[ntp:questions] Allan variance plots

Piotr Trojanek ptrojane at mion.elka.pw.edu.pl
Wed Sep 17 11:49:31 UTC 2003


Dear NTP experts,

I have been reading NTP papers for a long time and now I wanted
to make Allan variance plot of my machine's clock. I have collected
time differences measured beetwen free running system clock and
PPS signal from GPS receiver using modified ppsctl.c from PPSkit.

My data is sampled once per second, so I have offsets like:

.007540355
.007513540
.007486726
.007460749
.007433934
.007412148
.007380305

Now I'm confused how to process this data to plots like those in
NTP papers. I have now about samples from few days of non-stop operation.

I feed this data directly to dataplot package (NIST software) and use
'AV PLOT' command in log-log scale -- result is straight line with
slope +1, so now I think it is wrong method. I can't figure out the
right way.

Thanks in advance for any help!

-- 
Piotr Trojanek



More information about the questions mailing list