[time-nuts] Second FTS4060 shows Drift, is it me?
Tom Van Baak
tvb at leapsecond.com
Tue Feb 7 23:27:55 EST 2006
> I've noticed that the emphasis on this list seems to be placed on
> fast measurements of short-term drift. I'm more interested in long
> term drift, so I use a clock accumulator to measure drift over
> months (no setup has lasted a year, yet).
> Why is the focus on short-term drift?
> Bill Hawkins
Sounds like you're doing the right thing.
I think the key point is not so much short- or long-term
but how long do you have to wait before you get reliable
results. And this is especially true if you're testing or
repairing some surplus gear from eBay. You want to
know as much as possible about the performance in
as short a time as possible.
We could put a Cs standard in a closet with a Z3801A,
TI counter, UPS, and logging PC and take one reading
against GPS each day. A year later, assuming nothing
went wrong, we could look at a very nice plot and make
very solid statements about frequency drift rate, etc.
But sometimes it's nice to see trends as soon as they
are statistically accurate. And that may take much less
than a year. Yet it may take longer than days or weeks.
Everything is relative. Quartz drift rates can be measured
in a matter of hours or days; accurate rubidium drift rate
measurements take longer; maybe days to weeks. And
Cesium even longer than that.
What started this thread, and one similar to it almost a
year ago, was Brooke's Excel plots showing anomalous
frequency drift in one of his FTS 4060's. He wasn't out
to measure Cs drift as far as I remember; it just showed
up in his plots.
More information about the time-nuts