Slashdot Mirror
NIST's New Atomic Clock Is So Precise Our Ability To Measure Gravity Constrains Its Accuracy (vice.com)
dmoberhaus writes: Researchers at the National Institute for Standards and Technology (NIST) have developed an atomic clock that is so precise that our models of Earth's gravity aren't accurate enough to keep up with it. As detailed in a paper published this week in Nature, the atomic clock could pave the way for creating an unprecedented map of the way the Earth's gravity distorts spacetime and even shed light on the development of the early universe. "The level of clock performance being reported is such that we don't actually know how to account for it well enough to support the level of performance the clock achieves," Andrew Ludlow, a physicist at NIST and the project lead on the organization's new atomic clock, told me on the phone. "Right now the state of the art techniques aren't quite good enough so we're limited by how well we understand gravity on different parts of the Earth."
So now it's device that can measure not just time, but gravity too.
If it weren't for deadlines, nothing would be late.
Sure internet lag, and performance of our PC especially in terms of the time clock wouldn't matter... But darn it, I want my PC to have accurate time.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
He'll figure it out.
More like his daughter will.
If it weren't for deadlines, nothing would be late.
Because I want my clock to be accurate.
It is like having speakers that go to 11
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
Since you can measure gravity extremely accurately, and gravity varies with height, does that mean they've just invented an extremely accurate altimeter?
If you pay us a couple million a year, we probably won't create any black holes around it. Just insurance ya know? /s
Because I want my clock to be accurate.
It's a security feature. It allows you to require krb5_deltat < .000000000000000001
#DeleteChrome
And behind Door #2 time is running out for WWVB.
The low frequency WWVB standard and short wave clock time standards seem have time running
out for them.
https://www.voanews.com/a/time...
It may simply be that we will know with more precision when infrastructure has its plug pulled.
GPS time is likely better than NTP time for computers.
Clocks like this may allow for the elimination of almost all Olympic timing errors and ties.
I can see headlines... runners fail to best Usain Bolt's best time by one Picosecond +/- 2.7 Femtoseconds.
Truth is stranger than fiction, but it is because Fiction is obliged to stick to possibilities; Truth isn't. Mark Twain.
So get cracking on that spacetime gravity problem!
You can get time accurate to within about a thousand to a millionth of a second with PTP, using network time as your reference. That's the time standard used by thr vast majority of people, network time aka ntp time aka internet time, which is closely synced to GPS time.
So which time do you want to consider "correct" - the time used by 98% of the precision clocks, or the time used by one clock in Boulder?
NTP is super easy to use and pretty darn accurate as well. PTP is quite a bit more accurate, with microsecond synchronization within your network.
But darn it, I want my PC to have accurate time.
The optical lattice clocks they're talking about "would take longer than the age of the universe (13.8 billion years) to lose a second." (from: Scientists Build Atomic Clocks Accurate Enough to Measure Changes in Spacetime Itself). If you think your PC will last that long, I want to know more about it than the clocks... :-)
It must have been something you assimilated. . . .
But darn it, I want my PC to have accurate time.
You can buy an atomic clock on a chip, such as the SA.45S from microsemi.
A dev kit version with serial output runs just under a thousand bucks for quantities of one, and would give you a legit strata-0 time source.
It's not as accurate as this new atomic clock NIST has, this one will drift by 10^-11 seconds per day, but it's far far cheaper.
The accuracy would only be valid for its own time frame, though. As soon as you need to send the time anywhere else, and you have to, even if it's just to a circuit to read the time from, the accuracy depends on stability of the environment. Any kind of change in acceleration, including but not limited to gravity, will cause a cumulative difference over, well, time, for lack of a better word.
Someone storing large amounts of metal in a nearby building, or Boring a tunnel underneath might be enough to cause drift that must be compensated for. To say nothing of factors like tidal forces, and even how the earth's own orbital speed decay over the aeons cause time variations.
But it's darn accurate for its own time frame.
when worn by obese or anorectic people :-)
I have good news and bad news for you.
The offset value we log isn't the actual difference between your clock and the peer clock, and the logged value is seconds, not milliseconds. If you see 0.08 in the log, that means we're *assuming* that you're about 0.08 *seconds* off from that peer.
If you ping a server "thousands of miles away", you'll notice each ping takes a different number of milliseconds. That difference in ping times, or jitter, is pretty much a hard limit on sync accuracy with the NTP protocol. If you get a response from a server thousands of miles away saying "it's exactly midnight", we don't know if that packet was generated 30ms ago or 35ms ago. So that loses 5ms of precision.
Back to your peerlog. The last four fields show the assumed offset, delay, dispersion and jitter (root mean square). Take delay, multiply by 0.25, then add the jitter, and that's roughly the accuracy you can expect. If you're using ntpclient, add the dispersion, as that's the dispersion of the server. If you're using ntpd, your dispersion value is arough estimate of your accuracy - in seconds.
The good news is, your ISP's DNS server and gateway are many milliseconds closer, and ine or both are probably NTP servers. You CAN get much more accurate time, you just need to use servers that are much, much network-closer to you, such as your ISP's DNS or gateway.
The reason for choosing the closest server may merit more explanation. If you're on the west coast and you query a server on thr east coast, your accuracy suffers with every router and switch along that path, so you end up with some unknown offset from the NTP server's time. ISPs and geographically disperse companies don't have that problem. Here's why.
Comcast will have a DNS server / ntp server in Houston. It may have a GPS, tdma, or LTE reciever locally connected, but most importantly it peers with their servers I Miami, Denver, and San Jose. The one in Miami peers back with Houston - and Miami, PLUS San Jose and a tier 1. All of the Comcast NTP servers have to estimate the time sync, but *they all arrive at the same estimate*, because they peer with each other multilaterally. So they all end up with the same time, network time.
By getting your time from a nearby server that does mutual peering, you're getting network time from a server that is only a few milliseconds away, amd probably a *consistent* number of milliseconds, so ntp can make accurate estimates of delay.
Imagine if you had a array of these clocks spread out in a 2d grid, level with each other to within less than a millimeter, It is possible some would run faster or slower than the others because of subterranean variations in density, and thus slightly different amounts of gravity.
Interesting stuff, thanks for sharing.
Not quite. How do you set time of day? It's only a frequency reference, not a time reference. For time references, there's really only GNSS, dialup time services like ACTS and the various radio services (and maybe Iridium). The value of the CSAC, synchronised with GNSS, is that it will give you better holdover than the very ordinary crystal oscillator in your PC if GNSS is unavailable.
As for the frequency offset, the CSACs 'age,' that is their frequency offset increases with time, so that the time offset increases quadratically.
I know you were just joking but ... it's not a clock. It's a frequency reference. A clock needs to run continuously and the state of the art 'clocks' like lattice clocks are not reliable enough to do this yet. Very good frequency references can be used to correct the frequency of less accurate clocks and this is what happens in the computation of Coordinated Universal Time. Time of day is computed from 400 or so continuously running hydrogen masers and caesium beam standards and then a very small correction is made using 'one second' as calculated from half a dozen or so extremely good frequency standards like caesium fountains.
Now we will FINALLY know what time it really is!
A man with a watch can tell you what time it is.
A man with two watches is never really quite sure.
This new clock demonstrates a sort of high-tech version of this problem. Two of these clocks might not agree because they are in different locations, where gravity is stronger or weaker. At least in this case, the clock can help solve its own problem by helping refine the map that it needs to be more accurate.