicle. The boom would have ejected and the vehicle pitched up to enter blunt end first. After reentry, it would have ejected a parachute for recovery. This capsule could have carried either the mapping camera or a high resolution camera looking out the side of the vehicle, but it would have been a cramped fit for the latter. Lockheed’s preliminary design for the capsule would have placed the film supply and takeup reels inside the capsule, next to the camera, against the ablative heat shield.

The second proposed capsule was much bigger and consisted of a rounded cone atop a cylinder. It too used the telescopic retrorocket design. The cylinder would have contained the camera, which would have looked out the side of the vehicle, with the film supply and takeup reels on either side of the camera. This capsule would have been much larger than required for the mapping camera, but would have provided more room for the bigger high resolution camera. It also looked similar to Lockheed’s earlier manned Sentry proposal of April 1958, but with a larger diameter. Both the first and the second capsule proposals would have returned the cameras inside the reentry vehicle.

Lockheed’s third proposal was even stranger. A truncated cone-shaped reentry vehicle would have been mounted backwards, underneath a nosecone. Attached to its rear, under the nosecone, would have been the telescopic retrorocket. Unlike the two other designs, however, the reconnaissance camera would have been outside the reentry vehicle, between it and the Agena spacecraft, looking out the side of the spacecraft. Like CORONA, the camera would have been disposed of at the end of mission along with the rest of the spacecraft, a very expensive piece of trash. Although the briefing slides provide no justification for this design, clearly one major advantage was that it allowed for a simpler and smaller reentry vehicle and presumably a larger camera, at the cost of discarding the camera after each mission. All of the spacecraft would have been actively stabilized in three axes, using the same control system that was developed for CORONA.

At some point in this process, Lockheed added another requirement to the spacecraft design—the capsule had to be pressurized. “It was just specified: Thou shalt pressurize,” camera designer Jack Herther explained. This requirement had nothing to do with the reconnaissance mission, and whether Lockheed had added it at Air Force request or not remains unclear. Pressurization only complicated the spacecraft design. The pressure shell and pressurization system added weight that could have been devoted to payload.

Pressurization was unneeded for a film return system—CORONA was unpressurized. But the pressurization requirement allowed the Air Force, and Lockheed, to develop key technologies necessary to a manned space program. Although the Air Force was out of the manned space effort, building a large pressurized capsule capable of carrying a man kept the Air Force half a step behind NASA—as opposed to out of the race completely. They were cloaking a manned spacecraft program in the veil of a military reconnaissance satellite.


In April 1959, after reviewing the BMD development plan, the ARPA leadership gave specific approval to the new higher-resolution Sentry E-5 camera system and its recoverable capsule. But by implication ARPA withheld authorization for the mapping and charting camera that the Air Force also wanted to build. What was unknown to many Air Force and Lockheed officials was that at that same time the CIA was acting to incorporate a budding Army mapping satellite program known as VEDAS into the CORONA management system, where it was renamed ARGON. ARPA’s disapproval of the Air Force’s mapping camera was undoubtedly due to the CIA negotiations over VEDAS.

On May 25, 1959, ARPA formally canceled the Air Force mapping camera. Barely a month later, on June 23, ARPA also canceled the Sentry E-5 recoverable satellite program, directing that it be “deferred” pending a complete program review. ARPA also cut the Sentry budget by $25 million for 1960.

ARPA had become very unpopular in the sixteen months since its creation in early 1958 to manage military space programs. ARPA decisions had often seemed capricious and contradictory to leaders in the Army, Navy, and Air Force.

Exactly why ARPA officials made this decision killing the Sentry E-5 is unknown. According to an official history of the satellite program, ARPA cut both programs due to budgetary concerns, diverting the money to other space projects, although the history also implied that ARPA officials were not satisfied with Lockheed’s recovery approach. Even though the high resolution system had been proposed as an adjunct to the E-1 and E-2, it still represented an entirely different way of returning data and was a direct competitor for funding. ARPA director Johnson and others perceived shortcomings to the E-1 and E-2 and saw the E-3 camera—and not the high-resolution recoverable camera—as the solution to these shortcomings. Money for Sentry was finite, and since ARPA officials were more interested in the E-3 than the recoverable capsule, they canceled the recoverable capsule.

Air Force officials were incensed by ARPA’s decision to terminate the E-5 camera. Major General Osmond Ritland, Commander of Ballistic Missile Division, complained to Air Force Chief of Staff General Thomas D. White. Lieutenant General Bernard Schriever, the head of BMD’s parent organization, Air Research and Development Command, and a former commander of BMD, also complained to White. Schriever sent a letter to White declaring “should the ARPA decline to continue the recovery program… it is recommended that the Air Force immediately support this urgent development.” In other words, with ARPA refusing to fund the Sentry E-5, the Air Force would have to find additional money in its own budget to pay for the recoverable capsule, taking it from another non-space program. More importantly, because ARPA was officially in charge of all military space money, Schriever was essentially advocating a bureaucratic reorganization to remove ARPA from control of military space programs.

Responding for General White, Deputy Chief of Staff General Curtis LeMay told Schriever “I am completely sympathetic with your point of view and have taken action through Secretarial channels to restate the Air Force requirement to the director of ARPA and request reconsideration of its support in FY 60.”

LeMay took the issue to the member of the Air Force civilian leadership who was most knowledgeable about space issues. Air Force Assistant Secretary for Research and Development Dr. Joseph Charyk was then rapidly rising through the civilian Air Force leadership. Charyk, who had a Ph.D. in aeronautical engineering from Caltech, had previously served in an advisory role to Generals White and LeMay as the Air Force Chief Scientist, and had just taken over the position of Assistant Secretary for R&D. Charyk was an engineer with years of experience in missile guidance and rockets and was the chief Air Force civilian contact for the CORONA reconnaissance satellite, working closely with CIA official Richard Bissell to keep the program on track. Although his technical qualifications were impressive, Charyk also proved to be a skillful manager and an able bureaucratic warrior.

As Schriever’s letter indicates, ARPA had become very unpopular in the sixteen months since its creation in early 1958 to manage military space programs. ARPA decisions had often seemed capricious and contradictory to leaders in the Army, Navy, and Air Force. Director Roy Johnson also had an annoying habit of changing his mind about the projects under his authority. But opposition to ARPA was reaching critical mass in the Department of Defense.

In 1959 a law reorganizing the Defense Department had created the Directorate of Defense Research and Engineering (DDR&E), which like ARPA, was another independent Department of Defense organization. But ARPA had been created only through an executive order by President Eisenhower, whereas DDR&E had been created by a law, and was given greater power and authority. DDR&E was created to oversee all advanced research and development programs among all the armed services, whereas ARPA was consigned to space programs. In late August Charyk took his complaint about the recoverable capsule to the head of DDR&E, Dr. Herbert F. York. He pointed out that the recoverable capsule program was then the only way to meet the higher resolution requirements established by the Air Force’s GOR 80 intelligence statement in September 1958, and would only cost an additional $17 million in 1960, an amount that he argued was not extreme.

Apparently Charyk’s complaints to York worked, for on September 4, 1959 ARPA approved award of an E-5 camera contract to Itek, but did not approve the development of the Lockheed recoverable capsule. After further complaints by the Air Force, ARPA also approved development of spacecraft subsystems, including the recoverable capsule.

Meanwhile, in August, the Sentry program had been renamed Samos, apparently in order to de-emphasize the military mission of the reconnaissance system. The delays between March and September 1959 meant that Samos E-5 had made relatively little progress. General Schriever, with ample justification, blamed this on ARPA.

But the Pentagon agency’s resistance to the E-5 had not been completely without merit. ARPA officials apparently had qualms about the recovery approach selected by the Air Force, probably because the spacecraft was so much bigger than the existing, although still unproven, CORONA, and because a bigger spacecraft was naturally more complicated.

In November 1959, responding to the overwhelming complaints from the various military services, the Secretary of Defense formally removed control of military space programs from ARPA and turned the agency into a small technology development organization, which it remains to this day. But Air Force leaders found little time to rejoice, for the Directorate of Defense Research and Engineering, staffed with a number of highly-trained scientists and engineers, immediately took ARPA’s place with responsibility for making major program decisions for advanced space systems. What was worse was the fact that DDR&E Herbert York was a better bureaucratic warrior than ARPA’s Roy Johnson.

The struggle between readout versus recovery was beginning to break out into open conflict.

As York began to flex his bureaucratic muscles, the Samos E-5 spacecraft’s prospects improved, but not in a way that made Air Force officials happy. Whereas E-5 had been created to supplement the E-1 and E-2, York essentially viewed the primary question to be which form of data return—recovery or readout—was most likely to prove successful. In other words, E-5 now posed a threat to the readout spacecraft, which many in the Air Force still strongly supported. In November York ordered that the Air Force place primary emphasis on the E-5, and relegate work on the E-1 and E-2 to secondary status.

Ballistic Missile Division objected to this increased emphasis on recovery over readout, claiming that it would delay availability of an early operational satellite reconnaissance system—a readout system based on the E-1 and E-2—by 14 to 20 months, or until the first half of 1963. BMD asked for more money for the E-5 to keep it on track without jeopardizing the E-1 or E-2.

In January 1960, BMD submitted a revised development plan that was really a half-hearted effort to conform to the new directive from DDR&E. In essence, BMD kept the original E-1 and E-2 plan and simply added seven E-5 flights. The struggle between readout versus recovery was beginning to break out into open conflict. But in the meantime, the optical and engineering wizards at Itek, like Jack Herther, had started work on the E-5 camera. They were having their own problems.