Ch5-2

Computers in Spaceflight: The NASA Experience

- Chapter Five -
- From Sequencers to Computers: Exploring the Moon and the Inner Planets -
Fixed sequencers: "Computers" on Ranger, Surveyor and the early Mariners

[141] Whether the final mission destination is as close as the moon or as far as Neptune, probe spaceflights consist of the same milestones and activities: launch, mid-course maneuver, cruise, and encounter. Spacecraft are launched in a stowed position dictated by the geometry of the booster vehicle. Most space probes look like multiarmed Hindu gods in flight due to the need to expose solar panels, point antennas, and deploy imaging equipment, but they must be folded to fit into the nose fairing of a rocket. During the launch period the spacecraft is injected into its transfer orbit to intercept the target, deploys its various appendages into their proper positions, and orients itself. A decision was made early at JPL to build spacecraft that would be stabilized in three axes during flight¹. Spacecraft would be oriented by using the sun, earth, and/or a star as a reference. If kept from tumbling they would always be pointed in a specific direction. A key advantage of this plan is that a directional antenna could be used for earth-space communications, reducing power requirements. Imaging equipment could also be more stable than on a spin-stabilized spacecraft such as a Pioneer. A disadvantage of three-axis stabilization is that a fairly sophisticated attitude control system must be carried, including a sensor system to find the sun and a guide star. Part of the launch phase, then, is spent scanning the sky for Campus, Vega, or whatever star has been chosen for aligning the spacecraft.
The mid-course maneuver phase often comes only a day or two after initial transfer orbit insertion in order to correct relatively large [142] injection errors. Consisting of a timed burn of the spacecraft's propulsion system in each of three axes, it serves a number of purposes. Early launches could not depend upon the launch vehicle to establish a adequate flight path. Later, as booster guidance improved, probes were purposely aimed to miss the target so as to avoid contaminating planetary atmospheres with earthly bacteria hitching a ride on a spacecraft if the spacecraft ceased to function during launch and could not change its path to miss the planet. Therefore, the mid-course burn took place to correct the path of a "live" spacecraft. On long-duration missions with several targets, such as the Voyager probe to Jupiter, Saturn, Uranus, and Neptune, this maneuver might be repeated before and after each encounter. Engine firings are made before encounter to improve the accuracy of the trajectory to achieve a better gravity assist from the target planet to the new trajectory and reduce the size of the post encounter maneuvers.

Less is done on the spacecraft during the cruise period than in any other mission phase. However, recent larger and more complicated spacecraft have particle and fields experiments that run constantly and engineering calibrations that need periodic attention. If the spacecraft attitude is disturbed, reorientation may be necessary. This period of relative quiet ends when the encounter sequences begin as the spacecraft nears its target. Instruments must be turned on, calibrated and aimed. Imaging instrument pointing must be programmed and controlled. Data must be recorded and transmitted to earth. Of course, these activities are repeated during multiple encounter missions.

Initiating the functions done in each phase requires on-board control. This was unnecessary for Ranger missions to the moon, which were simple impact flights with televised imaging during the last minutes. Because maximum speed-of-light delay in radio signals to the moon is less than a second, near-real-time commanding could be done. Ground commands could fire engines, point the spacecraft, and turn on cameras. Ranger flights used a voice/manual commanding system for this. Desired instructions were developed and formatted at JPL and then delivered by telephone to the Deep Space Network station currently in contact with the spacecraft. An operator would thumb-wheel the octal codes into a panel called the "Read-Write-Verify Console," sending them to the spacecraft after verification². Such care was not always enough. On Ranger III, a guidance error caused the spacecraft to miss the moon by 23,000 miles. Although JPL flight controllers were able to get images during the flyby, a documentation discrepancy between the command set developed during the ground testing of the spacecraft and the flight set caused Ranger to point the wrong way, returning images of open space³.

Ranger carried a "Central Computer and Sequencer" to back up the direct command system. Activated before lift-off, it counted the hours, minutes, and seconds until a specified mission event was to [143] occur and then executed a set of commands that performed the required functions. If the uplink radio channel failed, the mission would proceed according to a prepared plan. This assumed optimum performance, turning on the cameras regardless of where the spacecraft might be actually pointing. Still, it provided a bit of insurance for the mission.

At the same time that the Rangers were being built, JPL designed and flew the first Mariners. Mariner's initial mission was a Venus flyby launched in 1962. In the case of this spacecraft and its later brethren, the Central Computer and Sequencer was the prime source of commands, at least for cruise and encounter portions of the mission⁴. The time delay for commands to travel to Venus and Mars defeats real-time control from the ground. For Mariner II, at launch time minus 15 minutes, the clock was set so that the encounter sequence would begin at 12 hours from the closest approach to Venus. The sequencer's clock, a very accurate oscillator similar to computer clocks today, started at launch time minus 3 minutes⁵. Direct commanding capability was maintained. When the star tracker got confused and locked onto the wrong target, ground controllers could reinitiate a search⁶. Direct command could also be used for midcourse maneuvers. As a complement to direct command, "quantitative" commands could be sent to the sequencer for later use⁷. For instance, times such as "51 seconds of minus roll" and "795 seconds of minus pitch" or burn times could be inserted into the memory for later execution⁸. Mariners could abandon direct command and go to automatic command if a radio failure was detected. On the Mariner Mars 1964 spacecraft the sequencer contained a cyclic command that checked for such a failure at 66 2/3 hour intervals, effecting an auto switch-over⁹.

The Mariner II spacecraft to Venus (1962), Mariner IV to Mars (1964), and Mariner V to Venus (1967) carried the same Central Computer and Sequencer. Just one flew on each mission, due to space and weight restrictions, even though the machine weighed in at 11.5 pounds¹⁰. However, with the direct command capability intact, each had essentially the same level of redundancy as the Gemini and Apollo spacecraft, with their single-processor on-board computer systems and ground control computers. Plans for Mariner Mars 1969 called for a larger spacecraft and a more ambitious mission: two picture-taking flybys of different portions of the "red planet". JPL's Neil H. Herman, who had headed development of the Sequencer, saw an opportunity to improve the device for the upcoming flights¹¹. One aim was to give the new spacecraft more flexibility. If the first flyby turned up something special, it would be very useful if the second spacecraft could be reprogrammed in flight to take advantage of lessons learned on the initial pass¹². This actually happened during the missions when reprogramming was accomplished for Mariner VII's [144] August 5, 1969 flyby in response to Mariner VI's July 31 passage¹³. Another reason for more on board autonomy is that command sessions for the Mariners lasted as long as 8 hours! Mariner's command rate was l bit per second, so long sequences were expensive both in personnel time and Deep Space Network time¹⁴. The availability of more space and weight plus the desire for flexibility and greater autonomy caused JPL to change the Sequencer to make it more of a computer and less of what it really was, a fixed-program counter.



Figure 5-1. Mariner Mars 1971 carried a programmable sequencer with an expanded memory. (JPL photo P12035)