Clocks in relative motion cannot remain synchronized for anylength of time.
in [Relativity]
|
Prev: The Infinitely Large Arch (was Re: Three times happening together)
Next: The problem of Position for massless particles
From: Sam Wormley on 8 Mar 2010 14:36 On 3/8/10 12:19 PM, kenseto wrote: > So how do they make the GPS clocks synchronized with the ground > clocks? See: Relativistic Effects on Satellite Clocks http://relativity.livingreviews.org/open?pubNo=lrr-2003-1&page=node5.html
From: Tom Roberts on 8 Mar 2010 16:39
kenseto wrote: > Clocks in relative motion cannot remain synchronized for any length of > time. This is true. For once, kenseto just happened to get something right. > So how do they make the GPS clocks synchronized with the ground > clocks? > The answer: They redefined the GPS second [...] True to form, kenseto got this wrong, along with the rest of his post. One second in the GPS is the same duration as one second of UTC, when both are measured on earth's geoid (which is where the latter is defined). What they actually did in the GPS is to construct a locally-inertial frame called the Earth-Centered Inertial frame (ECI). In the ECI, time on the geoid coincides with GPS time, which is kept essentially equal to UTC modulo leap seconds (which would screw up GPS computations). The ECI is constructed so the vacuum speed of light is isotropically c in the region between earth's surface and the GPS satellite orbits. To do this requires non-standard clocks, so they adjusted the satellite clocks to display ECI time wherever they are, by modifying the clock divider in the satellites to account for the effects of both gravity and orbital velocity. No "redefinition" is involved, because one of their requirements is to keep GPS time within 1 microsecond of UTC modulo leap seconds (in practice they do much better). Tom Roberts |