[time-nuts] Timing Distribution in Mountainous Terrain

[Ralph Smith]

On Thu, September 9, 2010 3:17 pm, John Anderson wrote:
> Hmmm...I design such timing systems for Moto data radios, and 30nS sync is
> going to be very hard to achieve in reality.  Over a few hundred miles
> you're going to have OTA time of flight issues, temperature dependencies,
> etc.  Over the years this has been tried, usually with dismal success in
> the reliability department.  You might get the system working in the
> summer, and then when winter hits and snow builds up on the GPS antennas
> your network goes down.  Or, spring hit with a lot of atmospheric
> turbulance and you get all sorts of reflection / refraction effects, and
> trying to sync up time slots that close will be almost impossible.  It
> might work oneweek, and not the next.

We are not syncing time slots for communications. The timing requirement
is for determining aircraft position by multilateration. Timing errors
translate into position uncertainty.

> What's your carrier freq?  In mountainous regions you'll probably have
> better luck at the lower end.  Please tell me you're not trying this at
> GHz freqs at higher, rocky elevations. 

Signal emitted by the aircraft is 1090 MHz, pulse-position modulated
transponder squitter (Mode C/Mode S). Radio stations receive signal,
determine time of reception, and forward the timestamped transmission via
landline to a central facility for correlation and position determination.

> You might find +-1uSec is a good number to shoot for is cost and
> reliability is a concern.  Just a suggestion.

1 us = 1000 ft, which is more uncertainty than they want to deal with. I
don't make the requirements, we just have to live with them. We definitely
do provide feedback on cost and reliability to those that do dream up the
requirements.

Ralph