[time-nuts] Terrestrial Tides and Land Movement

[Tom Van Baak]

> How is this achieved? Is there a coupled dead-reckoning system that updates the timing location, or something else?

Hi BobS,

No need for dead-reckoning -- GPS gives you the location. And the time.

Take a step back and remember how GPS works. You have sats flying above that essentially send coordinated time down. The receiver solves N equations in 4 variables (X,Y,Z,T). The solution is 3D position and time. In general, the larger N is (the number of sats received) the more accurate the position and time calculation, by sqrt(N). GPS has always worked this way, whether the receiver is moving or stationary.

In fact, "stationary" is a bit misleading. You may think of your home as 0 MPH and your car as 60 MPH or a plane at 500 MPH, but the sats themselves are moving at 8700 MPH and this motion is relative. Since the equations are already working doppler frequencies and velocities of thousands of miles an hour there's not much difference between what you call a stationary receiver and a moving receiver.

What you are probably thinking about is the "zero-D", or "overdetermined solution" timing trick that was created in the 90's(?). The idea is simply that *if* you already know X,Y,Z then the receiver only has to solve N equations in *1* variable (T). This gives you a more accurate time than in the 3D case. The other advantage is that the receiver will still output time even if N drops below 4.

Many of the cheap GPS/1PPS receivers we use these days (Sparkfun, Adafruit, Parallax) don't even bother with a 0-D mode. So theoretically they should work just as well in a car as at home.

Does anyone on the list already have 0D vs. 3D timing data for our favorite receivers like M12, ublox6, TBolt, etc.? As long as the survey position is ideal, we would expect 0D to outperform 3D. The question is, by how much.

/tvb