RELAY IN THE SKY: THE SATELLITE DATA SYSTEM DWAYNE A. DAY Space Studies Board, 500 Fifth Street, NW, Washington, DC 20001, USA

Posted with the permission of the author.  Originally appeared in JBIS, Vol 59, Suppl. 1, pp. 56-62, 2006.

 One of the key developments in real-time satellite reconnaissance was the Satellite Data System, or SDS. The SDS is a constellation of communication satellites placed in highly elliptical-inclined orbits that relay imagery from low-altitude reconnaissance satellites back to the United States. The original concept for satellite data relay dates to the late 1950s, but the modern concept of the system was conceived in the mid-1960s and the satellites entered development in the early 1970s. The satellites were built by Hughes and at least a dozen of them were built and launched through the mid-1980s. They have apparently been replaced by newer and larger satellites starting in the early 1990s.


In fall 1963 Albert “Bud” Wheelon was sitting in his living room in a suburb of Washington, DC, watching a football game being played in San Francisco. At the time Wheelon was the Deputy Director for Science and Technology at the Central Intelligence Agency. He was building his directorate into a powerful research and development organization that initiated and funded aerial and space reconnaissance systems for the United States [1].

While watching the football game Wheelon had an epiphany. “lt suddenly struck me that if I could do that, the technology was available to view the Earth’s surface from orbit and to observe that scene as it was being received in the spacecraft, Wheelon said. “ln other words, to develop a near-real-time imagery system.” The next day Wheelon called one of his deputies and assigned him to work fulltime on the development of such a system. “This turned out to be one of the best decisions of my career,” Wheelon said. “We ended up creating a new class of satellite imaging systems that revolutionized intelligence collection.”

It took a decade to develop the technology for such a near-real-time system, known as the KH-11 KENNAN and first launched in late 1976. The ClA helped fund work at Bell Telephone Laboratories on what eventually became charge-couple devices, or CCDS, which today are readily available in commercial digital cameras.

But there was another hurdle that they needed to overcome: getting the data back to a ground station. “This was a major issue because the optical data were so vast in each frame that we could not afford to store it on board the satellite”, Wheelon explained. “The data simply would pile up and overflow the limited available storage devices.“ In fact, an early attempt at doing this with the Samos E-1 and E-2 satellites was limited to the transmission of only a few photographs during each pass over a ground station, “We could not downlink the data as it was being collected, because it would have to be received in denied territory”. But Wheelon had a solution for that as well. “The right way to do this was to uplink the data to a relay satellite in much higher orbit, which could then pass it on to the ground station” [2].


Wheelon was not the first person to develop the idea of a data relay satellite. In August 1958 Lockheed Missile Systems Division prepared a slide depicting a “Sentry Data Relay Concept.” The slide depicted data and commands being relayed between three satellites in low Earth orbit and a ground station. This was a means of connecting a satellite with a ground station that was far below the horizon. But this method of relaying data between satellites in low Earth orbit was not practical in 1958 with satellite communications and electronics in their infancy. In fact, it was not until the late 1990s that the technology was perfected and utilized for the lridium low Earth orbit communications system.[3]

The biggest problem was that of locating another satellite to relay data through. Because all of the satellites were in low orbit and moving relatively fast to each other, no satellite would be visible to another for very long before moving below the horizon. Before the days of sophisticated computers it was impossible to compute what satellites would be in range, where they would be located, and then point an antenna at them, switching to the next available satellite when needed.

Another problem at the time was bandwidth, or the amount of information that could be sent over a data link, something hat every user of the internet is familiar with. The communications systems used with the early Samos satellites were severely limited in how much information they could transmit. lt took many minutes to transmit a single photograph.

Intelligence officials desperately wanted a near real-time reconnaissance satellite that could take pictures and relay them to the ground within a few minutes because it could be used to warn of immediate events, such as Soviet tanks about to roll into Czechoslovakia. But without a relay system, a satellite over the Soviet Union would have to travel a quarter of the way around the world or more before coming within range of an American ground station.

Wheelon’s original idea was to place the relay satellite in geosynchronous orbit above the equator. But the problem with this solution was that a single relay satellite would be insufficient to beam the data back to the United States and it would have to transmit it through another satellite or a ground station. The key was finding a way of sending the information through only a single relay satellite before sending it back to the ground.

According to Wheelon, Alexander Flax, the Director of the National Reconnaissance Office from 1965 until 1969 – after Wheelon had left the government – came up with a superior solution. Although it is not clear how Flax got the idea, it is possible that he had help from the Soviet Union [4].

In April 1965 the Soviet Union launched its first successful Molniya communications satellite [5]. The satellite was placed in a new orbit, which was soon named the Molniya orbit. lt was highly inclined to the equator, but also highly eccentric, meaning that it swung low over the Earth in the southern hemisphere before heading toward a distant apogee high over the northern hemisphere – like tossing a tennis ball high into the air, it would slow down on its way up and speed up on its way down, but appear to spend most of its time in a small area above one spot. A satellite in a 12-hour Molniya orbit would spend eight hours flying above the northern hemisphere, but only four hours flying over the southern hemisphere. From the satellite’s perch approaching apogee, it could see much of the northern hemisphere. A satellite in such an orbit could maintain a line of sight with both a reconnaissance satellite low over the Soviet Union and a ground station in the United States. The Soviet communications system required three satellites for a full day of coverage, but because the American reconnaissance system only operated during the day, only two SDS satellites would be necessary to support a reconnaissance satellite in low Earth orbit.


For the remainder of the 1960s the CIA sponsored the development of new technologies to make “near real-time” reconnaissance possible. Until the basic image creation and processing technologies were developed, there was not really a need for communications systems to relay the imagery to a distant ground station.

ln June 1971 President Richard Nixon approved the development of what was soon designated the KH-11 KENNAN reconnaissance satellite. The KH-11 was a big telescope with an image-forming electronic device at its focal point. The first KH-11s launched used light sensitive diodes, but the later ones used a linear CCD array [6]. Once the KH-11 was approved, a data relay system was also necessary.

The exact origins of the relay satellite are not currently known due to continued classification. Apparently in the early 1970s the U. S. Air Force developed a requirement for providing command and control information to nuclear forces in the Arctic where the propagation of radio waves is poor. The Air Force proposed a Data Relay Satellite System, or DRSS, which would transmit data to nuclear forces operating in the Arctic, such as Strategic Air Command B- 52s attacking the Soviet Union during a nuclear war, and also relay data from satellites over the horizon to a ground station. The early Defense Support Program missile warning satellites were proposed as candidates for this data relay mission [7].

At the same time, the CIA component of the National Reconnaissance Office required a relay satellite for its KH-11 data. This satellite, known as the Satellite Data System, or SDS, was undoubtedly approved sometime in late 1971 and it is possible that for a short period of time the American military was studying both DRSS and SDS simultaneously [8].

DRSS was formally cancelled in May 1973, but it seems likely that its missions were transferred over to the SDS by early 1972. DRSS had formally been designated “Program 313” and this designation was apparently then given to the SDS [9]. One mission that was not given to SDS was that of relaying DSP Missile warning satellite data. Instead, the Air Force continued to utilize an overseas ground station in Australia for several decades to perform the relay function.

The first payload added to SDS was an Air Force Satellite Communications System (AFSATCOM) UHF transponder used for communicating with strategic forces in northern regions. Another payload that was added to the satellite was known as the “Mission-22 packet.” It was essentially a high data rate communications system for connecting the Air Force Satellite Control Facility at Sunnyvale, in California, and its seven remote tracking locations situated around the globe [10].

One key question about the SDS concerned whether it would be managed by the intelligence community or the Air Force now that it included both highly classified and relatively unclassified missions. This decision was inextricably entwined with the complex way that the United States managed satellite reconnaissance. The National Reconnaissance Office, or NRO, was a Department of Defense agency headed by a civilian Air Force official. That official, the Undersecretary of the Air Force, had dual responsibilities. On intelligence satellite issues he answered to the Secretary of Defense as well as the Director of Central Intelligence. But he was also responsible for purely Air Force satellite issues as well, and on these questions he reported to the Secretary of the Air Force.

The NRO was composed of three offices known as “Programs.” Program A was an Air Force office based in Los Angeles along side unclassified Air Force space offices. Program B was the CIA office based in Washington, DC, and Program C was a small Navy office, also in Washington. The CIA’s Program B managed the development of the KH-l1, which was then being built by Lockheed.

Both the AFSATCOM and Mission-22 payloads were standard Air Force missions, not an intelligence community requirement. But once they were added to the satellite they created a question and an opportunity. Should the SDS development be managed by the Air Force Program A component of the NRO, or by the Air Force’s Space and Missile Systems Organization (SAMSO)? SAMSO was normally responsible for development of Air Force satellites and was located across the street from Program A. lf the satellite was developed by SAMSO, it would provide useful cover for the classified mission. An early plan for the near-real-time system was that it could be covert. Unlike previous reconnaissance satellites, the KH-11 would never have to drop recovery vehicles containing film, thereby giving away its mission. lf the data relay system had an effective cover story, then nobody would know about its link to the highly secret imaging satellite.

There was a drawback to giving the management of SDS to SAMSO, however. SAMSO and the NRO operated under different rules and regulations. The NRO’s Program A had a simplified chain of command and a “streamlined procurement system,” meaning that it did not have to abide by the same paperwork requirements as the rest of the Air Force. In contrast, SAMSO had to obey standard procurement regulations-meaning significant red tape — and all major decisions had to be approved by various committees and senior leaders.

The original schedule was to select a contractor to build the SDS by 1 March 1972. This selection was delayed until Deputy Secretary of Defense Kenneth Rush decided who would manage the program [11]. On 3 May Rush signed the Satellite Data System Management Plan, which placed the Space and Missile Systems Office in charge of SDS development and also described streamlined management, security and public release procedures to be used in running the program. According to a cover letter, the plan “also recognizes the potential need to deviate from or waive certain Department of Defense Directives and Instructions which conflict with the requirement for streamlined procurement. In other words, although the NRO’s Program A was not formally managing SDS, SAMSO could operate much like Program A. Although the details remain classified, presumably Program A managed the intelligence communications payload for SDS [12].

A source selection board selected a contractor in early April and then briefed senior Air Force officials, but the winner was not announced until 9 May 1972. The announcement was classified at the “confidential” level-the lowest level of security classification. The formal public announcement of the award was delayed until negotiations were completed and “the program objective was realigned and the public release policy established” [13].

In June 1972 Philip Klass wrote in Aviation Week & Space Technology that the Air Force was procuring a new class of data relay satellites that would “enable search-and-find type reconnaissance satellite photos” to be transmitted directly to the U.S. for speedy analysis and the transmission of commands back to the reconnaissance satellite to take close-look pictures of targets of opportunity discovered on the earlier photos.” Klass also quoted the USAF’s prepared statement for the service’s fiscal 1973 budget request indicating that it was developing a relay satellite. He reported that TRW and Hughes were competing for the satellite contract. In reality, Hughes had already won it. Nevertheless, his article probably unnerved members of the intelligence community who were not used to seeing their missions mentioned in publicly released statements, or the pages of Aviation Week [14].

The first launch was scheduled for January 1976, followed by a second launch in March, with $17.8 million allocated immediately, and $23 million for fiscal year 1973 [15].


On 5 June 1972 Hughes Aircraft Company was given a “Letter of Intent” from the Air Force indicating that the government would formally sign a contract with the company in the near future and which allowed Hughes to begin working on satellite development. On 7 July the Air Force publicly announced the contract, worth $36 million [16]. The initial plan was to launch the first two satellites in January and March 1976.

Almost nothing is known about the technical design of these satellites other than that they were cylindrical, weighed approximately 700 kg, and were apparently based upon Hughes’ commercial Intelsat lV satellite. The Intelsat lV which weighed approximately 730 kg, was derived from an experimental comsat developed for the Air Force in the late 1960s known as Tacsat and some reports at the time stated that SDS was based upon Tacsat (Fig.1). However, Hughes had a policy of designing its commercial comsats to military specifications and the differences between Tacsat and Intelsat were not major. lt seems likely that advances incorporated into Intelsat lV were also adopted for SDS. The satellite was drum-shaped and spun in orbit. At its top was an antenna “farm” or platform that was de-spun by an electric motor so that its antennas pointed at the Earth [17].TACSAT

Because the SDS satellite operated in a considerably different orbit than Intelsat lV, it undoubtedly had numerous differences from its progenitor. Probably the most significant difference was that the SDS did not require an apogee kick motor to place it in its final orbit. The Molniya orbit that SDS operated in was similar to the initial transfer orbit that geosynchronous comsats were placed in by their upper stages. But this was the final orbit for the SDS and no powerful rocket was necessary to circularize it. The apogee kick motor had considerable mass, and eliminating it from the SDS undoubtedly had a major impact on the satellite’s design, allowing it to carry extra mass for other purposes, such as station keeping fuel.

Another major change was that SDS operated in an orbit that repeatedly took it in and out of the Van Allen Radiation Belts, which regularly cooked its electronics. This required extra radiation hardening.

But undoubtedly the most important change in the satellite from its commercial predecessor was the communications package. The KH-11 KENNAN’s designers wanted to make the satellite as covert as possible and one means of accomplishing this was to prevent its transmissions over the Soviet Union from reaching the ground. The way they achieved this was to use an uplink transmitter that operated at 58 GHz, a frequency which was absorbed by the oxygen in Earth’s atmosphere and could not reach the ground. The KH-11 would transmit up to the SDS satellite at this frequency and the SDS would then retransmit the information down to a ground station at 22 GHz [18]. Because the SDS was already being used to transmit KH-11 imagery, which would have taken much bandwidth, it probably was also used for less bandwidth-intensive telemetry, command and control, relaying commands and satellite health data back and forth to a ground station. In addition, all of this information would have to be encrypted [19].


Although little is known about the technical design of the SDS satellite, more is known about its early administrative history. In the second half of 1972 the launch dates for the first two satellites were slipped to June and August 1976, with the system achieving operational availability by November. Funding was also realigned as better cost estimates became available. The Air Force had initially allocated $49.1 million for fiscal year 1974, but reduced this to $43 million and then to $40 million [20]. But $1.9 million was later added back to the budget [21]. However although these numbers were declassified, it is highly likely that additional money was also provided through classified channels to support the development of the communications package devoted to the intelligence mission.

The communications subsystems Critical Design Reviews (CDRs), an intensive review of all aspects of the communications payload, were conducted in November and December 1973 [22].

Hughes initially started work on two satellites. The first, designated X-1, was a structural model, designed to prove that the spacecraft structure was sound during launch. The second, Y-1, was the qualification model, which was equipped with most of the electronic systems and intended to demonstrate that the satellite could perform the functions it was designed for. The system CDR was accomplished in March 1974 [23].

The initial plan was apparently to procure four flight spacecraft plus to refurbish Y-1 to serve as a flight qualified spare if any of the production spacecraft was destroyed.

Early in 1974 the production schedule for the spacecraft was revised to accommodate delays experienced in the delivery of some equipment as well as increased manufacturing and testing costs. Delivery of the first flight vehicle was slipped from November 1975 to February l976 [24].

The Air Force also initiated a study to upgrade the satellites for Survivable Satellite Communications, known as SURVSATCOM. Although the specifics are unknown, this probably involved nuclear hardening of the electronics and other systems [25].

In the second half of 1974 refurbishment of the Y-1 spacecraft as a flight spare was slipped from fiscal year 1976 to 1977. The first flight spacecraft, designated F-1, was under construction, with F-2 starting construction and F-3 scheduled to start in fiscal year 1975 (26).

In the first half of 1975 testing of X-1 was completed, assembly of Y-1 was completed and testing was started, and fabrication and assembly of F-1 continued. During this period the budgets were increased slightly due to increased costs for the booster and some increases in the costs of F-2 and F-3 (27).

In August 1974 the Secretary of the Air Force approved a change in spacecraft to “augment” the polar coverage of the Atomic Energy Detection System which warned of nuclear explosions in the atmosphere. The F-3 and F-4 spacecraft were to be retrofitted with nuclear detection (or NUDET) devices known as “bhangmeters.” NUDET was already carried aboard DSP missile warning satellites by this time [28].

By late 1975 system level qualification testing on the Y-1 spacecraft was completed, with all critical specifications met or exceeded and the overall design validated. Assembly of the first flight spacecraft was also completed and final acceptance testing was started [29].

ln November of 1975 the Secretary of the Air Force for Research and Development approved a plan to procure two additional spacecraft, F-5 and F- 6 in fiscal 1978 and 1979 respectively. These spacecraft were to be modified to be compatible with Space Shuttle launch. In addition, their anti-jamming protection features and other aspects of their AFSATCOM payload were also to be improved. Funding also included Space Shuttle launch support through fiscal year 1982 [30].


The first two satellites were launched on schedule in June and August 1976 from Vandenberg Air Force Base atop Titan lllB Agena D launch vehicles. The first KH-11 KENNAN was launched in December 1976 and was operational by January 1977. Satellite F-3 was launched in August 1978. The fourth and fifth satellites were delivered in May and October 1980. Satellite Y-l, the configuration test model,was apparently redesignated as F-5A and delivered in May 1980 [31].

The launch dates of the satellites after F-3 are not accurately known because the National Reconnaissance Otfice launched a classified signals intelligence satellite known as JUMPSEAT into the same orbit. There were six classified American launches into Molniya orbit during the 1980s. In 1981 the Air Force proposed buying an additional SDS satellite, F-7. This meant that the Air Force had a total of four SDS satellites, plus the F-5A spare to launch during the 1980s, although it is not clear how many of those five satellites were actually launched during this period.

Two different sources have indicated that the early operations of the data-relay system were troubled. Apparently the KH-11 KENNAN satellites had problems with the Travelling Wave Tubes that generated the 58 GHz signal used to send data to the SDS satellites. The KH-11 TWTS burned out earlier than expected [32]. Another problem was that the SDS satellites were apparently not in their proper positions when the KH-11s were uplinking data. This ultimately resulted in the NRO taking over management of the SDS system by 1983. Apparently it was Program B, the CIA component of the NRO and not the Air Force component, that took over management authority from SAMSO.

Instead of procuring a new batch of identical satellites, in the early 1980s the Air Force chose to develop a new and more capable “SDS-B” class of satellites based upon the larger Intelsat Vl commercial satellites. Congress appropriated money for a “shuttle-compatible” SDS satellite in the fiscal 1984 budget. The first launch of this bigger satellite took place aboard the Space Shuttle Columbia on STS-28 in August 1989. Two satellites were placed in Molniya orbit, although a third was placed into geosynchronous orbit [33]. That class of satellites was probably replaced by a new class of satellites beginning in 2001. lt remains to be seen if further information about the early years of this program will be released by the National Reconnaissance Office, or whether they will remain partially obscured in the remaining shadows of the Cold War [34].SDStable


  1. “Albert D. Wheelon”, in Robert A. McDonald. ed., Beyond Expectations–Building an American National Reconnaissance Capability, American Society for Photogrammetry and Remote Sensing, 2003, p.333.
  2. Ibid.
  3. Lockheed Missile Systems Division [Slide], “Sentry Data Relay Concept”,’19 August 1958. [declassified in response to NRO FOIA request] The name Sentry was changed to Samos only a few months later.
  4. “Albert D. Wheelon,” Robert A. McDonald, ed., Beyond Expectations-Building an American National Reconnaissance Capability, American Society for Photogrammetry and Remote Sensing, 2003, p.333
  5. Bart Hendricks, “The Early Years of the Molniya Program.”Ouest 6, pp. 28-36, t998.
  1. Jeffrey T, Richelson, The Wizards of Langley, Westview Press, 2001, p.199.
  2. “DMS Market Intelligence Report, Satellite Data System,” Greenwich, CT,1979, p.1.
  3. See: Jeffrey T Richelson, “The Satellite Data System”, JBIS, 37, pp. 226-224, 1984.
  4. “DMS Market Intelligence Report, Satellite Data System,” Greenwich, CT,1979,p.1.
  5. Both the AFSATCOM and Mission-22 payloads were publicly acknowledged in congressional budget documents. See: Jeffrey T. Richelson, “The Satellite Data System”, JBIS, 37, pp.226-228,1984. This article contains many contemporary citations concerning the openly acknowledged missions.
  6. “Semi-Annual History of the Directorate of Space, 1 January 1972-30 June 1972”, [declassified excerpt], Air Force Historical Research Agency, Maxwell Air Force Base [hereafter AFHRA], pp.18-19.
  1. Kenneth Rush, Memorandum for the Assistant Secretary of Defense (Comptroller) et al, “Satellite Data System (SDS) Management Plan”, 3 May 1972. [declassified in response to FOIA request to the Department of Defense. The SDS Management Plan is currently under declassification review and has not been released as of July 2006.]
  2. “Semi-Annual History of the Directorate of Space, January 1972-30 June 1972”, [declassified excerpt], AFHRA, pp.19-20.
  3. Philip J. Klass, “USAF Plans Data Relay Satellite”, Aviation Week & Space Technology, l9 June 1972, p.12.
  4. “Semi-Annual History of the Directorate of Space, 1 January 1972-30 June 1972”, [declassified excerpt], AFHRA, pp.19-20.
  5. “Hughes to Develop USAF Data Relay Syslem”, Aviation Week & Space Technology, 17 July 1972, p.14.
  6. USAF Brigadier General Henry B. Stalling testified before the House Armed Services committee that the SDS would be based upon Tacsat. “Two Spacecraft Planned for USAF Satcom Syslem”, Aviation Week & Space Technology, 6 August 1973, p.16.
  7. Comments of a former intelligence official.
  8. These radio crosslinks had apparently been proposed for the Defense Support Program satellites in the 1980s, but the DSP program office selected a laser-based crosslink instead, which proved impossible to develop at that time. See: Jeffrey T.Richelson, Ameica’s Space Sentinels, University Press of Kansas, 1999, pp. 278-279, fn. 55.
  9. “Semi-Annual History of the Directorate of Space, 1 January 1972-30 June 1972”, [declassified excerpt], AFHRA, pp.19-20; the operational date is from “Two Spacecraft Planned for USAF Satcom Syslem”, Aviation Week & Space Technology, 6 August 1973, p.16.
  10. “Semi-Annual History of the Directorate of Space, 1 July 1972-31 December 1972”, [declassified excerpt], AFHRA. p.31.
  11. “Semi-Annual History of the Directorate of Space, July 1973-31 December 1973”, [declassified excerpt], AFHRA, p.41.
  12. “Semi-Annual History of the Directorate of Space, 1 January 1974-30 June 1974”, [declassified excerpt], AFHRA, p.40.
  13. lbid.. p.41.
  14. lbid., p.42.
  15. “Semi-Annual History of the Directorate of Space, 1 July 1974-3’l December 1974”, [declassified excerpt], AFHRA. p. 30.
  16. “Semi-Annual History of the Directorate of Space, 1 January 1975-30June ‘1975”, [declassified excerpt], AFHRA, p.37.
  17. Ibid., p.38.
  18. “Semi-Annual History of the Directorate of Space, 1 July 1975-31 December 1975”, [declassified excerpt], AFHRA. p.37.
  19. lbid., pp,37-38. One document lists as a “milestone” the upgrade of vehicle #6 in September 1978, but this was long before the delivery of the fourth vehicle.
  20. The launch dates for the first three satellites are from: “Satellite Data System (SDS)”, C3l Program Management Structure and Major Programs, USDR&E (ASD/C3l), 10 December1980.
  21. The KH-11s had another problem as well. Their Control Moment Gyros, which were used not only to point the satellites but to slew the satellite in order to remove image smear from the photographs, burned out earlier than expected. Comments of a former intelligence official.

33, Dwayne A. Day, “Out of the Shadows: The Shuttle’s Secret Payloads”, Spacellight, 41, pp 78-84, 1999.

  1. Craig Covault, “Launch Surge Begins for Secret NRO Missions”. Aviation Week & Space Technology, 10 September 2001, p.43; Craig Covault, “Secret Relay Satetlite Launched as USAF Weighs Surge Options,” Aviation Week & Space Technology, 15 October 2001, p.44.



This entry was posted in Projects/Missions by Jack Fisher. Bookmark the permalink.

About Jack Fisher

Jack was a systems engineer at Hughes from 1961 to 1992. He contributed to various programs including Surveyor, Pioneer Venus, Galileo, Intelsat VI and innumerable proposals. He was the manager of of the Spacecraft Systems Engineering Lab until his retirement. Upon retirement Jack taught systems engineering at a number of national and international venues.

1 thought on “RELAY IN THE SKY: THE SATELLITE DATA SYSTEM DWAYNE A. DAY Space Studies Board, 500 Fifth Street, NW, Washington, DC 20001, USA

Comments are closed.