Dr. Marc Rayman
December 17, 2007
Dawn is climbing away from the Sun on a blue-green pillar of xenon ions as it begins a new chapter in its mission. After the remarkably successful initial checkout phase, the project is now in the interplanetary cruise phase.
When last we visited Dawn, it had superbly demonstrated that it was ready to fly the way it will for most of the mission. Although there will be many special activities during its journey to Vesta and from there to Ceres (and be sure to visit this site again to be among the best informed on your planet of what Dawn is doing), the probe will spend most of its time doing what it is doing today: patiently reshaping its orbit around the Sun with its amazingly efficient ion propulsion system.
Without taking any more time than was needed for some high fives (and a few high sixes for the more mathematically avant-garde) following the flawless execution of the test of long-term thrusting, the team turned its attention to more checkout activities. Ion thruster #2 was the focus of tests on November 13 - 15, and the excellent results matched those of thrusters #1 and #3. (The identities and locations of the thrusters were revealed in a previous log before such information could be publicized by more sensational, vulgar sources.)
November 19 was the first in-flight use of the spacecraft’s main antenna. (In fact, that antenna was used in some of the thruster #2 tests, in which the spacecraft was oriented so the antenna cast a shadow on the thruster. This permitted operation at the highest throttle level without overheating while still relatively close to the Sun. We recommend even to official Dawn scorekeepers that that not count as the antenna’s first use.) Because of the celestial geometry earlier in the mission, pointing the main antenna to Earth would have exposed sensitive components elsewhere on the craft to too much solar heating. All communications prior to the use of the main antenna relied on 1 of the 3 smaller antennas that do not emit as tight a radio beam. As the distance to the spacecraft increases, the smaller antennas will allow only very slow communications, while the main antenna, with a diameter of 1.52 meters (5 feet), will permit the return of many pictures and other scientific data even from distant Vesta and Ceres.
Much of the rest of the initial checkout phase was dedicated to updating software on the spacecraft. Many units onboard, both in engineering subsystems and in science instruments, incorporate their own computers, but the command and data handling subsystem contains the master computers. One main computer runs operations aboard the ship, and another computer supports it (and has a few other responsibilities of its own). Each of these computers has an identical backup, able to take over should the primary unit experience problems. With a planned 8-year mission in the forbidding environment of deep space, Dawn may face challenges that require more than such backup hardware. So each main and each support computer holds both a primary and a backup copy of the software. If radiation or some other problem corrupts one version of the software, the system can detect that and resort to another.
Engineers recognized well before launch that new software would need to be loaded during the initial checkout phase. As extensive as ground testing was, the team anticipated that the need for some updates would be identified once Dawn was operating in space. In addition, during the last few months before launch, when ongoing testing ferreted out bugs, only those changes that were essential for the beginning of the mission were made. Modifying complicated software is – well, complicated; and seemingly simple changes can have unintended consequences. To allow thorough testing of the spacecraft’s capability to complete the complex and critical steps after separating from its Delta rocket, as described on September 21, late prelaunch changes to the software were kept to a minimum, and an improved version was planned for November.
Following the popular trend of giving software a snazzy name, the project denoted the latest suite “flight software 7.0.” We fully expect this to inspire new toys, movies, fashions, and even lifestyles (particularly among readers in the Pleiades), and Dawn’s marketing department is standing by to work with you.
On November 20, the new software for the support computers was radioed from Earth and installed separately on the primary and backup units in the command and data handling system. The software for each computer requires 67 files occupying about 135 kilobytes with a total of 22,800 lines written in the programming language C and in assembly code. The backup copies of the 7.0 software were transmitted to both computers on December 6. Each of these activities required great care, verifying that the computer memory remained healthy throughout, that no bits were lost or altered in transmission or storage, and that if an unrelated problem arose on the spacecraft during the process, the computers would be able to handle it, never being left in a vulnerable state. Each step was tested and verified repeatedly with the indispensable Dawn spacecraft simulator at JPL (down the hall from mission control and around the corner from the very popular Dawn ice cream freezer and the less popular Dawn fruit bowl).
Loading software into the main computer was still more complex than doing so in a support computer. To start running the new software, the computer would have to be rebooted. While that is a familiar and straightforward procedure for most terrestrial readers in the early 21st century, it is considerably more complicated when the computer is in control of a spacecraft in flight.
On November 26 and 27, the main computer’s memory was checked, its health was verified, and the updated software was sent from mission control to the Deep Space Network stations in California, Spain, and Australia (the work took long enough that all 3 communications complexes were required), and then 17 million kilometers (11 million miles) to the main computer. The 2.6 megabytes of 7.0 software for that computer consists of 591 files totaling more than 410,000 lines of C and assembly code. After loading all the software, operators began preparing the spacecraft for rebooting, scheduled for November 28.
Whenever the main computer reboots, it commands the spacecraft into “safe mode.” Other conditions can trigger this mode as well, including separation from the rocket on September 27. The probe does not know whether safe mode was planned, as it would be with the installation of new software, or was called in the process of dealing with a problem when ground controllers were not available to intervene.
Most of the characteristics of safe mode remain as they were on launch day, but there are some differences now. Dawn no longer has to wait for xenon to stop spinning, and it does not have to deploy its solar arrays. It still points one face to the Sun, the only easily identifiable spatial reference throughout its flight in the solar system. For the first few weeks of the mission, the relative locations of the Sun, Earth, and spacecraft required Dawn to sweep its safe mode radio signal like a searchlight that would periodically illuminate Earth, as explained before launch. For the rest of the mission, from Dawn’s vantage point well outside Earth’s orbit around the Sun, the planet and star always will appear close enough together that the spacecraft can use the broad beam of an antenna that points at the Sun in safe mode, instead of an antenna at right angles to it. While extremely weak (as we will see in a log early next year), the faint radio whisper that would reach Earth would still be loud enough to be heard. In October, controllers modified safe mode so it would employ this other antenna.
After the reboot preparations were completed on November 27, most team members went home to get some rest for the activities scheduled for the subsequent 2 days. After they commanded the reboot and Dawn established itself in safe mode, the Deep Space Network would need to capture the radio signal, they would have to verify that the software was operating correctly, and then the long process of bringing it out of safe mode to normal operations would begin. Exquisitely detailed plans for each step had been formulated with great care, but when the team left for the day, they did not know that a single surprise lay in store.
Around 10:00 pm PST on November 27, the spacecraft’s main computer rebooted, just half a day before operators planned to command it, and Dawn entered safe mode. As soon as the Deep Space Network detected the corresponding change in the radio signal and the small night-shift team in mission control realized what had happened, a different plan, known dryly as “anomaly response,” was put into action. Some team members were called back in to JPL and spent the entire night investigating this unexpected event; some others were in occasional contact by telephone or Internet. As many team members as possible were not disturbed, so they would be fresh the next day to pick up after the anomaly team’s overnight work. (In the same vein, although your correspondent was among those who went to JPL, he opted not to contact you right away; rather, he chose to let you read about it now, under more leisurely circumstances.)
The team first used the trickle of data from the distant probe to verify that it was indeed safe. Then, proceeding with great diligence, given the unanticipated behavior of the computer, they downloaded certain diagnostic files. All indications were that Dawn was quite healthy, with no apparent signatures of a lingering problem. By the middle of the day on November 28, engineers had determined that the new software (automatically loaded when the computer rebooted) was doing well and the spacecraft was ready to resume its work. That day and the next had been planned already for the time-consuming transition from safe mode to normal configuration, so most of the team followed those plans while others continued analyzing why the reboot had occurred. On November 29, all steps were complete, just as they would have been had the reboot been commanded the previous morning.
Even now the investigation into the unplanned reboot continues with simulators. Meanwhile, the mission has progressed very smoothly. The backup main computer received its primary copy of 7.0 on December 7 and its backup copy on December 14. The backup copy of the primary main computer’s software will be loaded in January during a hiatus in thrusting.
While flight software 7.0 had been tested extensively before being sent to the spacecraft, ever-prudent mission controllers had planned to conduct one additional test, this time on the spacecraft. Because Dawn will spend most of its life thrusting with the ion propulsion system, engineers wanted to verify that the new software did not introduce any bugs that would interfere with this essential capability. Under the guidance of the main computer, all systems operated well during the “thrusting validation” on November 30.
In the subsequent 2 weeks, as they were finalizing plans for the beginning of interplanetary cruise, the team conducted another round of instrument tests. Following the excellent results of the first set of tests in October, this month the science camera and the visible and infrared mapping spectrometer were pointed at specific targets to allow more thorough characterizations of their functions and capabilities. Observations included Saturn (too distant to appear as more than a bright spot to Dawn’s camera, but still useful for tests), Arcturus, Vega (of special significance to your correspondent and his wife), and other stars and star fields. While the primary science camera was operated for the first time in October, the backup had its first in-flight exercises on December 10 and 12.
As the instrument and outreach teams find time, more views from the cameras and spectrometers will be posted at http://dawn.jpl.nasa.gov/multimedia/spacecraft.asp. (Certain ones may take longer, while 3-way negotiations drag on among the Dawn project, as-yet unnamed celebrities in the imaged star systems, and tabloids in those regions.)
Several lovely scenes have been captured. While the brief encounter with Mars is still 14 months away, arrival at Vesta is 2.5 years after that, and rendezvous with Ceres takes place in 2015, the excellent instrument tests whet our appetites for what will be revealed. Readers are particularly recommended to see the image of a star field in Cepheus as well as the Eta Carinae Nebula (known to some readers as NGC 3372 and to others as “home”), captured in a calibration image of the fine cluster of bright stars NGC 3532, a popular sight for observers in Earth’s southern hemisphere. Eta Carinae is a massive and unstable star in the gaseous star-forming nebula that bears its name, and it has displayed a highly variable brightness since it was cataloged 330 years ago. For a time in the 19th century, it outshone all but one of the stars in Earth’s nighttime sky, despite being much much farther away than most. It has faded and brightened several times since then, sometimes being too faint for naked-eye observation.
Dawn’s initial checkout phase was remarkably productive and has served extremely well to certify the systems on the spacecraft and in mission control for interplanetary cruise. On December 14, in addition to loading the backup copy of 7.0 into the backup main computer, controllers radioed to the probe all the instructions and data it would need for the first 37 days of the next mission phase. The files were stored for use beginning on December 17.
At 12:01 pm PST on December 17, Dawn obediently began executing the instructions to reconfigure in preparation for long-term cruise. Three hours later, it initiated the prestart sequence for the ion propulsion system. While preparing for thrusting, the spacecraft also turned to aim ion thruster #3 in the required direction. This took the main antenna away from Earth-point, but the commands directed the craft to switch to one of the smaller antennas with broader coverage, allowing the team at JPL to monitor progress using signals received at the Deep Space Network complex in Spain.
After the spacecraft’s fine performance during initial checkout, engineers observing the beginning of thrusting expected no less. During initial checkout, the ion thrusters were started a total of 16 times and accumulated about 11 days 14 hours of thrust. (Some of the thrusting was for tests of the ion propulsion subsystem, and some was for tests of other subsystems or the entire Dawn system while thrusting.) With this experience, there was little reason to be concerned, but prudence dictated verification that Dawn got underway smoothly.
Telemetry confirmed that thrust began on schedule at 4:08 pm, as Dawn began propelling itself deeper into space, farther from Earth and the Sun. When the spacecraft turned its main antenna away from home, it set its sights elsewhere, on uncharted worlds, as it embarks on the next phase of its extraterrestrial expedition.
Dawn is 27 million kilometers (17 million miles) from Earth or 70 times as far as the moon. Radio signals, traveling at the universal limit of the speed of light, take 3 minutes to make the round trip.