Then (1881) and Now (1979), first person account.

[rdamon]

I used to listen to WWV as a kid.  :)

[Paul of Covington]

rdamon

I used to listen to WWV as a kid.  :)

 

   Must have been boring after a couple of hours.

   But seriously, in my teens I used to like working on clocks, and I used to use WWV to set and adjust them.

[CMStPnP]

BaltACD

My Android phone gets its time automatically - I don't know its source, however, it seems to be on the same program as my W10 computer.

 

Not exactly sure how cell phones work it but office phones and the like get their time display from the phone PBX which should be synchronized to the local Central Office Switch which the PBX is a child of.    Phone switches are basically Unix or Linux server boxes.

[Erik_Mag]

The digital modulation schemes used by cell phones require very tight synchronization between the towers. Each one has a GPS receiver to provide a very stable time and frequency reference. In addition, the phones themselves have to be synchronized to the towers, so part of the handshaking between the tower and phone would be time information.

Thirty years ago, the best way of getting accurate time and frequency references was via Loran-C, where the receivers would use the phase of the 100kHz signal to extract sub microsecond timing information.

[Overmod]

Erik_Mag

The digital modulation schemes used by cell phones require very tight synchronization between the towers. Each one has a GPS receiver to provide a very stable time and frequency reference. In addition, the phones themselves have to be synchronized to the towers, so part of the handshaking between the tower and phone would be time information.

In the original analog cellular radio, handoff between cells required some tracking of location of the various devices.  As I recall this was done by sending and receiving a coded signal which would give TOF, and having phased antennas that would give angle.  That is how the original 'phone locator' system was supposed to work.

Thirty years ago, the best way of getting accurate time and frequency references was via Loran-C, where the receivers would use the phase of the 100kHz signal to extract sub-microsecond timing information.

And of course that timing precision needed to be provided via something, which I recall being three cesium clocks per station, which was synced to CUT and all the other stations within 100ns.  I think in practice this was slipped about 300% but is still pretty good...

Remember to put this in perspective that the speed of light is essentially a foot per nanosecond.

 

[Erik_Mag]

Position error with Loran-C is about a half foot per nanosecond, as moving a half foot from station A to station B causes a half nanosecond delay from station A and a half nanosecond advance from station B. At 100kHz, the propagation is by surface wave so changes in the ionosphere do not have a major affect on the signals.

My learning about Loran-C for timing was to come up with a frequency reference for a proposed Josephson junction voltage standard.