[time-nuts] Counter OCXO behaviour

[Fabio Eboli]

Hello,
recently I tried to trim the Racal Dana
1992 04E internal reference, using the
GPS pps as a reference. I'd like to ask
a pair of questions...

- First is about the method.
I'm using the counter TI to measure it's
own OCXO. The GPS is starting the count,
the internal reference 10MHz (on the rear
there is a reference output connector)
stops the count and I log the result
as I've done before with other oscillators.
Is this a correct procedure?

- Second, is about the behaviour of the OCXO
after trimming. The OCXO seem not that stable
after the trimming, like if the crystal started
to age faster than before trimming, and now is
slowly stabilizing.
It's like the retrace of the crystals I've read
about, but the instrument was never powered down
in the last month. And it's oven has been on for
the last year. Will a crystal retrace also
after retrimming?
The frequency of the crystal is slowing, in the
first hours after the trimming rapidly, and now
more slowly. After few days the frequency seem
to be slowing somewhere around 1x10^-10 per day.
Unfortunately I havent logged the counter for
enough time before the trimming, but it measured
the same Rb with less than 3 digits of difference
(3x10^-10) in last 6 monthes.

Thanks,
Fabio.