On Apr 10, 2010, at 11:31 AM, Eric Garner wrote:
My question is why is it done that way at all? It seems impractical
in the extreme. Why not use a meter? Or the afore mentioned
vibrating reeds?
What happens when the clock rates sync up again, since the wall
clock time will still be ahead/behind the sat clock? Do they reset
the wall clock? Or is it only the rate that matters?
Given the sophistication of the rest of the network there must be a
reason.
They want our synchronous wall clocks to always be within a
reasonable time to accurate. So they use one such clock to act as a
surrogate.
If the clocks have slowed down due to a drop in frequency, when the
load goes away they need to run a bit faster to catch up.
In the US, at one time there was a specification that required the
average over a (I think 24 hours) to be within three seconds. It
always was when I was growing up in the 60's and 70's. Now I've seen
it get as much as 4 seconds (once 5) out.
They're providing a service (at least they were...today, I wonder if
they care?).
Tom Frank, KA2CDK
_______________________________________________
time-nuts mailing list -- [email protected]
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
