[time-nuts] Why would I want a rubidium.. More

J. Forster jfor at quik.com
Mon Oct 5 19:41:40 UTC 2009


I neglected to mention there are very, very good reasons to have a single
master frequency standard and distribution amp for a lab. IMO, that is far
more important than the standard's absolute accuracy.

-John

==========

A GOOD crystal oscillator will stay w/in parts in 10E10 for days at least.
And I don't mean a DO. Just run it with a continuous comparison to LORAN
or WWVB or GPS and MANUALLY tweek it as needed. A benefit is that you KNOW
when you made an adjustment and not have to worry about random changes
from the oscillator. That's a reason I stopped using mine.

>From an engineering point of view, you rarely need to be that close to
frequency in absolute terms. Being very stable is often enough. As with
voltage and current, often what really matters is the ratio, not the
absolute value. Only if you are doing fundamental physics research or
tracking crustal movements is such ultra high precision really needed.

Why do you really care if your clock is off a few nanoseconds, or your
position off a few feet? If you are going to put up a fence, the legally
defined property line is the issue, not what GPS says.

My main gripe w/ Rb and especially Cs is they have very limited lives.
IMO, it's just not worth paying extra money for stability I don't need. If
I ever do need their capabilities, there are several on the shelf, but
until then they can sit unpowered.

FWIW,
-John

==============


> Here's a reason why you might "want" (as opposed to "need"!) another
standard beyond the GPSDO.
>
> An Rb or Cs is self-contained.  The GPSDO relies on an external factor,
the presence of GPS.  Holdover mode may be good to keep thing close for
a while, but over the long term the GPSDO without GPS is no better than
its crystal oscillator.
>
> The Rb is a secondary standard and therefore isn't "correct" by
> definition, but it has a low aging rate and gives you something
> independent of GPS to use for measurements.  Monitor it against GPS for
a while to learn its offset and drift characteristics, and then you can
extrapolate its performance out over a much longer time than you could
with an OCXO.
>
> You may not require that independence, but it gives you additional
measurement capability.  For example, comparing the output of two GPSDO
may not be meaningful because their frequencies could be correlated by
their common view of the GPS constellation.  Using an Rb reference would
eliminate that common mode error and reveal information about the
GPSDO's short and medium term stability that would otherwise be hidden.
>
> John








More information about the Time-nuts_lists.febo.com mailing list