Now there’s no noticeable delay between when they update.
Here’s the process I used https://alexsci.com/blog/preparing-to-change-the-clocks/
Now there’s no noticeable delay between when they update.
Here’s the process I used https://alexsci.com/blog/preparing-to-change-the-clocks/
I believe that phones can get time from the cell tower or GPS, and both are probably more-accurate than NTP, since they don’t have to try to compensate for network latency.
NTP is accurate within a few milliseconds. GPS is accurate within a few microseconds. However, you need to hook into the PPS (pulse per second) on the GPS signal to get proper accuracy. Think of it like a super-precise drummer keeping beat. NTP servers typically use GPS as a primary time source.
Source: I operate a stratum-1 time server in the NTP pool.
I have a GPS app on my phone, and when I run it, I can see the difference between the GPS time and phone time quite clearly
searches
For Android:
https://source.android.com/docs/core/connect/time
“Network” will presumably be NTP, “Telephony” the cell tower time, “GNSS” GPS, and I dunno what “External” is — maybe there’s some way to have a USB-connected time device or something.
This is all good stuff, maybe at some point I will get back into some Android development and make an app that exposes some of the underlying NTP configuration. Really, I just want to see the ntpd (or chrony) config file (if it exists) and tweak it.
I did some searching of my own today, and found this guy’s blog
https://jitinsharma.com/posts/adventures-in-android-timekeeping/
And he seems to think the poll interval is 16, or 18 hours. I don’t know why these poll intervals are so long. I bet even 2^10 (17 mins) would work much better across all phones, and not really consume that much bandwidth.
And also, if the phone has access to time from the cell network, why can’t it use that more often?
This is the first phone I have had that is regularly more than a second off, and it bugs me…