Problems with Galileo navigation system

[mr friendly guy]

http://www.bbc.com/news/science-environment-38664225

Galileo satellites experiencing multiple clock failures
By Jonathan Amos
BBC Science Correspondent
18 January 2017

The onboard atomic clocks that drive the satellite-navigation signals on Europe's Galileo network have been failing at an alarming rate.
Across the 18 satellites now in orbit, nine clocks have stopped operating.
Three are traditional rubidium devices; six are the more precise hydrogen maser instruments that were designed to give Galileo superior performance to the American GPS network.
Galileo was declared up and running in December.
However, it is still short of the number of satellites considered to represent a fully functioning constellation, and a decision must now be made about whether to suspend the launch of further spacecraft while the issue is investigated.
Prof Jan Woerner, the director general of the European Space Agency (Esa), told a meeting with reporters: "Everybody is raising this question: should we postpone the next launch until we find the root cause, or should we launch?
"You can give both answers at the same time. You can say we wait until we find the solution but that means if more clocks fail we will reduce the capability of Galileo. But if we launch we will at least maintain if not increase the [capability], but we may then take the risk that a systematic problem is not considered. We are right now in this discussion about what to do."
Still operational
Each Galileo satellite carries two rubidium and two hydrogen maser clocks. The multiple installation enables a satellite to keep working after an initial failure.
All 18 spacecraft currently in space continue to operate, but one of them is now down to just two clocks.
Most of the maser failures (5) have occurred on the satellites that were originally sent into orbit to validate the system, whereas all three rubidium stoppages are on the spacecraft that were subsequently launched to fill out the network.
Esa staff at its technical centre, ESTEC, in the Netherlands are trying to isolate the cause the of failures - with the assistance of the clock (Spectratime of Switzerland) and satellite manufacturers (Airbus and Thales Alenia Space; OHB and SSTL). It is understood engineers have managed to restart another hydrogen clock that had stopped.
Esa is also in contact with the Indian space agency which is using the same clocks in its sat-nav system. So far, the Indians have not experienced the same failures.
Mitigating actions
A statement issued by the agency late on Wednesday gave additional details.
It appears the rubidium failures "all seem to have a consistent signature, linked to probable short circuits, and possibly a particular test procedure performed on the ground".
The maser clock failures are said to be better understood, with two likely causes, the second of which has caused most grief.
The Esa statement said this second scenario was "related to the fact that when some healthy [hydrogen maser] clocks are turned off for long periods, they do not restart due to a change in clock characteristics".
Actions are being taken to try to prevent further problems. These involve changing the way clocks are operated in orbit. Clocks about to fly are also likely to be refurbished, and future devices yet to be made will have design changes, the agency says.
Esa is hopeful it can still launch the next four satellites in the constellation before the end of the year.

Precise timing is at the core of all satellite-navigation systems.
Atomic clocks generate the time code that is continuously transmitted to users on the ground to help them fix a position.
The passive hydrogen maser clocks in Galileo are determined to be accurate to one billionth of a second per day, or one second in three million years.
This performance ought to contribute to giving users fixes that have errors of a metre or less - significantly better than the standard open service from GPS.
A fully operational Galileo system is regarded as a constellation of 24, split across three orbital planes in the sky. But spares are required also, and with one very early satellite in the constellation already considered very close to complete failure - for different reasons - there needs to be near-continuous production of spacecraft.
The four latest satellites went up as a quartet in November; more are set to follow later this year.
Galileo is a project of the European Commission, the EU's executive branch. The EC employs Esa as its technical and procurement agent.
The development path to a "European GPS" has been a tortuous one. The project is years late, and the completion cost - expected to be some 7bn euros by 2020 - is substantially higher than that originally foreseen by EU member states.
line break
Galileo's atomic clocks by the numbers
First four satellites launched were called In Orbit Validation (IOV) platforms
The next 14 were referred to as Full Operational Constellation (FOC) satellites
Three of the rubidium clock failures have occurred on Galileo's FOC satellites
Five of the hydrogen maser failures have occurred on the IOV spacecraft
One maser has stopped on an FOC satellite, giving nine failures in total
Three of the four IOVs are affected; two of the 14 FOC satellites are affected
Every satellite has two hydrogen maser clocks and two rubidium clocks
That means a total of 72 atomic clocks are currently in orbit
All Galileo satellites presently have at least two working clocks

I am no expert, but I haven't heard of this problem affecting GPS or the Russian global navigation system nor the Chinese one. What is going on?

[Sea Skimmer]

If you want random speculation, since if anyone knew the truth they'd have said it, the two forms of clocks share nothing in common except the satellite databus and the power distribution system. Problems with both can usually be diagnosed from the ground, if not fixed, but often that requires basically reverse engineering the the problem to make sense of the data. All satellites that aren't pure experiments use full scale full feature ground test objects for the R&D phase that are never meant to be orbited (big reason why satellites cost so much, they might build several different ground test versions first) and those are kept in storage in case of problems like this so they can be experimented on. They also serve as a way to validate software patches.

From what the ESA has said though in other articles I read it sounds like the problem is thought to be directly with the clocks themselves, and they may simply not have been sufficiently conditioned before testing, or that the conditioning may have caused damage in conjunction with the ground tersting. Space is a cold harsh place if your going to be in very high orbit for years and years. To give an idea of what harsh means, a lifetime radiation exposure limit of 1 million rems is not unheard of for spacecraft like this. You can't just plug an item into a lab bench at 60F and 14.7psi and say it will work in space. This is a big deal with certain aircraft, missile/torpedo and engine parts too. They must be tested like they will actually be expected to function, and even with a good design some products off the assembly line are going to fail testing. And if none fail testing....then the argument is you could have designed a lighter product in the first place (much math will be employed arguing over the economics involved in this, nav sats are in very high and thus expensive orbits). So you have to condition, and then test, and test for extremes, and NOT damage the damn thing in the process. The focus is on an error on that process. A pity, since if testing damaged stuff then every clock may be damaged, and nothing can be done about it from the ground.

But also keep in mind, since one form of clock only failed on the early semi test satellites, and the other on the later models so this could be two completely different problems.

Meanwhile this program was the boondoggle from the start with all the problems joint European led programs always have (and people wonder why the EU needs reform!?), and that means problems like FOD contamination or even more direct errors in assembly are just inherently more likely then on a single nation program where much closer control can realistically be enforced on suppliers. Errors like this can usually be deduced from orbital data given sufficient time to implement more or less backwards engineering of orbital sensor data but it can be very hard to do this.

Lets remember when Galileo got going Europe was deluded enough to think they could get commercial companies to pay for it! After about six or seven years that fell apart because it was blatantly stupid, given that the US was already giving away GPS for free, and that GPS receiver costs (for low end ones that only listen to one satellite at a time) fell below 100 USD by the end of the 1990s. Today its approaching the point you could put GPS in a sticker. It then became a fully EU funded program instead with all the need to hand out contracts to doners that involved. For bonus they began building the hardware without firm commitments to actually fund the complete system! Is it any wonder something might break?

As far as the long term implications no satellite has yet failed completely, and the completely funded program as it stands now includes three spare satellites, the GPS system also has in orbit spares, so while this may require delaying future launches its probably not going to be crippling unless they did in fact damage every clock. But it almost certainly will mean that the satellites wont exceed their rated lifespans, some GPS satellites have lasted over three times design life. That's not even that uncommon with spacecraft. If they work, they often work for much longer then they were tested to work, until the batteries get so weak they can no longer function on the dark side of the planet and kind of just wither away rather then overtly breaking down. If you orbited crap though, no helping it except to learn and move on.

As far as GPS goes I think one or two early satellites did prematurely die back in the 1990s, and they've had some random software problems that degraded accuracy, but most of that was with the all important ground control system. Very few satellites are capable of functioning without ground control, and navigational satellite vitally need a ground control central to keep them all coordinated.

GLONASS is a joke in comparison, it works sure, but only by spam. The design lifespan of the early satellites was only 3 years and none of the earlier ones lasted more then 6 years. I dunno on the current series but I doubt it much better. But this was just a reality of all Soviet era and 90s Russian spacecraft design, they were never meant for long lives and Russian was incapable of building better. The only saving grace was the Soviets also has some pretty cheap space boosters which helps mitigate the cost somewhat. Still this is not good engineering, in the cold war even those nuclear reactor powered Soviet recon satellites only lasted around 100 days in orbit, crap like that is exactly why they went bankrupt.

[K. A. Pital]

Still this is not good engineering, in the cold war even those nuclear reactor powered Soviet recon satellites only lasted around 100 days in orbit, crap like that is exactly why they went bankrupt.

I've read that the "Oko" satellites lasted around 2 years in case of troublesome operation and their designed lifespan was 5 years in orbit.

Where do you get your information from? Just wondering.

http://en.wikipedia.org/wiki/US-KMO

[Broomstick]

I suspect early GPS had its share of problems, too, but because it started as a military program the public never heard about it, or at least not much about it.

[Sea Skimmer]

I've read that the "Oko" satellites lasted around 2 years in case of troublesome operation and their designed lifespan was 5 years in orbit.

Where do you get your information from? Just wondering.

http://en.wikipedia.org/wiki/US-KMO

The US-KMO series is powered by solar panels. Astronautix would be my default choice for space, though its not perfect. Other then that, I constantly collect information and it can be hard to tell you what is what without work. I don't think anything I've mentioned or will mention here though couldn't be found via google in seconds though. Space is not a subject I buy books on.

The Oko series were a nightmare because to function they had to transmit optical data back to earth in real time; while in comparison no Soviet ground imaging photorecon satellite ever had a image return by radio capability, pure film return to the fall of communism IIRC. The US did not have an easy time with its first generation early warning satellites either, it demanded a lot of processing capability in orbit. This is a tech field where its kind of impressive Soviet spacecraft worked at all.

The nuclear reactor satellites were the RORSAT style radar satellites which had to be nuke reactor powered because at the very low orbits they used the drag of Soviet solar panel tech to power them would have dragged them out of orbit. Design lifespan was IIRC 6 months but they often lasted less. It did not help that Soviet radar transmitter tech also sucked on lifespans in general. The RORSATs actually worked by batches, they'd store up data, and then dump it all on command, rather then constant broadcasts of actual target locations. That bit I know from Friedman's book on networked warfare.

Ultimately the Soviet method of rigidly dividing R&D labs from the production facilities ensured stuff like this would suck. Read about any high tech western program and you'll find whenever the west did that it also got shit for results.

I suspect early GPS had its share of problems, too, but because it started as a military program the public never heard about it, or at least not much about it.

Doubtful. Remember GPS was already the third US series of navigational satellites. The first series to go operational was Transit in 1964, something made possible by the utter US dominance in semiconductors and microprocessors back in that era. Accuracy was to 400m from a single fix (you could match GPS if you were stationary), which was not great but it was far better then anything else that would work over the open ocean. Primarily this system was used by SSBNs, so they could take navigational fixes at periscope depth, instead of using a periscope to navigate by the stars which had problems with cloud cover. 400m was better then the raw accuracy of a Polaris missile so it was very useful. Land based radio navigation couldn't work over the wide open Pacific to better then about 2km accuracy with OMEGA, which was rather less acceptable for employment of nuclear missiles.

The USSR orbited a system similar to Transit in the 1970s, but again, it employed a much larger number of spacecraft then the US system. In general though the gap in reliability of early spacecraft does not seem to have been large, but by the end of the 1970s a yawning gap opened up and the US began orbiting stuff the USSR just couldn't match at all like KH-11, the first photorecon satellite that downlinked digital images. The first was in orbit in 1976, though the US kept building some film return satellites until the mid 1980s because they had better very wide area coverage. Once the shuttle program got going that role was pointless.

[Elheru Aran]

Random, possibly non related question: what's the difference between GPS and say Google Maps Navigation on my phone? I presume the GPS thing communicates directly with satellite, while Maps uses the Internet, more or less?

[Zixinus]

Random, possibly non related question: what's the difference between GPS and say Google Maps Navigation on my phone? I presume the GPS thing communicates directly with satellite, while Maps uses the Internet, more or less?

GPS is hardware, Google Maps is software. GPS is a signal made by the satellites that your phone recieves. It is made by the US. Galileo is roughly the same but made by the European Space Agency. Russia (Glonas), China and India have their own satellites too. Some phones allow you to use more than once.

Google Maps often uses internet to get map data but has no real preference. Your phone converts the GPS (or other) data into location data that then Google Maps uses (I think). Google Maps "needs" an internet connection because that's how it gets map data. Other navigational software (like Sygic or pretty much anything else) allows you to download the map data so you don't need Internet connection to use them, just enabling GPS or "location services" on your phone.
Note that you can also use mobile network to get location. This uses mobile data but allows faster connection. Even Internet connection can provide location data somehow.

[Sea Skimmer]

Random, possibly non related question: what's the difference between GPS and say Google Maps Navigation on my phone? I presume the GPS thing communicates directly with satellite, while Maps uses the Internet, more or less?

Google maps uses the internet to source map data from a server for the image display. It is using a GPS receiver in your phone listening to the satellites for your actual locational data, and that function works fine with no internet, it's just not all phones will let you actually look at your actual gird coordinates without using a third party ap like Google Maps or something else you downloaded.

Its seriously in the tune of 10 USD to add a GPS receiver chip to a smartphone these days, and its antenna is integrated with the normal phone antenna. So all of them have it. I think standalone GPS capable watches are under 20 USD now.

The quality of GPS receivers varies though. GPS only works with at least four satellite signals, but it gets the best results with a much larger number, ideally at least eight and I believe the maximum that can be in view is twelve, but realistically a receiver is unlikely to pick them all up unless its on a high altitude plane.

However to save money on the hardware GPS can timeshare a single receiver channel, and so it listens to one satellite, records a bit of data, then listens to another ect... and calculates your position based on this memory stored data in batches. Any position produced in between batches is just being estimated by dead reckoning methods. For most civilian purposes this works fine, but it is a reason why cheap GPS mapping systems fritz out sometimes.

Additionally GPS can have its accuracy greatly upgraded by also receiving ground based augmentation signals that correct for atmospheric distortion of the radio signals (only an option in North America, Western Europe and Japan presently) and in some areas, I think only in the lower 48 states right now fixed ground beacons can also be used (this allows survey levels of accuracy). At least some phones do use the ground based augmentations.

Good GPS receivers have enough hardware to listen to at least eight channels in real time, and constantly compute the position based on all that information. This is why you can get a GPS chip for 10 dollars, and yet we still also still bother to sell units that cost hundreds and thousands of dollars. Phones vary on what they use, GPS built into cars is the expensive sort. Though even then GPS is kind of hard pressed to tell you which lane you are in on a moving vehicle (Google Maps uses educated guesses because it knows where the roads 'should be')

Galileo and the incoming Block III satellites for GPS will produce moving user accuracy nearly as good as the present ground augmented systems, without needing ground augmentation. This is the result of using much more accurate clocks on the satellites, improved transmitters with cleaner signals, and waveforms which deal with atmospheric distortion better. But mainly its just about the better clocks.

As is in the worst case, say four satellites and a high speed plane in a thunderstorm, civilian signal GPS can still only guarantee 100m level of 3D positional accuracy. Its just for civilians worst case is seldom relevant, though it has held back reliance on stuff like GPS landing aids for aircraft. Because you need those systems most in bad weather and that's when GPS works the worst.

Also if one is involved in ultra precise uses, GPS actually generates WGS-84 grid coordinates, not latitude and longitude, and conversion errors do apply. But unless your surveying or involved in map making that isn't a big deal. We have to correct our maps for continental drift as is, it can be up to several inches a year in some parts of the world and for things like self driving cars that's a problem if the maps aren't updated. That's where a system like google maps beats a standalone system, because Google will apply the error corrections from time to time without telling you.

For lol when they built the Channel Tunnel it was actually the first time we know where England was relative to France with any accuracy. All surveying prior to this had been done relative to fixed points in each country, and never carried across the channel. As the distance is rather far for optical survey methods it took a lot of work to come up with accurate measurements (they didn't want to trust satellite methods completely, 1980s and all ) and found out the 'assumed' relative locations were off by something like IIRC several meters. Which you know, when your boring a tunnel like five meters in diameter from both ends at once could have been a BIG PROBLEM.

[Sea Skimmer]

Note that you can also use mobile network to get location. This uses mobile data but allows faster connection.

That's based on the time signals from the cellphone towers, which are needed to synchronize the networks anyway or else they'd never work as packet switching/time sharing networks. Accuracy is not always amazing, something like 30 meters, but in the US its actually federal law that phones must be capable of doing this for emergency call locating purposes. They passed that law before GPS got so cheap every phone would have it though.

GPS time is really important to synchronizing military digital frequency-agile radio systems too. So much so that the US is now throwing money at designing ultra low cost atomic clocks to put into its latest tactical radios, so that they will still work if GPS is blown up or jammed. Earlier radios could rely on less accurate time signals, basically a really good Swiss wristwatch would do the job. But they only changed frequency several times per second, rather then every single pulse the way 4G can work.

Even Internet connection can provide location data somehow.

Just a variation of the above on a phone. If you mean to your house then that's just because the phone company knows where you live.

What holds this back is the celltower clocks are not as accurate as the clocks thrown into orbit (otherwise the towers would cost a fortune) and the known survey locations of the towers antennas themselves are not always hyperacurate. The very fact that a cell tower will wobble in the wind and bend with the uneven thermal expansion of the sun on the metal creates errors. Satellite systems have elaborate methods of correcting for their own error sources like solar radiation pressure, this is also why as I mentioned above they need constant ground control to keep everything aligned and working. So end result is the celltower data sometimes can be better then GPS, but other times it's not. Many phones cross reference the signals to get the best results they can.

[Sea Skimmer]

NASA Reference Publication 1390: Spacrcraft System Failures and Anomalies Attributed to the Natural Space Environment.
https://ntrs.nasa.gov/archive/nasa/casi ... 050463.pdf

Was trying to find a PDF on spacecraft reliability that was not 600 pages, not scanned on a 1990 scanner, and not so simplistic as to be useless. And.... I'm bored of failing at that, so I'll settle for this because it's interesting if nothing else. And I think it helps point to the kind of difficult to test on earth conditions spacecraft systems must withstand, and the thus the problems and dangers those tests could involve in terms of damaging hardware themselves. Also GRAPHS! And if you want minimal reading it has an appendix at the end listing many space environment caused failures including ones from Russia and Soviet Canada.

[K. A. Pital]

The nuclear reactor satellites were the RORSAT style radar satellites which had to be nuke reactor powered because at the very low orbits they used the drag of Soviet solar panel tech to power them would have dragged them out of orbit. Design lifespan was IIRC 6 months but they often lasted less

The US-A RORSATS had a low lifespan indeed - 120 days or so (the earliest ones had a lifespan of just 45 days). But the successors, equipped with Topaz reactors, lasted around a full year.
http://en.wikipedia.org/wiki/Kosmos_1867

Ultimately the Soviet method of rigidly dividing R&D labs from the production facilities ensured stuff like this would suck.

I thought that the reactors for RORSATs were developed and produced at the NPO "Red Star"... could be wrong, but in this case there was no separation of development and production.

[Enigma]

Soviet Canada?

[Tvpnbb]

Soviet Canada?

I made a careful analysis of the text and came to the conclusion that he was employing an obscure advanced literary technique called a joke.