[GPS_Standard] HP oscillator slowly moving

[Dave Platt]

On 09/06/2012 01:15 PM, Chris Howard w0ep wrote:
> 
> I've been running my GPSDO since May.
> It seems to work well.  We have had a couple of power
> outages and it gets sync'ed up again and continues on.
> 
> After some initial fiddling around to get the control
> voltage near zero, I have just let it do it's thing.
> 
> The control number is now 00710, steadily sinking over
> the weeks.
> 
> I'm wondering if this slow movement is a normal pattern?
> My oscillator is an HP from an old frequency counter.
> 
> I have log files from most of the period this thing
> has been in operation.  I put a local-computer time stamp
> on each line, so in theory I could graph the
> control voltage change if that would be of any use or interest.
> 
> I guess I'm kind of concerned about the oscillator and
> wonder if I should try to acquire another.

>From what I've read, and in my own experience with a VE2ZAZ
standard equipped with an Efratom OCVCXO, it is quite common
for a crystal to shift frequencies significantly as it ages.
The changing control voltage in your GPSDO could quite easily
be the result of this... the discipline is doing what it should,
and is counteracting the crystal's tendency to drift.

As I understand it, the drift is primarily a purely physical
phenomenon... the crystal structure, and (especially) the little
metal clamps which hold the crystal in place, "loosen up" over
time as a result of the crystal's vibration.

The effect is (I'm told) most pronounced when you switch an
oscillator on, after it has been off for a significant length
of time (weeks or months), and it drops off after the oscillator
has been running for a few weeks.  There will continue to be some
long-term drift as the crystal ages, but it's much less than what
occurs during the initial "loosening of muscles after a nap" period.

My own Efratom oscillator drifted quite a bit for a few weeks after
I got my DO running, and there was a steady downwards drift in the
average control voltage.  After a few weeks, the drift became less
than the day-to-day variation (which I believe is almost entirely
due to temperature change in my garage), and nowadays the average
control voltage has no long-term drift that I can see (whatever is
there, is a lot less than the day-to-day temperature-related cycle).

Here's the most recent month or so of data:

   http://snulbug.mtview.ca.us/dave/fll.png

If your oscillator is drifting fairly steadily, and isn't
undergoing sudden jumps in frequency, and as long as the DO
controller is tracking the drift well enough to keep the frequency
error within the range you're willing to accept, I would not
think you'd need to replace the oscillator.  If you do, expect
whatever new oscillator you install, to drift significantly
until it gets "broken in".

You may well find that your oscillator, like mine, will drift
less and less as time goes by.