16
$\begingroup$

I have been looking at different threads. It seems like the receiver clock on Earth doesn't have to be synchronized with the satellite clocks.

But I believe the satellite clocks need to be synchronized with each other in the earth's frame of reference.

How is this achieved?

$\endgroup$

2 Answers 2

13
$\begingroup$

The satellites' location is very accurately determined using ranging techniques, where a pulse is sent from a ground station to the satellite and the satellite responds very quickly, with a known time between receive/transmit. This allows one to determine exactly how far away the satellite is. Combining this information with orbital tracks allows one to determine exactly where the satellite is.

The amount of time required to do a time update from the time the signal is sent to the time it is received and in use is also well known. What is actually sent to the satellite is the time deltas, to tell the clock to jump forward or backward a certain amount of time. For GPS, this update can occur every 8 hours, but in practice is done every 24 hours.

The satellites each have onboard atomic clocks, for very accurately determining the time. Relativity is managed in essence by correcting for the clock based on the expected time. They slowed the clocks down to correct for this effect.

Signals are sent up to the satellite via the GPS ground network.

The time standard used for GPS clocks is a whole number of seconds from UTC. GPS time does not account for leap seconds, thus the whole number of seconds time difference.

$\endgroup$
12
  • 3
    $\begingroup$ Are you sure that is how clock correction commands to the satellites are formated? If I were building such a system it would feel a lot more robust if the commands were not, "set your clock to such-and-such NOW!", but instead "your clock is running 13 nanoseconds slow, please correct". Then the time it takes to transmit and decode the command itself would not be critical anymore. $\endgroup$ Commented Mar 7, 2016 at 18:17
  • 1
    $\begingroup$ @AmeetSharma: Added the comment to address your concern. The answer is every 8-24 hours. Relativity, while a concern, isn't as much of a concern because all of the satellites orbit at the same altitude. $\endgroup$
    – PearsonArtPhoto
    Commented Mar 7, 2016 at 19:58
  • 1
    $\begingroup$ It was, in the design of the satellites. $\endgroup$
    – PearsonArtPhoto
    Commented Mar 7, 2016 at 22:07
  • 1
    $\begingroup$ @HenningMakholm I think you have a misconception. The clocks are not "set" or "corrected". The clocks just keep running and no counts are changed. What is sent up to the satellites and periodically updated, purely to turn around and send to receivers on the ground, are files that provide the data needed to convert the clock time to the real time, and how to calculate the position of the satellite. "Subframe 1 contains the GPS date (week number) and information to correct the satellite's time to GPS time, plus satellite status and health." $\endgroup$
    – Mark Adler
    Commented Mar 8, 2016 at 1:08
  • 1
    $\begingroup$ @MarkAdler: That makes more engineering sense, yes (especially since it would be difficult to jump the LFSRs creating the various modulating bitstreams backwards and forwards on demand). Note that I was responding to the original version of the answer which explicitly claimed that clock correction commands have to be precisely timed such that the new setting would take effect that the right time. $\endgroup$ Commented Mar 8, 2016 at 12:12
8
$\begingroup$

Receiver clock synchronisation:

GPS or GNSS applications are only dependend on the travel time of the signal. To put it simply: You only need the travel time of four GNSS satellites to calculate the position. Three satellites for determination of the position (x,y,z) and one for the receiver clock error. This clock error is the synchronisation between the receiver (e.g. car navigation) and the GNSS satellites.

Satellite clock synchronisation:

The satellite clocks are synchronised regularly by an associated ground station of the provider.
For example Galileo: There is a ground station (Galileo Mission system) and there are several Sensor stations worldwide.

The GMS ( Galileo Mission System ) will use the GSS ( Galileo Sensor Stations ) network in two independent ways. The first is the Orbitography Determination and Time Synchronisation (OD&TS) function, which will provide batch processing every ten minutes of all the observations of all satellites over an extended period and calculates the precise orbit and clock offset of each satellite, including a forecast of predicted variations (SISA - Signal-in-Space Accuracy) valid for the next hours. The results of these computations for each satellite will be up-loaded into that satellite nominally every 100 minutes using a scheduled contact via a Mission Up-link Station

Stolen here

Best regards Ben

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.