[time-nuts] UTC and the speed of light?

[Tom Van Baak]

> The counter of the frequency is subject to phase shift caused by the
> speed of light and the distance to the measured source. Each counter
> must be given an initial value of time in order to be called a clock.

Hi Bill,

When not in motion it's easy to synchronize remote clocks --
you just send pulses back and forth and split the difference.
It doesn't matter what the speed or distance is.

But on a rotating planet there is also a clock correction for the
Sagnac effect. It's not something you need to apply in your home
lab but it does affect global time transfer, including GPS signals
(the corrections are automatically made in your receiver), any
precision signals sent east or west, or even clocks transported
east or west.

Try this. At the equator you send a pulse to yourself around the
earth, eastward. Given a circumference of about 40,000 km it
should take 40000km / 299792.458m/s = 0.133 seconds to go
all around and get back to your clock and time interval counter.

But Earth's rotation speed (equator) is 40000km / 1day = 460
meters/second (that's just over 1000 mph) and so during those
many 133 milliseconds that the pulse is flying east, the earth is
rotating 460m/s * 0.133s = 60 more meters. Not fair; a moving
target! So it will take an additional 60m / 299792458m/s = 200 ns
for the pulse to finally get back to your clock/counter.

Similarly, if you send a pulse to yourself westward it will arrive
some 200 ns sooner than you expect.

Note that as long as you take the Sagnac effect into account
(knowing the lat/lon of the clocks and signal path) you can still
use simple split-the-difference clock synchronization.

/tvb