[time-nuts] Fundamental limits on performance

[Magnus Danielson]

Lux, Jim (337C) wrote:
> Precisely... The GPS satellites don't transfer between each other, but from
> the satellite to earth.. But get out past the moon's orbit, and you can't
> "see" the GPS signals anymore.

Actually, they do have that capability to some degree. They use 450 MHz 
Yagi links for inter-sat communication links and pseudorandom ranging. 
However, it comes to little use in practice.

> We do a lot of science with sending very quiet (low phase noise) signals
> from a probe back to earth: e.g. You can do occultations to look at the
> atmospheric structure of moons or the planet, you can do precise gravity
> measurements (GRACE and GRAIL, for instance).  The problem is that (as time
> nuts are aware) a quiet, stable, low power, reliable frequency reference
> isn't easy to come by.  JPL is working on a 1 kilo/1 liter trapped Hg ion
> clock, but that's still in the future, it's big and heavy, and consumes
> power. 

Out of curiosity, what kind of requirements do you guys have? What kind 
of environmental aspects is there? What kind of physical limits would be 
typical?

> So, if I have an orbiter with a good clock around Mars, and I want to
> discipline a bunch of cheaper oscillators on the surface (or perhaps
> orbiters), what is the fundamental limit on how good you can do, given the
> radio link available from orbiter to ground station.

That depends. You can afford doing bi-directional ranging, as you have 
fairly low amount of space and mars surface nodes. The benefit would be 
that the surface nodes has high stability in position but not as stable 
in longterm, while the space nodes can provide frequency stability. 
Pseudo-ranging aids in orbit tracking and the relative position of the 
surface nodes can be established. Additional space nodes can use the 
resulting pseudolite-satelite constellation for tracking of orbit and 
landing position.

I suspect that the weak atmosphere of mars does not call for as advanced 
correction of delays as here on earth, but that could become a research 
field in itself by using two or three frequency rangings.

The key to GPS ability to handle low signal strength is the synchronous 
modulation and ranging codes that helps to decorrelate noise. The use of 
coarse and fine ranging codes and use of modern FFT correlation methods 
would allow quick lockins. The data-modulation is synchronous to the 
carrier and ranging codes and of low rate. Another benefit is the 
repeating aspect of the data, which allows even deeper correlation gains 
when data is know and integration bins can range over several symbols.

The surface nodes would benefit from using smaller rubidiums for 
long-term stability.

For such a system would the satellite orbits be of great importance, as 
always. Inclined orbits with sufficient heigth to allow common mode 
views between surface nodes as well as multiple observation angles of a 
sat from surface nodes, just as we can do for GPS. Onces surface nodes 
can observe the sats, orbit stabilization can be performed.

Cheers,
Magnus