[time-nuts] Frequency based on mobile services
magnus at rubidium.dyndns.org
Tue Jan 25 19:39:16 UTC 2011
On 25/01/11 18:12, Bob Camp wrote:
> Several of the "big guys" have tried this approach.
> The problem seems to be that you are at the mercy of the guy who is
> transmitting the signal. He may (or may not) be very good at keeping things
> running right. Regulations may (or may not) actually require him to lock his
> signal to GPS.
> A lot of these systems work just fine at> 1.0 x 10^-8 off frequency. You
> have to keep the slots lined up between them, but as long as everybody
> agrees on alignment, they keep running fine. There's a lot that you can get
> away with and still be providing a dial tone.
GSM and hence UMTS has 50 ppb requirements in the base station
frequency. For GSM no timing is required, but synchronisation of several
base-stations has been shown to improve hand-over performance.
Depending on the operator you tune in to, it may or may not be locked to
GPS or similar... but may also run from caesium beams somewhere
completely different, if so you are only guaranteed 1E-11 in frequency
There is many systems, and you also need to tune in to the right signal.
GSM for instance may frequency jump regularly.
Digital TV signals may be also be used. Depending on which mode the
network operate in the transmitters may be undisciplined or locked to
GPS directly or indirectly. DVB-T for instance has regular pilot-tones
in the spectrum. A DVB-T receiver is a nice exercise... :)
It's possible, but not trivial.
More information about the time-nuts