Atomic clocks and GPS

Tinstafl

Tinstafl

Pre-Flight
Joined
Oct 17, 2019
Messages
73
Display Name
Tinstafl
apple.news

How a power outage in Colorado caused U.S. official time to be 4.8 microseconds off — NPR

Officials said the error is likely too minute for the general public to clock it, but it could affect applications such as critical infrastructure, telecommunications and GPS signals.
apple.news apple.news

This is kind of an academic interest, and I doubt it has real world downstream effects in the aviation world, but it made me curious about how this might link into the tools we use in flight.

I’m guessing the level of error might be too small to introduce any appreciable error in any field except scientific research or something. But do the satellites in usage for GPS source their error corrections from this facility directly or indirectly? If so, what amount would it take to impact our inflight accuracy? Is this actually a legitimate concern?

To be clear I’m not flying around worried my LPV is suddenly going to fly me into a building. Just curiosity.
 
Light, or in this case GPS radio signals, will travel about 1.5 meters in the 4.8 microsecond error referenced in the article. Someone who better understands the intricacies of GPS will probably be along shortly to explain how much of a positioning error this might cause with 4+ such signals used to determine the position of the receiver.
 
I wonder how they compensate for different gravity? Time itself runs slower with more gravity. A clock on the 10th floor will run faster than a clock on the 1st floor. Nothing those of us would ever notice, but a scientist running an experiment might.
 
I'd want more data on what the article is based on, it doesn't quite make sense. My initial reaction was that it was unlikely that an Atomic Clock would fail due to a power outage, and later in the article it confirmed their 16 clocks kept "ticking" as it were.

And I'm pretty sure that no one that needs time that is that accurate relys on one single time sourse. When the NIST time went away they just relied on their other sources. So why then did the NIST acually let their system send out an inaccurate time? The system should have resynced with the 16 still running clocks and THEN they should have started sending out a signal.
 
  • Like
Reactions: WDD
Fun Fact: They (the international UTC people) have to add a second every now and then to the time it broadcasts as the earth doesn't rotate in exactly a perfect 24 hours.
 
WDD said:
I wonder how they compensate for different gravity? Time itself runs slower with more gravity. A clock on the 10th floor will run faster than a clock on the 1st floor. Nothing those of us would ever notice, but a scientist running an experiment might.
Click to expand...
They absolutely do compensate for this.
It has been told that if they didn’t, the system would be useless within an hour.
 
GaryM said:
Light, or in this case GPS radio signals, will travel about 1.5 meters in the 4.8 microsecond error referenced in the article. Someone who better understands the intricacies of GPS will probably be along shortly to explain how much of a positioning error this might cause with 4+ such signals used to determine the position of the receiver.
Click to expand...
Sorry, light in a vacuum travels at about 1 ft/nanosecond. There are 1000 nanoseconds in a microsecond.
Therefore light travels 4800 feet in 4.8 microseconds.

186,000*5280 = 982,080,000 ft ~= 1 billion

From our human point of view, light is very, very fast.
From a universal POV, light is very, very slow.

Ask me about the "now."
 
GaryM said:
Light, or in this case GPS radio signals, will travel about 1.5 meters in the 4.8 microsecond error referenced in the article.
Click to expand...

Light travels very much faster than that — at about 3 nanoseconds per meter (or 30cm per ns), in fact (in a vacuum, etc.), or — to regress to the Imperial units of my very distant youth — it travels about a foot per ns. So in 4.8 microseconds it's gone about 4,800 feet — a good portion of a statute mile, if I can remember my Imperial units properly. That's significant. (I haven't read the article itself, though…).

Edit: Domenick beat me to it (by a nanosecond) :)
 
Domenick said:
Sorry, light in a vacuum travels at about 1 ft/nanosecond. There are 1000 nanoseconds in a microsecond.
Therefore light travels 4800 feet in 4.8 microseconds.

186,000*5280 = 982,080,000 ft ~= 1 billion

From our human point of view, light is very, very fast.
From a universal POV, light is very, very slow.

Ask me about the "now."
Click to expand...
You're looking at now, sir. Everything that happens now, is happening now. We passed then just now. We're in now, now.
 
Several have correctly pointed out the precision necessary to make GPS work. It's kind of amazing that it works at all given those complexities.

Now, keep in mind that the first launch was in 1978. Incredibly ambitious project for the technology of the day!
 
It seems that if our clocks on Earth and those onboard the satellites got out of sync, the practical impact is that the on-board GPS would see different positions coming from solutions computed with different subsets of the satellites it had found. That would trigger RAIM flags and you'd have to land using VOR/DME/ILS or in VFR conditions.

This makes me wonder what would happen if the time errors were consistent yet different: all the Earth GPS devices and clock sources (group A) agreed, all the satellites (group B) agreed, but group A differed from group B. Is it possible that all the positional solutions computed by the on-board GPS agree, while being incorrect? In which case it would be wrong with no RAIM error. The system is designed to make this impossible, but often the term "impossible" really means "unlikely".
 
WDD said:
Fun Fact: They (the international UTC people) have to add a second every now and then to the time it broadcasts as the earth doesn't rotate in exactly a perfect 24 hours.
Click to expand...
Fun fact #2: The earth rotates 360 degrees in 23 hours 56 minutes, not 24 hours.

It all depends on your frame of reference. In 23:56, the Earth makes a complete rotation in relation to the universe. However, in that time, the Earth has traveled around the sun far enough that it takes an additional four minutes to rotate so that the sun is again directly overhead. See "siderial time".

Ron Wanttaja
 
GaryM said:
Light, or in this case GPS radio signals, will travel about 1.5 meters in the 4.8 microsecond error referenced in the article. Someone who better understands the intricacies of GPS will probably be along shortly to explain how much of a positioning error this might cause with 4+ such signals used to determine the position of the receiver.
Click to expand...
Yeah I think it’s clear that this is nowhere close to a significant impact on our accuracy.

I think the only question I would have is what occurred that caused the error to be broadcast without being corrected by other units or locations. I would expect that there is a dispersed network of time sources that ensures stability in case there is a larger, longer, more significant event. But … what is it?

This falls well into the category of things we don’t need to worry about. But I’m a nerd . If anyone is knowledgable on this, I’m still curious.
 
  • Like
Reactions: WDD
AV8R_87 said:
When will then be now?
Click to expand...
Your now and my now will never be the same. No absolute simultaneity is possible, it's all relative. Except when speaking about anything with your wife. Then the only valid frame of reference is what she wants it to be.
 
So wait … … maybe it’s much more significant an error level than I fist assumed.

But I flew an ILS that day … so we got lucky and didn’t die… :-)
 
Doh! I used kilometers/sec in my calculation but expressed the resulting distance in meters.

Only off by a factor of 1000. Carry on.
 
TCABM said:
Each satellite has onboard atomic clocks that get an occasional sync with a DOD source.

GPS and Telling Time | GPS.gov

www.gps.gov www.gps.gov
Click to expand...
Specifically, the almanac is usually uploaded every day but may delay up to 6 days. Why 6 days? Because one week is an important time slice within an epoch. For example, we're in epoch week 2398, day 358. Epoch 1 was Jan 06, 1980. There's been 3 epochs since time began...officially Epoch 3 began in April 2019. People begin to worry as the rollover gets closer.

Fun fact....when everything started in 1980, number of bits in a computer word was a whole lot less than those 64 bit computers we use now. Only 10 bits to track an epoch, which means 2^10 or 1024 weeks in an epoch, when the week number begins again. Or not quite 20 years. Epoch rollover is can be a software burp if the system can't handle it. What system? Any system that needs GPS. Your cell phone. Your bank. We're pretty much dependent on GPS these days.

I didn't waste those 4 years on the Space Weather program nor the few months at [company that shall remain nameless] on the GPS system system for naught..

NIST in Boulder may have had a power failure but GPS command & control is run out of Schriever, which is east of Colorado Springs (KCOS)
 
Tinstafl said:
Yeah I think it’s clear that this is nowhere close to a significant impact on our accuracy.

I think the only question I would have is what occurred that caused the error to be broadcast without being corrected by other units or locations. I would expect that there is a dispersed network of time sources that ensures stability in case there is a larger, longer, more significant event. But … what is it?

This falls well into the category of things we don’t need to worry about. But I’m a nerd . If anyone is knowledgable on this, I’m still curious.
Click to expand...
As noted in the article, there are a dozen clocks located around Boulder, and all 12 feed into the NIST time system. What's not in the article is the fact there are other clocks around the world that also send data to NIST. Don't forget the other 3 GPS systems, Baidu, Galileo and Glonass. They all play somewhat nice with each other. Actually, not the systems themselves but most contemporary GPS receivers can take signals from 1 or more. But then there's the military M-Code US GPS signal but that's another discussion entirely and not here.
 
FWIW, from a typical IT system/network perspective I don't think microseconds matter. But milliseconds do matter, in terms of packets flows. So it's not unusual for a decent sized network to base their time off of a common internal set of time servers that typically get their time from GPS. So that even if GPS goes nuts for a bit or is unavailable, all of the internal systems will still run off of the same internal clock. Or in other words, it's not that the absolute time matters, it's that all the clocks in the system are all running from the same source.

This is even more important on smaller networks using Internet based time sources, where you're always going to get some delay in time in the ms level, and that may jump around enough to cause problems. Source for the latter, even if you're using good time sources:

 
AV8R_87 said:
When will then be now?
Click to expand...
<OnSoapbox>
Never. For two reasons, biological and physical, there is no "now."

We experience the universe through our senses. It takes a time lag of many milliseconds for sensory input to travel to your brain for processing before you experience the sensation. This is true even for spinal reflexes in which the signal input is so intense it jumps from sensory neurons to motor neurons in the spinal cord. This is why you hand jerks back from a hot stove.
Our sense of the now is actually the past. But not even that. Each of our senses has its own lag time measured in milliseconds. What we call "now" is actually a time smear of events in the recent past.

Physics makes this weirder. As mentioned above light travels at about 1 ft/nanosecond. Let's assume the screen at which you are staying is 2 feet from you eyes. It takes the screen light 2 ns to hit your retina. Now look at a wall, say 10 feet away. The light reflecting from the wall takes about 10 ns to hit your retina. Look out the window at an object, say, a mile away. It's light takes 5280 ns (that's a massive 5.28 milliseconds!) to hit your retina. Every object in you local reality is at a unique distance. Therefore your "now" is a smear of time. WDD above referred to this issue above, "No absolute simultaneity is possible, it's all relative." It's not just your now is different than my now, it's actually there is no "now," mine or yours.

But it gets weirder. Since the speed of light, that is the speed of information, is so damned slow, even the order of events can change from observer to observer. Say astronomers see a distant supernova(1), then 50 years later see another supernova(2). To us, event 1 occurs 50 years before 2. However, for a different observer somewhere else in the universe, they may see 2 occur before 1 because they are closer to 2. At astronomical distances that's a simple concept, but because light is so damn slow--5.28 milliseconds/mile, this reordering of event actually takes place in our own lives, but we rarely if ever notice. At quantum levels this is an issue and results in some damned weird effects.

We experience the universe in a timeless smear which raises all sorts of questions as to the nature of time and the meanings of past, present, and future.
<OffSoapbox>
 
Domenick said:
<OnSoapbox>
Never. For two reasons, biological and physical, there is no "now."

We experience the universe through our senses. It takes a time lag of many milliseconds for sensory input to travel to your brain for processing before you experience the sensation. This is true even for spinal reflexes in which the signal input is so intense it jumps from sensory neurons to motor neurons in the spinal cord. This is why you hand jerks back from a hot stove.
Our sense of the now is actually the past. But not even that. Each of our senses has its own lag time measured in milliseconds. What we call "now" is actually a time smear of events in the recent past.

Physics makes this weirder. As mentioned above light travels at about 1 ft/nanosecond. Let's assume the screen at which you are staying is 2 feet from you eyes. It takes the screen light 2 ns to hit your retina. Now look at a wall, say 10 feet away. The light reflecting from the wall takes about 10 ns to hit your retina. Look out the window at an object, say, a mile away. It's light takes 5280 ns (that's a massive 5.28 milliseconds!) to hit your retina. Every object in you local reality is at a unique distance. Therefore your "now" is a smear of time. WDD above referred to this issue above, "No absolute simultaneity is possible, it's all relative." It's not just your now is different than my now, it's actually there is no "now," mine or yours.

But it gets weirder. Since the speed of light, that is the speed of information, is so damned slow, even the order of events can change from observer to observer. Say astronomers see a distant supernova(1), then 50 years later see another supernova(2). To us, event 1 occurs 50 years before 2. However, for a different observer somewhere else in the universe, they may see 2 occur before 1 because they are closer to 2. At astronomical distances that's a simple concept, but because light is so damn slow--5.28 milliseconds/mile, this reordering of event actually takes place in our own lives, but we rarely if ever notice. At quantum levels this is an issue and results in some damned weird effects.

We experience the universe in a timeless smear which raises all sorts of questions as to the nature of time and the meanings of past, present, and future.
<OffSoapbox>
Click to expand...
And, there is exactly zero time between "I'm being careful" and "Oh ****, I don't want to look" when your fingers get sucked into the blade of your table saw.
 
Domenick said:
<OnSoapbox>
Never. For two reasons, biological and physical, there is no "now."

We experience the universe through our senses. It takes a time lag of many milliseconds for sensory input to travel to your brain for processing before you experience the sensation. This is true even for spinal reflexes in which the signal input is so intense it jumps from sensory neurons to motor neurons in the spinal cord. This is why you hand jerks back from a hot stove.
Our sense of the now is actually the past. But not even that. Each of our senses has its own lag time measured in milliseconds. What we call "now" is actually a time smear of events in the recent past.

Physics makes this weirder. As mentioned above light travels at about 1 ft/nanosecond. Let's assume the screen at which you are staying is 2 feet from you eyes. It takes the screen light 2 ns to hit your retina. Now look at a wall, say 10 feet away. The light reflecting from the wall takes about 10 ns to hit your retina. Look out the window at an object, say, a mile away. It's light takes 5280 ns (that's a massive 5.28 milliseconds!) to hit your retina. Every object in you local reality is at a unique distance. Therefore your "now" is a smear of time. WDD above referred to this issue above, "No absolute simultaneity is possible, it's all relative." It's not just your now is different than my now, it's actually there is no "now," mine or yours.

But it gets weirder. Since the speed of light, that is the speed of information, is so damned slow, even the order of events can change from observer to observer. Say astronomers see a distant supernova(1), then 50 years later see another supernova(2). To us, event 1 occurs 50 years before 2. However, for a different observer somewhere else in the universe, they may see 2 occur before 1 because they are closer to 2. At astronomical distances that's a simple concept, but because light is so damn slow--5.28 milliseconds/mile, this reordering of event actually takes place in our own lives, but we rarely if ever notice. At quantum levels this is an issue and results in some damned weird effects.

We experience the universe in a timeless smear which raises all sorts of questions as to the nature of time and the meanings of past, present, and future.
<OffSoapbox>
Click to expand...
You're 100% correct. But man, when people only understand part of the concepts you're laying out they can come up with some really goofy ideas.

Case 1 - Customer that was convinced that they're printouts would come out quicker if the CAT 5 network cable at the end was 5' long instead of 20' long. Twenty something me may not have been kind in pointing out their error was along the lines of 10 orders of magnitude out of whack.

Case 2 - Working as a ham as part of a local charity run. We sync the race clocks along the way via a 2m radio link, all analog, to the starting gun. The clocks are just there as a guide for everyone, it's not the official time. The race guy beside me, who must have paid attention to a week of physics somewhere, was convinced that the radio time between the starting point, the repeater, and back to my radio, about a 10 mile circuit, introduced a significant delay. Well, unlike a Webex call, which is intentionally delayed by a couple of seconds to allow packet reassembly, analog VHF radio has only the delay of the electrons flowing through things... In that loop, it was about 100 microseconds. Going 30 mph, that's about 60 thousands of an inch. Meanwhile, if you were listening to the starting gun a mile away, that would be about 5 seconds, or about 200 feet at 30 mph. That's only about 4-5 orders of magnitude wrong, so maybe I should've cut him some slack.

Meanwhile, if I screw up my budget estimates by a factor of just 10 people get very bent out of shape.

Back to your post, I'd add that a person's understanding and perception of time is one of the reasons why we're generally so bad at managing it. Even for someone that understands the meaning of the numbers, things with 9 zeros to the left or right of the decimal can be hard to wrap your head around.
 
It matters not what the difference between GPS's idea of time and the official time standard. It only matters that the GPS constellation itself is all operating on the same time standard. The only effect you'll get from a disagreement is that the time in your GPS receiver will not be the official time but the GPS time. Most of our GPS receivers don't display things to the resolution would show such a discrepancy (and not that you'd be able to resolve it if it did).
 
You must log in or register to reply here.
Back
Top