Timeline for Why should a clock be "accurate"?
Current License: CC BY-SA 4.0
11 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Mar 3, 2022 at 20:20 | comment | added | DKNguyen | @JanKanis That's the name! SONET. I had forgotten what it was called and was trying to Google random names like SONUS to talk about it here but never did find it. | |
| Feb 23, 2022 at 14:56 | comment | added | JanKanis | Additionally, the internet (well, the fiber optic communications backbone using the SDH and SONET protocols) requires atomic clocks to synchronize the two endpoints of an optical fiber. So without atomic clocks there won't be NTP, or the internet as we currently know it. | |
| Feb 22, 2022 at 23:29 | comment | added | Peter Cordes | @Peteris: Radio protocols provide ways for clocks to sync with each other, which is why they don't need atomic clocks to hit that 0.5 ms transmit window after running independently for a year or something. You need a clock with some stability, e.g. a quartz crystal, and some logic to figure out a ratio between it and the actual frequency / timing interval other devices are using (a PLL). But that's fine, you don't also need GPS receiver. Sub-millisecond accuracy after seeing something to sync with only a few seconds ago just means everything is happening fast (high frequency). | |
| Feb 22, 2022 at 15:50 | comment | added | fectin | @Peteris NB that there are multiple flavors of GPS, and the one you usually think of (CA) is just to bootstrap into the more accurate versions. As such, it provides MUCH more accurate time than it needs for the coarse positioning we all use for navigation. | |
| Feb 22, 2022 at 15:37 | comment | added | The Photon | @Peteris edited | |
| Feb 22, 2022 at 15:37 | history | edited | The Photon | CC BY-SA 4.0 |
added 24 characters in body
|
| Feb 22, 2022 at 15:34 | comment | added | DKNguyen | @CGCampbell In his defense, it will not matter much longer. | |
| Feb 22, 2022 at 13:20 | comment | added | CGCampbell | @Peteris But why should 1000 ft matter, he asks, as he stares at the cargo ship that just ran aground, or the self-driving vehicle that self-drove-itself off a cliff. | |
| Feb 22, 2022 at 11:50 | comment | added | Peter - Reinstate Monica | Another concern is that for many applications it is important that two measurements even with the same clock are as close as possible. For example during the Olympics, or when measuring the same physical phenomenon repeatedly with an error estimate, etc. | |
| Feb 22, 2022 at 2:30 | comment | added | Peteris | This is a nice answer, but it's probably important to note your examples of required accuracy are multiple orders of magnitude too lax. For GPS a microsecond will make a difference in your results (300 metre/1000 feet difference), a millisecond offset means your distence metrics are hundreds of km/miles off. The GPS time system expects to ensure an accuracy of <30 nanoseconds (gps.gov/systems/gps/performance/accuracy). Also, you need much-better-than-millisecond accuracy for many other things e.g. modern cellular comms may need you to hit a specific 0.5ms transmit window. | |
| Feb 21, 2022 at 17:50 | history | answered | The Photon | CC BY-SA 4.0 |