[time-nuts] Data Collection for Allan Deviation

Poul-Henning Kamp phk at phk.freebsd.dk
Sun Aug 2 02:13:57 EDT 2015

In message <1502052008.304624.1438465781711.JavaMail.yahoo at mail.yahoo.com>, Bob Stewart writes:

>Here's an example of what I'm talking about.  See description and
>notes.  The blue trace is the one using the clock in the 5370A. 
>The other tests match pretty closely down to around 60s or so.

As a general principle:  Until you know why they differ, always trust the worst measurement
rather than the best measurement.

The "waves" on the blue line could indicate a spur which is not at
a multiple of your basic sampling frequency (1Hz)

Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

More information about the time-nuts mailing list