Order Tray | Contact Us | Home | SIG Lists

[time-freq] Frequency Standards - Which to believe?

Eric Lemmon wb6fly at arrl.net
Thu Jun 22 01:39:56 UTC 2006


That's an excellent question!  I believe that the answer was published in
some papers that described the "Flying Clock Experiment" of several decades
ago that proved Einstein's theory of time compression and expansion.

As I recall from one of the papers, cesium standards are precisely correct
in frequency over a sampling interval, but they suffer from transition noise
and jitter.  Rubidium standards, on the other hand, are extremely stable
over long periods, but they must be tuned to place them on frequency.  In
other words, your rubidium standard does not operate unequivocally on a
predictable frequency; cesium standards do.  These statements are vintage,
and may no longer be true today.

I bought two of the Z3801A standards a few years ago, and I was pleased to
note that after 30 days of operation, they were almost perfectly in step.  I
attribute any error in synchronism to differing lengths of antenna cables.

Since the GPS timing signals are derived from cesium clocks and are
constantly checked for accuracy, I would suggest that you might adjust your
rubidium standard to agree with the GPS-disciplined standard.  However, I
believe that a 70-second period at 10 MHz is about 1.4 times 10 to the minus
ninth power, which is an exceedingly small error.  It may not be possible to
"tweak" the rubidium oscillator by such a small amount without overshooting
the zero point.

73, Eric Lemmon WB6FLY

-----Original Message-----
From: time-freq-bounces at lists.tapr.org
[mailto:time-freq-bounces at lists.tapr.org] On Behalf Of Arthur Shulman
Sent: Wednesday, June 21, 2006 7:08 AM
To: time-freq at lists.tapr.org; Blair Batty
Subject: [time-freq] Frequency Standards - Which to believe?

For the past 3 years, I have been depending (blindly) on the accuracy of
a surplus Efratom Rubidium Standard as an external reference in my Lab.
My faith was based on the theory of 'What can go wrong with the natural
resonant frequency of the Rubidium atom?'

Recently I was fortunate to acquire a HP Z3801A  GPS Standard. It seems
to work correctly, but I have concerns. Those more knowledgeable will
know the next line! "The man who has two clocks doesn't know the time."

I set up an X-Y display on a scope to compare the phase of the two
signals. I am able to get quite a nice Lissajou figure, with a period of
about seventy (70) seconds. But my reasoning is, 'Why should there be
any difference at all since I am expecting short-term frequency accuracy
better than 1E-11 for both devices?'

Also, assuming there is in fact a difference, this gives rise to several
other questions:
a)    what part of the physics do I not understand?
b)    how does one tell which device is really correct (or more
accurate)?, and,
c)    how can one get both instruments to agree with what they output?

Arthur Shulman ve3zv
Waterford, Ontario

time-freq mailing list
time-freq at lists.tapr.org

More information about the time-freq mailing list