[time-nuts] Question re neutrinos and GPS

[Tom Van Baak]

> The idea that Qz (time on the quartz clock, no?) drops out in
> the subtraction seems to me to require Qz to be invariant.

Hi Bill,

You might be making this more complicated intended. Antonio
was asking about *resynchronizing* clocks.

Here's another example. You're at home at 7:45 AM and see
that your wristwatch is 7 minutes ahead of your kitchen clock.

You get to work at 8:30 AM and notice that your wristwatch is
5 minutes ahead of your office clock. What conclusions can
you reasonably make from this?

Unless your wristwatch is really sick, the simple conclusion is
that your home and work clocks differ by 2 minutes. You have
used your wristwatch for time transfer.

But now here's the good part. Notice that it doesn't matter what
time you actually left home or what time you got to work. It also
doesn't matter if your wristwatch is ahead or behind or spot on.
Your wristwatch does not have to be accurate; yet it was still
usable for remote clock comparison.

Notice also that you did not have to synchronize your wristwatch
at home or resynchronize it at work. All you needed was two
measurements; the 2 minute conclusion is the difference of two
time differences.

So that's the point I was making. The actual time displayed by
the portable clock drops out of the calculation. You get a nice
time interval measurement without having to worry about time.
This is the beauty of differences of differences. It's why ADEV
is immune from phase or frequency offsets, etc.

So remote clock comparison using a portable clock works even
if none of the three clocks is accurate. Note the accuracy of the
measurement (think error bars) is a function of their stability; in
this example, the TDEV at tau 45 minutes.

/tvb