[time-nuts] Raw GPS signal samples

[Magnus Danielson]

Tom,

On 09/04/2013 11:05 PM, Tom Van Baak wrote:
> Hi John,
>
> It's not off-topic at all. Thanks for posting. Someday it would be cool if an amateur precise time project emerged from this effort.
>
> What is not clear from your link is the accuracy or stability of the LO that was used. It seems to me, for timing purposes, that one would want to have a very precise timebase for the binary samples that are collected.
>
> I suppose it depends if someone just wants to make a homebrew (50 baud) GPS subcode decoder and navigation receiver, or if someone is interested in extracting as much precise time from the code and carrier phase as possible. For our timing niche, SDGPS (Software-Defined GPS) doesn't even need to be real-time. It's the after-the-fact corrections to the LO that are of interest.
Naturally the LO and filtering is important. There is a bit more than
the 50 baud data you need to grab, but once you have your observables,
just putting them onto a file works. However, doing fair amount of
processing on that stream of data isn't very daunting. Getting down to 1
kHz is a bit cycle-consuming, so a good front-end processing for that is
a good idea.

I've naturally toyed with grabbing a gig or so of samples from the air
and put them on file, and then post-processing them, locking up to birds
and extracting data.

Cheers,
Magnus