[time-nuts] Spacecraft Timekeeping

[jimlux]

On 3/8/11 9:57 AM, Kevin Watson wrote:
> Hi Jim,
>
> As part of my research into keeping time on rockets and spacecraft, I
> joined
> this list to see what I could learn from the masters. Of course I'm a
> knuckle-head for not assuming that you'd be one of the resident masters
> <grin>. Anyway, as my accuracy needs are modest (~10uS across many onboard
> computers), have access to GPS most of the time and don't really need to
> worry about relativistic effects (yet, anyway <grin>) or radiation effects
> (due to redundancy), I thought I'd use a GPSDO that can handle a decent
> amount of holdover and then use PTP to distribute time across the vehicle.
> Do you, or anyone else, have a recomendation for the GPSDO? Jackson Labs'
> (http://jackson-labs.com/) DROR seems like it might work, but I wonder if
> there might be better alternatives.
>

I don't know that GPS is your big problem, more the holdover when you 
can't get a GPS signal because your antenna isn't pointed in the right 
direction or the high dynamics of the rocket cause it to loose lock.

So, you want sync to 10 microseconds over how long a time span?  1ppm 
would be 10 seconds, or are you looking at minutes or days?


And, do you need to sync to the "outside world" or just within your 
rocket/spacecraft?  That is, as long as "local time" stays synced, so 
that simultaneous events on different boxes are simultaneous, even if 
the local clock drifts, relative to UTC, that's ok.

I assume you're looking for small and relatively inexpensive, as opposed 
to just buying, say, a space qualified Rubidium reference from one of 
the various suppliers.

You also mentioned PTP... so you're interested in "distributing" time 
(or at least synchronization), and not adding extra wires for, e.g. IRIG 
timecode or a 1 pps.  PTP can certainly do the microsecond scale sync 
across Ethernet.  If you're using SpaceWire, the timecodes can get 
microsecond scale sync, as well.