[LEAPSECS] UT1 offset

[Jim Lux]

Most spacecraft don’t know any sort of time in an absolute time scale.  They have a free running counter at some rate, and everything is done in terms of SCLK. (Spacecraft Clock).  Someone on the ground does a process called “time correlation” to relate local clock on spacecraft to some other system (TAI, UTC are popular).  You send commands up with a “do it at this time” in terms of the local clock ticks, and telemetry comes down time stamped with the local clock. That may get turned into TAI or UTC on the ground, or even if you’re working in TAI on board, there’s some stuff required to deal with the jump from “no fix” to “have fix”.  Once you have a fix, if the GPS solution goes away, you just flywheel on forward, and either have a transient when GPS comes back or try to smoothly pick it up.

Recently, (last 20 years) with the advent of inexpensive GPS receivers, a lot of cubesats use them in some way, but a lot don’t actually use GPS time on board - after all, you need to have some time scale to use before you’ve got a GPS fix. That GPS time (and position) would probably be reported in telemetry.  It might also be fed to your attitude control and determination subsystem, so it can be used with the star tracker to determine attitude and position, which then gets propagated forward.  But that’s a fairly low precision need (1 second), and there have been plenty of satellites where the GPS receiver failed, and they just uplinked the “current GPS time” to replace it.

For satellites that are making precision measurements (e.g. radio occultation with GPS) the GPS onboard data is just sent to the ground as raw observables, often without ever trying to get a solution on board. The manipulation to some Earth based time scale is all done on the ground in post processing.

It’s pretty unusual for there to be a need for time knowledge on board in a standard time scale to “better than a second” kinds of precision.  My own SunRISE mission needs to know time to a microsecond onboard, but it doesn’t have to be any particular time scale, just the same across all six spacecraft. It happens to be GPS time.  We do determine time in post processing to a bit better than a nanosecond. 

This is actually becoming more important - if you want spacecraft autonomy, AND you want multiple spacecraft working together, having them all on a common time scale is important.  But it’s still sort of done in an ad hoc way - that is, while there’s plenty of CCSDS standards on how to report and keep time, the implementations vary.


Sent from my iPad

> On Jan 3, 2024, at 6:28 AM, Seaman, Robert Lewis - (rseaman) <rseaman at arizona.edu> wrote:
> 
> 
> Hi Tom and Mike and all,
>  
> I suppose we weren’t talking about DUT1 time signals?
>  
> See http://futureofutc.org/2011/program/presentations/AAS_11-675_Malys.pptx.pdf for details about the flipside question of operating a GNSS constellation (current as of a dozen years ago).
>  
> One shouldn’t find it surprising (at least, I don’t find it surprising) if navigating and calibrating constellations of Earth-orbiting satellites requires knowledge of Earth orientation. At some point during the 2011 Exton workshop, there was a discussion of GPS being able to detect motions due to plate tectonics. Earth orientation doesn’t necessarily need to be provided in terms of UT1, and the temporal geophysicists presumably need higher-order moments as well. One doubts the majority of satellites need such precision.
>  
> Rob
>  
> On 1/2/24, 7:45 AM, "LEAPSECS" wrote:
>  
> External Email
> 
> Hi Mike,
> 
> > the system needs an estimate of current UT1
> 
> Can you give some references to your observation? I don't recall seeing UT1 mentioned in the first couple of decades of GPS documentation. The system runs on GPS time, the WGS84 coordinate system, broadcast ephemeris including SV clock corrections. Where does UT1 appear in those?
> 
> > That estimate is applied internally so the end user does not need to know the details
> 
> Right, the user is shielded from many details. But I didn't think even GPS receivers had knowledge of UT1, nor the satellites themselves. So where in "the system" does UT1 apply?
> 
> Thanks,
> /tvb
> 
>  
> On 12/28/2023 1:23 AM, Mike Hapgood - STFC UKRI via LEAPSECS wrote:
> Jim outlines a calculation I've done many times. But there's a similar calculation for GNSS systems (GPS, Galileo, Beidou, etc). If you want to use GNSS to determine positions on Earth's surface to accuracy of a few metres, the system needs an estimate of current UT1 accurate at least to a few milliseconds. That estimate is applied internally so the end user does not need to know the details, just as that user does not need to know about the relativistic clock corrections or corrections for ionospheric signal delay that also underpin safe use of GPS. But the bottom line is that knowledge of UT1 (i.e. the spin phase of the Earth) is essential for GNSS - and many other space systems.
>  
> Mike
>  
> _______________________________________________
> LEAPSECS mailing list
> LEAPSECS at leapsecond.com
> https://pairlist6.pair.net/mailman/listinfo/leapsecs
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pairlist6.pair.net/pipermail/leapsecs/attachments/20240103/57b1d04a/attachment.htm>