[time-nuts] Re: constructing a moon base

[Lux, Jim]

On 9/29/21 7:30 AM, Chris Howard wrote:
> My employment involves the design and manufacturing of construction 
> equipment.
>
> I had a conversation today about what will be needed for 
> time/positioning when GPS is not in sight, like road-building on the 
> far side of the moon.
> The context was timestamp coordination between multiple data sources 
> in the vehicle environment, ntpd, and similar things.
>
> No GPS probably means cesium on the road grader? 


There's been a lot of research on PNT (Position, Nav, Timing) for lunar 
applications over the last 20 or so years. Google "Lunar PNT" as a start.

Typically, the proposed approaches use fixed monuments (pseudolites) 
which are disciplined by reference to some orbiting source, and then 
your moving things use fairly conventional GPS-like strategies.  There 
are the same discussions about two way vs one way, smart receivers vs 
smart transmitters, etc. that were all hashed out in the early days of GPS.

A complication on the Moon is that the visible horizon is pretty close 
(~2km for something a couple meters high), so the fixed stations need to 
be relatively close together. In a "road building" scenario, you can 
push the survey net forward as you build the roads, just as with 
conventional surveying.  I will note that we built the transcontinental 
railroad in the US, as well as most roads, with conventional surveying, 
and it works just the same on the Moon.   So "relative position" is 
something you don't need clocks for (in general).  It's when it comes to 
timing that satellites are handy for time transfer.

If you're on the near-side of the Moon, then you can receive Earth GNSS 
signals with a suitably large antenna. Or, if you're creating a system 
from scratch, you can set up a ground based transmitter to allow fairly 
small antennas (10s of cm) at the Moon - after all, a fairly small 
antenna (3m) on Earth will illuminate the Moon and surroundings within 
its beamwidth (1/2 -1 degree, depending on frequency), so you don't need 
much Tx power (and Tx power on Earth is a LOT cheaper than in space or 
the Moon).  I seem to recall 50-100 Watts is more than enough, with a 3m 
dish, at X-band (7 GHz) to get nanosecond timing at the Moon with an 
Omni receive antenna.

The challenge with "position" at the Moon using GNSS is that the 
geometry is terrible.  The Moon is 300,000 km away, so the GNSS 
constellation is < 10 degrees across. Talk about Geometric Dilution of 
Precision (GDOP).  That's why most schemes have some orbiter(s) to get 
the geometry better, and then periodically update the clocks on the 
orbiters with reference to an Earth source.  So that puts a constraint 
on the Allan Deviation of the orbiting clock, because it has to ride 
through the hour or so when it's out of view.

Anyway, lots of interesting stuff.  I've been involved with this at work 
for at least 10 years, so feel free to ask questions.