[time-nuts] Second FTS4060 shows Drift, is it me? Good Links
Tom Van Baak
tvb at leapsecond.com
Tue Feb 7 18:46:07 EST 2006
Correct, with a Cs standard you can expect that there
will be no long-term drift in frequency. However if you
look at the short-term you will often see trends that
look like drift. In your S/N 1227 plot both red and blue
show trends on the order of 3 to 4 days. You are wise
not to call this frequency drift.
But from my perspective, the 3 weeks of data that
is both red (Loran-C) and blue (GPS) are nice lines
with no parabolic component. It's not fair to "fit" a
2nd degree polynomial to 3 weeks of Cs data that
has more than 50 ns of noise.
For those 3 weeks, the slope of a *line* through the
blue points is about -4e-13 and the slope of the red
line about +4e-13.
I think you have a minor sign error. One counter is
measuring Cs to GPS and the other Loran to Cs,
instead of Cs to Loran, or something like that.
With the sign error corrected you'll also note both
lines are parallel. And that both blue and red show
a common bump around DOY 108. That suggests
it was your Cs going off a bit for 2 or 3 days, not
something unique to GPS or to Loran-C. There are
other visible common-mode excursions from a linear
trend as well.
I can't explain the parabolic trend in blue prior to
DOY 102. But it looks to me like as soon as you
started Loran-C logging you've got clean, linear
data from both GPS and Loran-C. Do you have
data for dates beyond DOY 120 to verify this hunch?
More information about the time-nuts