[time-nuts] Followup (still want a GPS-type NTP refclock)

[Tom Van Baak]

> OK, unless you have the coordinates of your antenna position... and here
> comes the difficult move: how can I have the coordinates of an indoor
> antenna that can't receive the required satellites? Starting from a
> suitable outdoor position then using precise length and angle measurements,
> yes, it can be done.

Azelio,

Assuming this thread is still about NTP -- you wouldn't need an accurate indoor position; any nearby outdoor position would work fine. Use google maps, for example. Realize that an error of a couple tens or even a hundred meters will not affect NTP performance.

> Take care that the 1 satellite timing mode comes after having seens more
> than 4 satellites for at least 10000 seconds (usually, or greater). You
> can't start a timing mode receiver with 1 satellite.

Note that all the original NBS GPS receivers were single channel receivers. The beauty of using satellites for time transfer is that you only need one. Sure, more is usually better, but one is sufficient. Even today many labs use GPS common view, where the position is hard-coded and timing is derived from one SV alone, or each SV independently, or a post-processed combination of all SV in view.

I say all this to emphasize the difference between satellite(s) for navigation and for time transfer.

Another timely analogy for you is Galileo. GIOVE-A was launched in 2005 and several years later GIOVE-B. The first two real SV in the system were launched just a week ago! So it will still be some time before you can use Galileo for car navigation, but since even the first SV was in orbit it can be used for time transfer. BTW, here's a great opportunity for one of you EU-based time-nuts -- be the first to report some timing from the first two Galileo satellites.

/tvb