[time-nuts] More Allan deviation data
didier at cox.net
Sun Oct 29 09:21:50 EST 2006
Ulrich Bangert wrote:
> Hi Didier,
> you may use the DRIFT section of the graphic editor of PLOTTER to remove
> a phase drift due to a constant frequency offset as seen in this data,
> with the result that you get more "visual resolution" out of it, as in
> the attached PDF.
> 73 de Ulrich, DF6JB
I ran the OCXO with the new antenna location overnight, except that the
laptop went to sleep some time in the night, so I have lost several
hours of data, which may explain the sudden slope change in the plot
below (I need to save timestamps I guess). However, the remaining data
is good I think. There was no sudden noise, and about 5 phase jumps due
to the OCXO drifting during the observation period.
I downloaded the latest version of Plotter (10-27) and it's all there. I
was able to quickly remove the phase jumps when the OCXO drifted past
the divided down period, and then remove the remaining slow drift.
What a nice piece of software!
(even the menu now has a *normal* File->Open :-)
The latest plots are at the usual place:
Now that the data collection method seems to be getting close, I wonder
how I should use the data, and what it means.
The Allan deviation seems to taper around e-10 at 1000 seconds, then it
looks like it wants to keep getting lower.
As a reminder, this is the HP10811 against the GPS receiver. I am not
sure what I should expect?
PS: I am still planning to measure ambient temperature at the same time
as I collect TI, via the GPIB, using a HP 3478A voltmeter and a
thermistor. This way, I should be able to correlate frequency changes
and temperature changes. I assume I can tell Plotter to ignore the
temperature field or display it with its own scaling. Should I use a
specific character as field delimiter? (such as a comma, or just spaces?)
(In the short term, temperature data will be thermistor values, until I
write the code to convert in degrees. )
Also, if I use time stamps, should I use field delimiter between the
time stamp and the data?
What is the time stamps are discontinuous (as if the laptop falls asleep
again), will that mess up Plotter?
More information about the time-nuts