The Space Reviewin association with SpaceNews

Samos 5 launch
The launch of Samos 5 on December 22, 1961. (credit: USAF)

A square peg in a cone-shaped hole: The Samos E-5 recoverable satellite (part 2)

<< page 1: clever compromises

Slow progress

In November 1959 the Secretary of Defense had called for speeding up the Samos E-5 schedule so that the first launch would occur before June 1961. But by January 1960 the first of seven E-5 launches was still scheduled for the second half of 1961, apparently having slipped a few months to August. In other words, despite the urgings of the Secretary of Defense, the schedule was slipping. The long time it was taking to develop the E-5 became even more troubling after Gary Powers’ U-2 spy plane was shot down over Siberia and President Eisenhower promised no further flights over Soviet territory. Satellites became the only method for returning reconnaissance photos, but the E-5 was still over fifteen months from first flight, and likely a year and a half from becoming operational. In July 1960 an Air Force working group evaluating problems with the Air Force reconnaissance satellite program stated that other recoverable payloads could possibly be developed faster than the Samos E-5.

At the same time that the E-5 was making slow progress, Air Force and Department of Defense officials were arguing over a far more basic issue: whether film-readout or film-recovery was the better approach for reconnaissance. Readout was immediate, whereas recovery took days. Many Air Force officers, particularly in Strategic Air Command, which flew the nuclear bombers and operated the ICBMs, wanted to use the satellite to warn of imminent attack. They essentially wanted a satellite burglar alarm system announcing that the Soviet Union was about to roll its tanks into Germany or launch a sneak bomber attack. This meant that timeliness of the data was more important than anything else, and therefore the Samos E-1, E-2, and moribund E-3 spacecraft deserved highest priority. Other officials, particularly civilians within the Department of Defense, but also apparently including some uniformed members of the Ballistic Missile Division’s Samos Program Office in Los Angeles, felt that the few pictures that a readout system would return would not really be valuable. They argued that the Air Force should emphasize the E-5 system, with its higher resolution and greater territorial coverage of the Soviet Union. In early 1959 Air Force officials had struggled to justify the Samos E-5 as a supplement to the E-1 and E-2 readout systems. Now it was a direct competitor.

At the same time that the E-5 was making slow progress, Air Force and Department of Defense officials were arguing over a far more basic issue: whether film-readout or film-recovery was the better approach for reconnaissance.

On February 15, 1960, the Samos Program Office in Los Angeles delivered its development plan to the Director of Defense Research and Engineering, Dr. Herbert York, who was now the Pentagon official with overall authority for military space. The new plan called for 25 Samos test flights from September 1960 to December 1962. Three would be for initial component testing (the E-1 satellite), eight for the E-2, seven for the E-5, and seven for electronic eavesdropping, or “ferret” systems. In response to a command that it place greater emphasis on the E-5 over the other satellites, the Air Force had simply added more E-5 flights to its existing schedule, rather than canceling some of the E-1 and E-2 flights. This was not what York had really wanted.

On April 20 York asked for adjustment of the Samos research and development plans. He wanted photographic payloads emphasized over ferret payloads, and film recovery—i.e. the E-5—emphasized over readout (the E-1 and E-2). In a second letter on the same day, York stated that the Samos readout program was moving too quickly towards operational capability without enough emphasis on research and development. What troubled him was that the Air Force was building expensive ground processing equipment for the photos from E-1 and E-2 before anyone knew if the satellites would actually work. He also specified the order for development of the individual elements of the system: the E-5 had top priority, followed by the E-1, E-2, and E-3, in that order.

Dr. Joseph Charyk, who as Assistant Secretary of the Air Force for Research and Development had appealed to York to save the E-5 from ARPA cancellation the previous summer, by spring 1960 had been promoted to Undersecretary, the number two civilian job in the Air Force. Charyk agreed with York that E-5 was being shortchanged by the Air Force. He recommended that E-5 get first priority in flight test. In summer Charyk directed BMD to revise its plan to provide for eight readout tests and seven E-5 tests. The latter included two diagnostic flights.

But not everybody who was critical of readout was enthusiastic about the Samos E-5 design. Bruce Billings, a member of DDR&E Herbert York’s staff, prepared a study of Samos in summer 1960. Billings argued that bandwidth limitations of the E-2 made it uncompetitive with recovery—it simply could not return nearly enough photographs to be worth the cost. Another E-2 drawback was its inability to take oblique shots—i.e. shots to either side of its ground track.

But Billings was also critical of the E-5. He felt that its recovery system was too complex. This tended to be a common criticism of the E-5; those who supported recovery over readout felt that the Air Force should have picked a different recovery system instead of the big, complex spacecraft that Lockheed was developing.

This entire debate took place amidst the continuing failure of the CORONA program. CORONA had suffered a string of failures—over a dozen by August 1960—and was proving frustrating to many within the intelligence community. Its managers were making progress toward success, but CORONA was not returning photographs.

In summer of 1960, following the May 1 U-2 spyplane shootdown, which suddenly halted the flow of aerial reconnaissance imagery of the Soviet Union, the White House undertook its own review of the overall Samos program. Dissatisfaction with Samos at the White House level had been spreading ever since direction of the program had been returned to the Air Force from ARPA the previous fall, but the U-2 was the catalyst for an overall review. A special review panel headed by White House Science Advisor George Kistiakowsky evaluated whether the existing Samos satellites then under development could satisfy intelligence community requirements, particularly the requirement to accurately measure Soviet ICBMs on the ground. They made this evaluation in light of a new list of reconnaissance requirements established by the United States Intelligence Board (USIB). The USIB had stated in July that the primary reconnaissance requirement was for high-resolution photos and a complete search of the entire Soviet landmass for ICBM launch sites—pointedly ignoring the requirement for timely data return that was at the basis of Air Force support of readout satellites.

By August, the review panel concluded that the existing Samos designs, including the E-5, could not accomplish the goals established by the USIB in July. They recommended that the Samos E-1 and E-2 essentially be canceled after each had made a single successful flight.

While the panel was not so hard on the E-5, the members felt that the camera’s stereo capability was too limited. Its small search area also probably factored into their criticism of the overall Samos program. They called for development of a new satellite system with resolution roughly equal to the E-5, but capable of covering far more area, in stereo. The Air Force had already decided that such a new system was needed and at the end of July had started initial work on it. It was designated the Samos E-6.

In the mid-1950s, as a graduate student studying satellite guidance at MIT, Jack Herther had come to realize that the biggest threat to a stable satellite was an internal moving mass.

Around this time additional program designations were applied to the various Samos efforts. The E-2 was designated Program 101A, the E-5 was designated Program 101B, and the E-6 was designated Program 201. These were their unclassified designations, used for situations where personnel without security clearances had to interact with those who had them. For instance, if an Air Force officer working on the Samos E-5 needed to travel from Los Angeles to Boston to talk to Itek officials, he had to file paperwork associated with his travel. The paperwork would state that he was traveling on business associated with Program 101B, and anyone who saw that paperwork would gain no clues about the nature of the work. But the camera designations E-2, E-5, E-6 were supposed to be secret. Itek had its own internal designation for the E-5 camera, which it called Project 9118.

By October 1960, Lockheed engineers had concluded that aerodynamic and thermodynamic analyses indicated that the reentry temperatures for the E-5 capsule would be higher than anticipated. They switched from titanium to beryllium for the nose cone fairing. NASA’s Langley facility was conducting tests to determine the aerodynamic stability of the E-5. Avco Corporation, whose expertise was in reentry systems, was also conducting tests of the ablative heat shield.


In the mid-1950s, as a graduate student studying satellite guidance at MIT, Jack Herther had come to realize that the biggest threat to a stable satellite was an internal moving mass. Any moving mass or rotating inertia on a spacecraft, such as a nuclear reactor turbine or the reels of a tape recorder, would impose a change in momentum of the satellite, which would start to turn in the opposite direction of the rotation. This movement had to be canceled out quickly to be within the allowable “angular dead zone” without exceeding the allowable blur rates—or the maximum amount that the image could move inside the camera when the film was being exposed. The movement had to be reduced mechanically, within the camera, almost as soon as it happened. But the spacecraft also had to reduce any motion that was left over. It sensed movement with its rate sensitive gyroscopes and damped it out with a cold gas jet stabilization system on the spacecraft.

Herther had taken this knowledge with him when he first worked on the CORONA at Itek and had used the expensive Pace Analog Computer to simulate the motions. CORONA’s panoramic camera lens assembly swung back and forth and moved linearly for image motion compensation inside the spacecraft. Its excess inertia was dampened out by the Agena control system without excessive blurring of the photos. CORONA also had film supply and take-up spools (reels) that turned, but the take-up spool was in the recovery bucket and was mounted 90 degrees to the roll axis of the spacecraft. It exerted its force against the long pitch axis of the vehicle, with its high moment of inertia, and therefore did not impart much movement to the overall vehicle. On the other hand, the supply spool was on the smaller roll axis of the vehicle and its inertia relative to the Agena roll axis inertia was such that every start caused a noticeable blur rate for a few frames until the Agena gas system counteracted the reaction.

Fortunately, the way CORONA was operated also negated problems caused by the initial activation of the camera and the rotating supply reel. “On the CORONA,” Herther explained, “they were generally search cameras.” Their mission was to “mow the lawn” over the Soviet Union, covering as much territory as thoroughly as possible. The camera ran during the entire south-north pass over the Soviet Union. “You turned them on as you came over and what we found was if we turned them on early enough, there was a little transient, but the gas system winds it down and then it goes okay until you shut it off. So generally the roll from the spool isn’t so bad.” In other words, the CORONA camera would quickly recover from this initial jolt and sail on to take smooth pictures during its transit over Soviet territory.

The E-5 was different. The Samos E-5’s big lens assembly was heavy and it moved, which also moved the spacecraft. First, it rotated back and forth, which caused the spacecraft to yaw slightly. This added a slight linear image blur along the length of the exposed film. But the designers considered the amount of blur to be negligible and did not compensate for it.

But the camera’s Image Motion Compensation System, which removed the image smear due to the movement of the vehicle over the ground, also moved the entire rotating lens assembly up and down while it rotated. “There’s a 250-pound [113-kilogram] lens going up and down, ten feet [three meters] from the center of mass of the spacecraft,” Herther explained. “So this would go up, the Agena would go down,” he demonstrated, tilting his hand up and down like a see-saw to show how the moving lens assembly would tend to move the entire spacecraft. The actual effect on the Agena was even more subtle than that: the combination of the lens yawing back and forth and also moving up and down actually exerted a rolling motion to the vehicle. Because the Agena had a low roll moment of inertia (meaning it could be rolled with less force than it could be pitched or yawed) this IMC motion force had to be canceled out. “We had two momentum balancers on either side of the lens assembly so that the vehicle would not get the pitch motions or the roll,” Herther explained. They were able to remove about 90% of the up and down motion of the IMC system.

But there was an even bigger problem. Just forward of the platen were the film supply and takeup reels and the various electronic systems needed to run the camera. The film reels had an inside diameter of 52 inches (132 centimeters) and an outside diameter of 54 inches (137 centimeters). Because the entire camera returned to Earth inside the capsule, the heaviest components had to be mounted as close as possible to the heat shield so as to ensure that the reentry vehicle descended heat shield first and did not tumble. To take its bursts of images, the camera had to transport six feet (1.8 meters) of film through the camera every second. It was far faster than the CORONA camera. “That’s movin’,” Herther explained. As a result, the heavy film spools turned fast as well.

What all this ultimately meant was that the images would be blurry and hence the E-5 could not possibly achieve the five-foot (1.5-meter) resolution it was supposed to.

Because of a lack of space in the confined capsule, the film supply and takeup reels also had to be mounted perpendicular to the direction of the vehicle, like coins inside a tube. This meant that they had the same rotation axis as the roll axis of the overall spacecraft. When they turned, the vehicle rolled in the opposite direction. It is easier to roll a long spacecraft around its long axis than it is to yaw or pitch it, just as it is easier to roll a pencil across a table than it is to flip it end over end. This rolling action had to be damped out immediately, or the movement of the spacecraft would blur the images.

This problem was also made worse by the fact that the E-5 camera carried a massive amount of film—250 pounds (113 kilograms)—and the entire system operated intermittently, which was a different manner than CORONA. At the time, the CORONA system, which had its first success in August 1960, carried only a paltry 16 pounds (7.3 kilograms) of film.

Dick Sementelli, the young technician who worked on the E-5, remembered how they removed some of the film spool inertia. “There was an ‘I-Omega Wheel’ associated with the supply and take-up spools that acted as a rotating inertia counterbalance as film transitioned from the supply to the take-up spool,” said Sementelli. “Omega” referred to the rate of rotation of the film spools and “I” referred to their inertia. He continued: “The spools rotated in opposite directions concentric to the roll axis of the capsule. Also mounted concentric to the spools was the I-Omega Wheel. The I-Omega Wheel was made of steel—stainless, I believe. My recollection as to the way it worked was that when the supply spool carried the full 250-pound film load, the I-Omega Wheel rotated in a direction opposite to it and at a velocity such that its rotating inertia momentum offset the combined rotating inertia of the full supply spool and the empty take-up spool. When both spools had equal amounts of film the I-Omega Wheel would be stationary. As more film moved to the take-up spool the I-Omega Wheel would begin to rotate slowly in the opposite direction to the take-up spool and would reach its maximum speed when the take-up spool was full.” This eliminated approximately 90% of the rolling motion.

With the CORONA, operators generally turned it on shortly before it crossed over Soviet territory. The gas jet system could then damp out any spacecraft motion before it reached the target and the cameras would continue running during the entire pass over the target until they shut down again after the spacecraft moved out of Soviet territory. But the E-5 camera in stereo mode would start up and take a burst of eight frames and then stop as the reflecting mirror at the front moved to a new angle. Then the camera would start again taking pictures at a new angle to provide stereo imagery. Thus the camera would be constantly starting and stopping the entire time it was over Soviet territory. The I-Omega wheel could remove some of the motions, but not all of them, and the gas jet system did not have enough time to damp out the remaining motions, whose reacting blur rates were much higher than on CORONA because of the heavy weights involved. “Because those things, when they start, for every burst of eight frames, every time you start, you’re going to get a transient to the Agena,” Herther said. “And the Agena roll axis inertia is low. So the roll rate is very high unless the gas system compensates for it very quickly and has it die out before a fraction of a frame.” The compensating system on the Agena had to be fast and had to work perfectly or the spacecraft would be in what the designers called “a continuous roll-rate transient” all the time it was taking pictures, rocking back and forth like a boat, control gas spewing out to correct the movements.

What all this ultimately meant was that the images would be blurry and hence the E-5 could not possibly achieve the five-foot (1.5-meter) resolution it was supposed to. Herther had used the Pace Analog Computer to simulate the motions and was startled by the results.

In this case, removing the remaining roll was not Itek’s responsibility, however. It was Lockheed’s responsibility, because they built the spacecraft and only the spacecraft’s cold gas jet control system could remove the remaining inertia. “As a camera configuration, this isn’t a good idea,” Herther said, tapping an illustration of his camera design with his finger over four decades later, “unless we worked very closely together.” It would require a group effort among the contractors. “And I didn’t see that,” Herther said. “In fact what I saw was that we had about 20 engineers on the job to check out equipment and the camera. Lockheed would come with probably 25 Lockheed guys and a couple of Aerospace [Corporation] guys. And they would be full of so many questions about ancillary things.” Later, Herther would concede that Lockheed probably did not have the money to seriously address the problem, because E-5 was always in danger of funding cuts. But at the time, in Herther’s estimation, the roll-rate transient problem was not getting fixed and it meant that the camera would not live up to its potential five-foot resolution goal.