[time-nuts] Question for any HP105 owners

John Miles jmiles at pop.net
Tue Aug 30 17:25:09 UTC 2011


> Right, the averaging time, tau, is critical when quoting ADEV
> numbers. In fact, the key feature of any ADEV plot is seeing
> how stability changes as a function of tau. But Mark mentioned
> "at 100 seconds" so that allowed me to compare his single
> ADEV number against the tau 100 second column of a full
> log-log plot.
> 
> Your data looks odd to me. Sample rate, or bandwidth, can
> have a noticeable effect for short tau but I would expect that
> by the time you're all the way out to tau 100 s that the points
> would be about the same. Instead in your case they differ by
> an order of magnitude.

Are you (Paul) using the counter in frequency mode, or are you using a
1pps/10pps divider?  The 5370 is noisier in frequency mode, and there's also
some dead time.  The best way to use the 5370 to compare two 10 MHz
oscillators is probably to use +TI mode with the sources applied directly to
the start and stop channels, and trigger the counter at 1 pps (or 10 pps)
with a divider.  I usually don't bother with the external trigger but I do
use TI mode over frequency mode if I actually care about ADEV fidelity.

-- john, KE5FX
 




More information about the time-nuts mailing list