[time-nuts] Spacecraft Timekeeping

[jimlux]

On 3/9/11 10:58 AM, Kevin Watson wrote:
> Hi Magnus,
>
> As I said in an earlier message, this is an experiment that I want to
> run and would rather not touch mission and safety-critical GNC
> components, like our navigation GPS receivers. Mass is not an issue.
>
> -Kevin

So this makes it pretty simple..

You have a GPS receiver that operates on the pad to get "initial sync" 
(and perhaps to calibrate your onboard oscillator... you don't actually 
need to discipline it, just have "knowledge")

Then, you launch, and do your timestamping with the oscillator...

When you get on orbit, and your GPS reacquires, you can recalibrate or 
resync.

For that matter, what you could do is very simple.. run your oscillator 
and run a counter.  Snapshot the value of the counter every time you get 
a 1pps tick from the GPS.  Log the time/position/status messages from 
the GPS as well.

You can then fairly easily post process to back-out the variations in 
your oscillator (which will be small) and recalibrate your timestamps to 
UTC.

Now.. if you need to have different *spacecraft* be synchronized, the 
post processing scheme won't work.


A separate issue is the "onboard distribution of time" to 10 
microseconds.  PTP over Ethernet, SpaceWire Timecodes, or IRIG on a wire 
would meet that easily.