[time-nuts] COMPARE GPS DERIVED 10MHz FREQUENCY OF TO LOCAL OSCILLATOR

Rex rexa at sonic.net
Thu Nov 9 00:19:49 EST 2006


On Wed, 08 Nov 2006 18:00:33 -0500, John Ackermann N8UR <jra at febo.com>
wrote:

>Another way is to use an oscilloscope to watch the drift of the Z3801A
>versus the Spectracom.  You can set the scope up to show a Lissajous
>display, or just use one signal to trigger the scope while viewing the
>other.  Either way will let you calculate the frequency offset by
>measuring how quickly the Spectracom signal drifts some amount (e.g., 1
>cycle) away from the reference.

Assuming you think the GPS reference is good, using an o'scope to
compare phase is what I have done. Just the simple sine waves on two
different channels to watch. If there is significant drift, adjust the
one you are measuring to stop the drift. During that alignment you can
put them in a known phase state.

On some of the OCXO's I have played with, the adjustment is too coarse
to ever set it properly. On a few others, I noticed a "jumpy" quality to
the signal I was trying to set. Can't like those results. In spite of
the good name-brand of those, they are now in my third string. 

On a few after syncing I was able to get what I thought was a good
stability over a couple of days. If they are staying close, a stopwatch
and measuremnent of drift over a period of time can give a relative
stability measurment.

Seeing bad things on several "good" by reputation OCXO's that I bought
on the internet, I'd say, first look and see if it is really working as
it should. After that and some tuning as described, you can then start
looking at the finesse that this group cares about.

This is my first-order approach. Having seen bad things from several,
what I thought would be good, oscillators I would recommend you look at
the actual signal and see if you can tame it and if it looks clean and
good on the oscilloscope before you waste time on statistical analysis.





More information about the time-nuts mailing list