About 3% of Starlink Satellites Have Failed So Far (phys.org) 165
According to Phys.Org, about 3% of SpaceX's Starlink satellites "have proven to be unresponsive and are no longer maneuvering in orbit, which could prove hazardous to other satellites and spacecraft in orbit." From the report: In order to prevent collisions in orbit, SpaceX equips its satellites with krypton Hall-effect thrusters (ion engines) to raise their orbit, maneuver in space and deorbit at the end of their lives. However, according to two recent notices SpaceX issued to the Federal Communications Commission (FCC) over the summer (mid-May and late June), several of their satellites have lost maneuvering capability since they were deployed. Unfortunately, the company did not provide enough information to indicate which of their satellites were affected. For this reason, astrophysicist Jonathan McDowell of the Harvard-Smithsonian Center for Astrophysics (CfA) and the Chandra X-ray Center presented his own analysis of the satellites' orbital behavior to suggest which satellites have failed.
The analysis was posted on McDowell's website (Jonathan's Space Report), where he combined SpaceX's own data with U.S. government sources. From this, he determined that about 3% of satellites in the constellation have failed because they are no longer responding to commands. Naturally, some level of attrition is inevitable, and 3% is relatively low as failure rates go. But every satellite that is incapable of maneuvering due to problems with its communications or its propulsion system creates a collision hazard for other satellites and spacecraft. As McDowell told Business Insider: "I would say their failure rate is not egregious. It's not worse than anybody else's failure rates. The concern is that even a normal failure rate in such a huge constellation is going to end up with a lot of bad space junk."
The analysis was posted on McDowell's website (Jonathan's Space Report), where he combined SpaceX's own data with U.S. government sources. From this, he determined that about 3% of satellites in the constellation have failed because they are no longer responding to commands. Naturally, some level of attrition is inevitable, and 3% is relatively low as failure rates go. But every satellite that is incapable of maneuvering due to problems with its communications or its propulsion system creates a collision hazard for other satellites and spacecraft. As McDowell told Business Insider: "I would say their failure rate is not egregious. It's not worse than anybody else's failure rates. The concern is that even a normal failure rate in such a huge constellation is going to end up with a lot of bad space junk."
Lower orbits = hopefully less debris (Score:5, Insightful)
IIRC, they need so many because they are at a lower orbit, and will drop out from normal friction far sooner then something in geosynchronous orbit.
At geosync, part of the design is they go into a parking orbit which is actually higher when they are EOL. Starlink, as annoying as it may be for astronomers at the lower altitudes means they are going to get more friction, and without orbital corrections will deorbit sooner. Still annoying until they do, of course.
Re: (Score:3, Informative)
they need so many because they are at a lower orbit
Indeed, the low orbit (~500 km) necessitates a mega-constellation. From any given point on the earth, any satellite more than about 20 degrees away (as an angle from the center of the earth) will be beneath the horizon.
And, to complicate matters, the ground terminals can probably only see to angles of +/-60 degrees from straight up* (without tilting the terminal), which means that only satellites within about 7 degrees (again, referenced to the center of the earth like latitude) can be "seen" by the ground
Re: Lower orbits = hopefully less debris (Score:2)
They said they were going to use a different, non phased antenna. Did they decide to go back to the phased one?
Re: (Score:2)
The ground units I saw in pictures of firefighter deployments were using a single parabolic antenna. That doesn't seem consistent with phased array at the ground station.
Back when I was doing RLL stuff, we liked phased array for going horizontally between BS and CPE. Cheap and effective dynamic beam steering, so the installer only needed to point it in the general direction of the base station.
As per your write up, it seems the trade offs work differently when pointing up to the sky.
Re: (Score:2)
Of course I can't find the picture now. It was from a deployment with Washington state wildfire fighters that was in the news recently.
Two pictures of two ground units sitting on the ground with the reflector pointing up and what looked like a central antenna pole. I automatically assumed is was standard parabolic thing. But of course I might be wrong and I'll look more closely if I see the picture again.
I did find this, which looks exactly like what you linked. So I was probably just wrong. https://twitter [twitter.com]
Re: (Score:2)
It's not a parabolic antenna, the top is flat (where the phased array is), and the bottom curved part contains the electronics. It looks like a parabolic antenna if you put a lid on the dish. That's just the industrial design, though.
Re:Lower orbits = hopefully less debris (Score:4, Interesting)
It's also hugely erroneous to assume a constant failure rate, in a constantly improving line of mass-produced satellites.
I noticed we've stopped talking about "SpaceX will destroy astronomy" FUD now that SpaceX has - as they said would - continue iterating their designs until the albedo was low enough to not create wash out patches of images if imaged at the worst times. So I guess it's inevitable that this will be the topic du jour for the next year or so...
Collisions between dead satellites make up only 2,5% of fragmentation events and 11,6% of debris [usra.edu] - and Starlink satellites are small. Starlink is not going to create an order of magnitude increase in debris generation even if they retain their current "meaningful-but-below-average" failure rate. Indeed, it's barely going to move the needle. NASA only requires a sub-10% rate [illinois.edu], and that's considered "optimistic". The above linked study uses an assumed 10% failure rate for Starlink and says "but it effectively creates a best-case scenario for the LEO environment so that the effects of Starlink can be more easily studied." 10% was assumed to be "best case". They then compared to other "potentially improved" failure rates As you can see in the graph on page 33 / figure 7, a 1% Starlink failure rate would be virtually indistinguishable from the no-Starlink baseline scenario, while a 5% Starlink failure rate is barely above it.
But of course, people are going to spread this "OMG STRALINK IS GONNA CAUSE KESSLER SYNDROME!" nonsense far and wide. Luddites gotta luddite.
Re: (Score:2)
"Starlink satellites are small"
yes, but space is big. Even trhe bit of space around the Earth where we orbit things, so small is so relative that it becomes irrelevant. Even a huge saturn-5 rocket is small in comparison to the space available in orbit, and yet even paint flecks are massively dangerous.
I'm more concerned about the 3% figure over time - is this a "it either fails at first boot, or it works forever" type of failure, or is 3% failure rate something that we will see entire clusters of sats start
Re:Lower orbits = hopefully less debris (Score:4, Insightful)
1) The odds of a collision is proportional to the satellite's cross section, e.g., it's size.
2) The amount of debris generated, and the size of the debris, is proportional to the size of the colliding spacecraft.
3) Heavier, dense debris fragments take significantly longer to deorbit than light objects.
When it comes to space collisions, size absolutely does matter.
And again, there's no need for conjecture. I linked a research paper specifically about Starlink. Even at a 10% failure rate there's no Kessler syndrome - there's just a less -than-doubling increase in the amount of LEO debris. At a 5% failure rate, there's a less than 20% increase in LEO debris. At a 1% failure rate there's virtually no change in the amount of debris.
It is not a problem.
Re: (Score:2)
Cheap, mass produced is SpaceX's specialty. That's why they have a Merlin engine on the F9 second stage, and will also have engine commonality in their upcoming Starship stages.
But that has no effect on reliability. The Merlin engines have flown more orbital flights than any other rocket engine in history except the RD-107, with only a single engine-related failure (which did not affect the primary mission). As per the fine summary, the 3% satellite failure rate is about average for the industry. And the St
Re: (Score:3)
You make some excellent points.
Historically it is not a given that mass produced products constantly improve over time... certainly not by the metrics of outsiders. It is usually the case that maintaining or even reducing cost is the metric of 'improvement' for the company producing the products which generally carries with it a reduction in quality and increased failure. That 1% failure may well raise to 10% and be considered improvement if the cost is perceived to experience a net negative effect.
Re: (Score:2)
The odds of getting diagnostic data to determine which ones failed and why is going to be a bit tough if they aren't communicating and they'll be destroyed on re-entry. I can't imagine how this will help them iterate to a better design.
Re: (Score:2)
Depends on the nature of the failure. If electronically it is responsive, but something mechanically 'sticks', then they may have telemetry showing what failed. It may be a bit more guesswork than being able to examine the mechanically stuck thing, but they at least know where to look and maybe make decisions to make something more robust if they know it is causing failures.
Re: (Score:2)
As a general rule, most failures in spacecraft can be tracked to their cause or at least a rough idea of the cause. And if for some reason you're getting failures you can't determine the cause of, you can instrument subsequent spacecraft to send more diagnostic data that can help you pinpoint the cause when you get a failure.
Re: (Score:2)
Yes, not as straightforward as being able to do a post-mortem, but particularly with such a large sample size they have lots of opportunity to get statistically significant failure data and fix and/or instrument to improve things generation to generation.
I'm personally still not convinced on the practicality of the overall business of this method of internet service, but from a technology perspective I think there at least strategies are sound for data gathering and improvement.
Re:Lower orbits = hopefully less debris (Score:4, Insightful)
It's also hugely erroneous to assume a constant failure rate, in a constantly improving line of mass-produced satellite
It's also hugely erroneous to assume a constantly improving line of mass-produced satellite when it is more cost effective to determine the cost of continually improving the MTBF for the satellites vs the minimum amount of time a satellite needs to last to return a decent profit.
See the 80/20 rule.
Re: (Score:3)
I don't remember the exact number or where I read it (one of their government filings, I believe), but I remember SpaceX recently boasting that they had gone some significant number of satellites with no failures (IIRC it was like two launches worth or something), indicating that their reliability was improving.
It's worth noting that even if 3% of their satellites failed, that is not equal to the number of satellites that will fail on station without the ability to de-orbit. They launch to a lower parking o
Re: (Score:2)
By and large, they really have. The albedo on new Starlink satellites is vastly lower than it was before. The "darksat" experiment cut brightness in half, while the sunshades (now standard) cut it by an order of magnitude. They've also changed how the satellites are angled at different times and how they ascend.
Starlink satellites, like all satellites and space debris (Starlink will ultimately basically double the number) will still leave streaks. But software removes streaks. The important aspect is to
Re: (Score:2)
(Sarcasm)What Starlink didn't kill of the remaining Astronomers yet? I am very depressed.(End of Sarcasm)
The modern Astronomer isn't spending days looking into they skies. But sitting in front of a computer crunching numbers.
Re: (Score:2)
IIRC, they need so many because they are at a lower orbit, and will drop out from normal friction far sooner then something in geosynchronous orbit.
No, they need so many because they need HUGE bandwidth. Although, in a non-geostationary orbit, having a shitload of satellites *also* helps with antenna beams as a bonus.
Re: (Score:2)
True, but debris isn't that much of a problem at geosynchronous orbit since the satellites are moving at a zero relative angular velocity, so there is little chance of them bumping into each other.
GEO [Re:Lower orbits = hopefully less debris] (Score:2)
True, but debris isn't that much of a problem at geosynchronous orbit since the satellites are moving at a zero relative angular velocity, so there is little chance of them bumping into each other.
Almost true. Once a GEO satelllite dies and stops stationkeeping, lunar perturbations and the non-sphericity of the Earth tugs it out of its spot, and satellites start drifting at non-negligible velocity.
Unfortunately, what perturbations do NOT do is change the semimajor axis, the orbital altitude. So they stay at synchronous altitude, just not in synchronous position.
Re: (Score:2)
In the immortal words of Mr. Knight: "Telstar! Wow! Isn't that the satellite that's raining debris all over Europe?"
Re: (Score:2)
There is an aspect in Science Fiction that people seem to miss. The fact that it is fiction.
The story plot that Modernization will bring doom on us all, is a very old story plot. This is the bases of many bible stories - modern ligature.
Adam and Eve get attracted to knowledge and suffer.
The folks who have created large powerful cities, have fallen, while the man of the land who rejected such city prosper.
The use of Magic comes with consequences.
Empires with advanced technologies are seen as daemons and mo
Re: (Score:2)
Re: (Score:3)
part of the design is they go into a parking orbit which is actually higher when they are EOL.
This kind of logic is exactly the opposite of what should be done. The spare satellites should be parked at a higher orbit where they can easily be maneuvered down into a place at a lower orbit whenever needed to fill a gap. Once a satellite is no longer useful or can not be communicated with then there should be a failsafe timer (dead-mans-switch) that will kick off an automatic deorbiting process. If you can't control the satellite it should deorbit itself. To deorbit properly all you need to do is somet
Re: Lower orbits = hopefully less debris (Score:2)
Interesting idea you have there. Now let's look at the math for the geostationary satellites you're complaining about.
To put them into an end of life graveyard orbit, it requires a delta-v of about 11 meters per second. To actually deorbit them, it requires a delta-v of about 1500 meters per second. And if they had that delta-v available, they could use it for station keeping and continue to remain in service.
So your idea really doesn't take into consideration reality.
Re: (Score:2)
It doesn't take a delta-v of 1500 meters per second to de-orbit them because that would seem to largely ignore drag even at their operational altitude. A delta-v of 0 m/s will eventually result in a de-orbit, and any delta-v you spend on the task simply accelerates the timeline.
Re: Lower orbits = hopefully less debris (Score:5, Informative)
Re: (Score:2)
If it is broken, their orbit will decay fairly fast.
"Fairly fast" is relative. The dead satellites are expected to stay in orbit anywhere from a few months to up to five years.
And with approval for about 12,000 satellites 3% gives about 350 dead ones in the upcoming round. All of them will eventually need to be de-orbited, although hopefully live satellites can be de-orbited over empty ocean where the damage is pollution rather than impact.
While the risk of cascading failures (Kessler syndrome) is low, each item increases the space debris pool. Dead satelli
Re: Lower orbits = hopefully less debris (Score:4, Informative)
Re: (Score:2)
It's more than a bit hyperbolic to talk about "debris fields raining down" from satellites this size. SpaceX launches 60 of them on a single rocket. I doubt they even have any parts large and solid enough to survive re-entry.
Re: (Score:2)
These satellites are not in a stable orbit, Orbiting in lower orbit without the aid of some correction is impossible. The Gravity from the Earth and the Moon. Also this isn't a 0 friction environment like in the physics book. There are some gases and particles that are in the same orbital path. So their orbit will decay faster than some in higher orbit.
The Moon is fairly steady as it is in very high orbit, has a lot of mass and momentum to deal with any little bit of friction, and the pull of the sun gr
Re: (Score:2)
they need so many because they are at a lower orbit
Not exactly.
Yes, exactly.
A low orbit can see only a small patch of the Earth. So you need more of them.
In the limiting case of altitude a (above the surface) being But, yes, it's also true that low orbits decay quickly.
Re: (Score:2)
The full constellation is roughly 42,000 satellite, with an operational lifespan of at most five years (probably less). That means that simply maintaining the constellation will require launching/expiring 8,400 satellites per year.
They are currently launching 60 satellites per Falcon 9 with a targeted frequency of two launches per month (though it looks like they'll only hit 15 this year), but plan to eventually switch launches to Starship, which will do 400 satellites per launch with more (full) reusabilit
Article is missing a lot of relevant info (Score:5, Informative)
Also the article does not include a link to the report, I checked the last reports and there's no such info there. I won' be surprised to see that the planned retirement of the initial 60 experimental satellites is included in the 3% count.
Re:Article is missing a lot of relevant info (Score:4, Interesting)
Re: (Score:2, Interesting)
It would be unusual if they completely burned up in the atmosphere. Disintegrated into very small parts perhaps, but completely consumed into vapour? Unlikely.
Since they have lost control there is no way to control where they de-orbit. With the very large number of these satellites de-orbiting on a regular basis we will have to watch carefully to see what sort of pollution there is.
Re: (Score:3)
Re: (Score:2)
I wonder what kind of liability insurance they have.
They have submitted paperwork for over 30,000 satellites but it's not clear if that is on top of the 12,000 they already planned for. Anyway, say 30,000, 3% failure rate, that's 900 satellites in uncontrolled re-entry. Per 5 years, since that's the lifespan before they need to be replaced.
3% might be a low estimate given the system is quite new. More will doubtless fail over time. Then again they might get more reliable.
Re: (Score:2)
The 30,000 satellites is in addition to the initial 12,000.
Re: (Score:2)
A lot of people will push for those events to be classified as acts of God - being that the satellites were sent up by Saint Elon himself :-)
Re: (Score:2)
v0.9 was only 95% demisable. v1.0, which is every launch after the first, has been fully demisable. So they burn up completely.
Re: (Score:2)
Re: (Score:2)
People who are worried about getting hit by satellite debris haven't spent much time outside of a city.
There is really a LOT of open space on the earth. Never mind the oceans.
Remember when Taco Bell put a target in the ocean for Mir to hit? That was awesome. Free taco for every person on the planet if it hit. They even bought insurance in case it did. Best marketing ploy I've ever seen.
Re: (Score:2)
Decades ago, I got a call from the FBI.
It seems that after I formed a corporation for a client, it sold a million dollar policy for a hole-in-one at a golf tournament, or some such.
Aside from not being licensed for such activity, it apparently just pocketed the premium.
And someone got the hole in one . . .
I'd had no contact with the client since the initial shareholder's meeting (more than half of these never came back at their second year meeting at the time . . .).
hawk, esq.
Re: (Score:2)
Re: (Score:2)
I won' be surprised to see that the planned retirement of the initial 60 experimental satellites is included in the 3% count.
No, because most of those initial sats still respond to commands. TFA specifies satellites that cannot be controlled any more, not obsolete sats.
Re: (Score:2)
TFA specifies satellites that cannot be controlled any more, not obsolete sats.
The article also uses the 3% number like it isn't speculation.
Re: (Score:2)
It's not speculation, he measured satellite orbits.
These fall into 3 categories:
1. orbit actively maintained against drag. These are operational sats.
2. orbit being lowered by more than drag alone. These are obsolete sats still under active control.
3. orbit predominantly influenced by drag. This is the 3% category. The only speculation is that these are all inoperative. Not a giant leap of imagination, IMO.
Many unknowns - this is an early warning (Score:2)
Re: (Score:2)
Re: (Score:2)
It depends; your characterization is completely true of some of the satellites in question, less true of others.
SpaceX plans to operate satellites in three different altitude bands: 340 km, 550 km and 1100 km. The lowest group will definitely de-orbit in a matter of months if not boosted, but the highest satellites might take years to deorbit naturally. For comparison, the ISS's orbit is at 410 km and decays at about 2km/month.
Dust in the Wind [Re: Article is missing...] (Score:3)
If they burn up in the atmosphere, that will cause more pollution, right? Hopefully Musk has a plan for that. Otherwise these may be contributing to climate change.
Oxidized metal: i.e., dust.
Trivial compared to the hundred metric tons of dust [astronomy.com] falling into the atmosphere each day from natural events.
Re: (Score:2)
Re: (Score:2)
In California, it's illegal to have a bonfire in your back yard. Because of the pollution they create.
[citation needed]
I won't claim there is no such state law because hey, it's California and I haven't looked, but given that you can have them on the beach [californiabeaches.com], I'd say that any reason for fire bans are either local ordinances based on neighborhood, or are typically due to weather conditions that make spread likely, and don't have jack to do with pollution (do you see how many darned fire pits there are in that beach photo?)
Re: (Score:2)
I didn't say it was State Law; in fact, most California counties with significant populations (and many a few thinly populated ones as well; especially those in forested areas) require permits for residential burning. In general, things you can't burn even with a permit include yard trimmings, trash, garbage, plastic, or treated lumber; those restrictions are imposed by Air Quality Management Districts, hence my reference to pollution. You can't have a fire more than 4 feet in diameter; that alone eliminate
Kessler Syndrome danger (Score:5, Interesting)
Re:Kessler Syndrome danger (Score:4, Interesting)
SpaceX had originally planned for their satellites to be higher (though I forget the exact orbital height), and they lowered the orbits in later filings to address the issue of space debris.
Re: (Score:2)
Re: Kessler Syndrome danger (Score:2)
Actually, a key thing to remember is that the altitude and velocity of a satellite is always correct for where the satellite currently is. A collision CAN NOT create debris in a higher orbit. It CAN create debris in an elliptical orbit that reaches a higher altitude, but that new elliptical orbit will also include the same altitude and velocity the debris had at the moment after the collision.
Re: (Score:3)
a dead Starlink satellite will deorbit in 1-5 years
Hmm ... let's see ... (source: SpaceX application to the FCC [arstechnica.net])
To this end, SpaceX will implement an operations plan for the orderly de-orbit of satellites nearing the end of their useful lives (roughly five to seven years) at a rate far faster than is required under international standards. Satellites in the LEO Constellation will de-orbit by propulsively moving to a disposal orbit from which they will reenter the Earthâ(TM)s atmosphere within approximately one year after completion of their mission.
Not the easiest thing to do when the satellite no longer responds to propulsion commands.
Still, there will be natural orbit decay regardless, so let us assume 1-5 years. The problem is you'll have a steady state of the constellation where a certain percentage of satellites will be down in the process of de-orbiting at all times. Now, if you want a back of the envelope calculation of the percentage, after ~2.5 years the percentage of non-responsive
Re: (Score:2)
Undoubtedly the failures follow a bathtub curve. So 3% "Every 2 years" isn't relevant.
Better or worse than Tesla cars? (Score:2)
Is this more or less reliable than Musk's other bit of technology, Tesla cars?
How often do Tesla cars need some non-trivial repair work in the first months of operation?
The problem with satellites is you can't just get your service agency to lend the user another satellite while you swap out the drivetrain, like Tesla does with the cars that break down.
Re: (Score:2)
How often do Tesla cars need some non-trivial repair work in the first months of operation?
Not often, according to customer satisfaction surveys. But when they do, it's a pretty simple matter to send a flatbed out and haul them to a garage for repairs.
Oblig. bad car analogy: If even a fraction of 1% of the vehicles on a road stalled each year but then were abandoned in place, roads would become impassible pretty soon.
Re: (Score:2)
"Oblig. bad car analogy: If even a fraction of 1% of the vehicles on a road stalled each year but then were abandoned in place, roads would become impassible pretty soon."
Indeed. Unfortunately, things in space tend to stay in space, for years to millennia, so it doesn't take a lot of breakdowns to clog things up. The ISS already has to dodge debris several times a year.
Re: (Score:2)
tend to stay in space, for years to millennia
Fortunately, Starlink is in LEO. So maybe 5 years to re-entry if no orbital adjustments are made. It'll be like that old car sitting on the shoulder with a red flag on it you've been passing for two weeks before it disappears.
Re: (Score:2)
How often do Tesla cars need some non-trivial repair work in the first months of operation?
Less than Jaguar, Aston Martin or Peugot?
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
A question about reliability:
Is this more or less reliable than Musk's other bit of technology, Tesla cars?
Your response about customer satisfaction which has nothing to do with the question of reliability:
They have the highest customer satisfaction in the industry by a wide margin.
Questions about reliability:
what is reliability rating? What is the MTBF?
Your response about customer satisfaction which has nothing to do with the question of reliability:
Customer satisfaction tends to be really important to me
Why are you avoiding the question of reliability
Re: (Score:2, Interesting)
Good for you. Your one experience is not data.
Here is some data from J D Power's "Initial quality study" - problems in the first 3 years of ownership. https://www.jdpower.com/busine... [jdpower.com]
Teslas had 250 problems per 100 vehicles, worse than any other manufacturer. The breakdown of power train issues (to compare with the rate of Starlink manoeuvring failures) is not available in the free data.
That doesn't give a precise answer but it seems like Starlink may well have less problems than Tesla cars during the initial months of operation.
Not some Tesla fanboi here (don't own cars or stock), but perhaps JD Power will present an accurate Tesla study one day, when they're allowed to survey all Tesla owners, not just the ones in 35 American states.
Seems the "initial" quality of this JD Power survey rather sucks when compared to their other studies.
Re: (Score:2)
Re: (Score:2)
Lol, 35 out of 50 is a fine sampling, unless you have reason to think there is something disproportionate about the quality about Tesla's cars in various states.... which would be a problem in itself.
Depends on which states, doesn't it? Purposely exclude certain states that could shift those results quite a bit when you're measuring problems per 100 vehicles. Teslas could have a hell of a lot less issues in milder climates, and that's but one variable. How long do you think the average ICE car lasts in Alaska vs. Florida? Obvious factors, are obvious.
An incomplete sampling can create obvious bias, especially when it comes to feeding the hatred and vitriol against Tesla, which is often unjustified.
Re: (Score:2, Insightful)
The hate is just a reaction to fanboys.
I want to love Tesla but we have that one guy at work who can't spend an hour without telling his coworkers how great his Tesla or any Elon Musk endeavor. We probably all know someone like him.
As a result, now I hate Tesla. Not because of their cars, they look like good cars, but because they turned normally interesting people into mindless cultists.
Re: (Score:2)
I'd buy one too if I had the cash but this one dickhead at work owns one. He walks around with his Tesla travel mug and his Tesla jacket. He doesn't talk about it much but he is sort of an asshole with his work.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Actually don't have an issue with their sampling method, it's the same as for every manufacturer they survey for (they send out voluntary surveys in the mail).
The big issue is that the large majority of issues in their initial quality survey have to do with infotainment and people not understanding how electronic functions work. For instance, "I couldn't get Sync to recognize when I say Szechuan Restaurant" (no seriously, go look some of these results up, lots are posted online).
So if your car has a LOT mor
that's much less than... (Score:4, Interesting)
.. percentage of my failed builds in our automatic build system.
From the other hand the build agents are cheaper than millions of dollars per each satellite.
Re: (Score:2)
Bathtub Curve (Score:5, Informative)
This is the start of what's called a "Bathtube Curve". check it out [wikipedia.org]
Basically says that among a group of new items, you'll have an initially high failure rate as products with manufacturing defects and materrial flaws die early. Once these flawed units have dropped out of circulation, then follows a relatively long span of time with low failure rates and high product reliability. Some time later, as products approach their expected lifespan, you'll see failures sharply rise back up as they rapidly die of anticipated age-related degradation.
Re: (Score:2)
Where this does explain the 3% failure rate, it still doesn't bode well for SpaceX's reliability.
I suppose, if you own the launch systems and are launching hundreds of these things, having a few of them fail is pretty much unavoidable, but if they don't drastically improve that 3% failure rate in future batches, it's going to impact their profit margin pretty significantly. The ISP business is one where you work in the margins and you have to carefully manage you cash flow. The infrastructure investment
Re: (Score:2)
if they don't drastically improve that 3% failure rate in future batches, it's going to impact their profit margin pretty significantly.
Quantify this please. How do you know their profit margin? The marginal cost of each satellite? The marginal launch cost? The taxable income is offset per lost satellite? How do you justify 3% being huge investment in infrastructure?
Until you know the numbers, you're just making stuff up.
Re: (Score:2)
Sadly doing something like that today is impossible because, well, greed.
Re: (Score:2)
That's why they send up spares. We always have GPS spares in orbit, ready to take over. (we don't just launch another one when one breaks, to replace it, we launch to get another spare in station)
With so many starlink sats, you can expect them to have enough for some redundancy.
And they almost certainly did some "burn-in testing" on the sats, but there's just some things you can't easily test for, like problems induced by microgravity, unlucky solar radiation hits, and soforth. The high G's and vibration
Re: (Score:2)
And vicious k'nids.
You never can tell which kind of satellite they find tasty . . .
hawk, apparently dating himself again.
Re: Bathtub Curve (Score:2)
Fix it (Score:2)
Couldn't they add a watchdog/suicide module with a standalone ion thruster that orients and de-orbits the sat when it doesn't get a refresh signal after some amount of time?
Re: (Score:2)
Effectively, they have just that. At the altitude they are using the air resistance is not negligible and over time will deorbit pretty much everything unless you add energy to keep the object in orbit. So, when one of their satellites goes inert, it's going to deorbit itself due to drag in reasonably short amounts of time.
Re: (Score:2)
No need for a thruster. They just need drag. There are some experimental designs which just pop out essentially a long streamer to increase air resistance.
Article is highly misleading and sensationalist (Score:2)
You need a watchdog, Elon! (Score:2)
And an automated one.
If the unit cannot get in contact with the control center anymore, automatically abort it and deorbit to self destruction.
Re: (Score:2)
The problem could be that the propulsion system failed, it's pretty new technology
Initial orbit is even lower (Score:3)
If you look at McDowell's actual data, you'll see that later launches have lowered the initial orbit to a bit under 300km. Any Starlink satellites that fail immediately will drop out of orbit in months. The working ones use their ion thrusters to raise up to the operational orbit of 550km.
https://twitter.com/planet4589... [twitter.com]
As others have noted, there have been substantial improvements to the design since the initial batch, so I'd expect the failure rate to decrease in later launches.
- Necron69
Seems like the satellites should have deadman's... (Score:2)
Seems like the satellites should have a deadman's subsystem. If the satellites can't receive commands or the rest of the control systems are otherwise not responsive, the deadman's subsystem should be able to initiate de-orbit. Of course many failures would inhibit this, but for some classes of failure such as failure to communicate with the earth, this would "cleanup the junk" faster.
Re: (Score:2)
If the system isn't capable of orienting itself to point at Earth and responding to a very particular signal to say even "I'm still here", what makes you think it's even capable of executing any such command, let alone manoeuvring, let alone know it's doing it in a safe direction that doesn't take out anything within range?
You'll be lucky if it ever responds to a lack of contact, let alone responds in a way that safely de-orbits it rather than just shoots off into a tangential orbit straight into the path o
Beware of propaganda (Score:4, Interesting)
First, I'll admit to being a fan of SpaceX, though not a fanboy of all things Musk. I'll also admit I have not had time to dig fully into the numbers in THIS story, but is smells a LOT like a previous one that claimed a particular failure rate for Starlink satellites - and that should strike anybody as odd, consider:
1. When's the last time you heard anybody publicly complain about the failure rate of any product (and more particularly satellites) while not being a customer of said product?
2. When's the last time you saw that same basic public complaint repeated? (and for SATELLITES? Really?)
It's similar to all the articles about SpaceX rockets always blowing up and being unreliable - which were traced back to a lobbying firm for another aerospace company and were in papers commonly read by members o0f congress at a time when new space launch contracts were being discussed.
Now, here's my reason for suspicion here: The last time we saw these same claims about Starlink satellite failure rates, they were being pumped by a competing US-based satellite operator (I'll not name here to avoid a "my team" vs "your team" devolution). The arguments last time mislead readers by merging several issues: They counted a bunch of early generation Starlinks that were always meant to be temporary and which were being intentionally de-orbited as "failed" Starlink satellites. They also distracted from a major design point of the Starlink constellation which was that most of the birds are in low orbits and each satellite is designed to de-orbit itself if it fails so that the constellation overall is far less likely to leave orbital debris of dead satellites than any other constellation ever launched. Counting satellites that fail in their primary mission, but then properly deorbit themselves, as designed, leaving no orbital hazard - and labeling them generically as "failed" has the effect of misleading people into believing Musk's constellation is "bad", dangerous, and polluting space. With any product there will be a failure rate, and with thousands of satellites, the Starlink constellation will have many even if it's only a 1% rate, so the constellation as a whole and the individual birds are all designed with that in mind. It's sad when competing companies who have done far less to mitigate such issues on their own systems play these PR games.
Re: (Score:2)