Europe’s Galileo positioning, navigation and timing program has suffered outages of nine atomic clocks and is uncertain of the cause. The problem has not resulted in a satellite failure as each satellite has four clocks, two each using passive hydrogen maser and rubidium technology developed by SpectraTime of Switzerland. Even so, ESA and Galileo’s owner, the European Commission, have agreed to delay Galileo’s next launch by 3-4 months, to late 2017, to give the inquiry more time to get to the bottom of the issue. Shown here is a Galileo satellite payload during integration. The maser clocks are at left, and the smaller rubidium clocks are at the center. Credit: SSTL
PARIS — European government officials have postponed the next four-satellite launch of their Galileo positioning, navigation and timing network by about three months to investigate atomic-clock failures on in-orbit satellites.
The launch, which had been set for August-September, has been rescheduled for November or December aboard a European Ariane 5 rocket.
As of mid-January, 10 Galileo atomic clocks, the key element in a navigation satellite, had failed on four different satellites. One was recently returned to service, leaving nine outages. With each satellite carrying four clocks, none of the failures has led to a satellite’s being removed from service.
Complicating the investigation has been that the failures have occurred on two satellite platform designs, one built by Airbus and Thales Alenia Space as part of a four-satellite system validation program; and the other by Galileo prime contractor OHB SE of Germany.
Both types of clock technology affected
What is more, the failures have been distributed between both kinds of clocks carried on board Galileo. Seven have occurred on hydrogen maser clocks, and three on rubidium clocks. Each satellite carries two of each kind.
One satellite has suffered two outages, meaning it still has a backup capacity in the event of a third failure, European Space Agency Director-General Jan Woerner said Jan. 18.
The common thread among them is their prime contractor, SpectraTime of Switzerland, which is the sole provider of atomic clocks for Galileo and has also provided the technology for the satellite navigation programs of India and China.
But even here the situation is not clear. Javier Benedicto, head of the Galileo Program Department at the 22-nation ESA, said the investigation has tentatively concluded that the issues do not relate to the core technology provided by SpectraTime.
In an interview, Benedicto said that while the root cause of the failure likely is to be found inside the clocks, it is more likely to be of a peripheral piece of circuitry or other unremarkable component that, when operated in a certain way, leads to failure.
“We have 72 clocks in orbit today,” Benedicto said. “Of these, we have had up to six failures. We have not been able to replicate those failures on the ground. These units all passed their qualification and acceptance tests.”
Benedicto said ESA, which is the technical manager of the Galileo program, which is owned by the 28-nation European Commission, has redoubled its supervision of the most recent clocks, including the 16 launched in November on four satellites.
The investigation’s early recommendation, in addition to an intense focus on clock components and ground test results for future launches, is to modify the in-orbit operating procedures.
“It is not a development issue with the clocks,” Benedicto said. “It is not a fundamental part of the unit that is failing.They were not bound to fail. But in certain conditions where you have certain parameters going in the wrong direction together, it could cause the unit to fail. You have to be particularly unlucky and have all the failure parameters drifting in a specific direction.”
Like the navigation constellations operated by the United States, Russia and China, Europe’s Galileo operates in medium-Earth orbit — in Galileo’s case, at 22,300 kilometers in altitude.
Ground controllers retain some flexibility in the temperature and voltage used to maintain the satellites based on what they suspect are risk factors.
India and China use Europe-built atomic clocks
Benedicto said Galileo officials have compared notes with managers of the Indian Regional Navigation Satellite System, which uses SpectraTime clocks, but that the issues confronted on the Indian system do not appear to shed much light on the Galileo experience.
Woerner said ESA has been in contact with India about the issue. He did not directly respond when asked whether similar discussions have occurred with China. Woerner said the Indian system has also had atomic-clock issues but that they appeared to be unrelated to the occurrences on Galileo satellites.
One European government official said managers of China’s Beidou navigation system, which before developing a domestic supplier also used SpectraTime technology, have been less willing to talk about any issues they have faced.
The four global navigation systems — the U.S. GPS, Russia’s Glonass, China’s Beidou and Galileo — are part of a United Nations Office of Outer Space Affairs group called the ICG Providers’ Forum. India and Japan, whose systems are regional, are also members.
Benedicto expressed some hope that the ICG might develop a common database to which all navigation system providers provide regular updates on operational issues without compromising proprietary information.
Several of the largest commercial telecommunications satellite fleet operators have established the Space Data Association to share, on a confidential basis, data on satellite flight safety and signal integrity. But the organization continues to battle concerns from some operators about submitting competitively valuable information.
“We have evidence that [the other system operators] have been confronted with similar problems and we all end up doing the same things — playing with the temperatures, the currents and voltages and so forth, and then cross-checking the telemetry and analyzing performance trends,” Benedicto said.
It also appears as though the failures are more likely to occur early in the satellite’s in-orbit life, or not at all, although the data set available so far is too small to confirm this.
The decision to launch four Galileo satellites in November was made only after an in-depth examination of their clock systems.
OK’ing November launch, delaying mid-2017 launch
With the European Commission eager to declare initial Galileo service before the end of 2016, there was pressure not to delay the launch. But Benedicto said ESA would have delayed the launch if it had any reason to believe the clocks were defective. The fact that each satellite has four clocks and needs only one also argued in favor of proceeding with the campaign.
“We have no evidence of anything wrong with these clocks and we would not know what to do with them on the ground,” Benedicto said, referring to the multiple prelaunch checks that were conducted.
“Of course there’s pressure to launch. But at ESA our point of view is simple:K If we are certain something is wrong in the equipment, we are not going to launch. We are going to repair even if it takes months, and then launch after. This is what we always do.
“On technical quality ESA does not make compromises. It is our reputation on that is at stake. We will not let satellites fly when we know something is wrong. We’re all accountable to our inspector-general and our DG [director-general] and that’s the way we operate.”
But with initial Galileo service declared following the November launch, program managers decided to delay the next launch by three months.
“We said: We clearly have a problem and a key resource we have to handle it with is time,” Benedicto said. “We need time to investigate and correct, and we want to implement a number of corrective procedures that have been recommended by [SpectraTime]. This will take a bit more time but our target is still to launch by the end of the year.”
Woerner said ESA and the European Commission are still debating whether to launch the next four satellites late this year or wait until a root cause has been found for the atomic clock problem.