[ntp:questions] Re: Allan variance plots

David L. Mills mills at udel.edu
Thu Sep 18 13:21:38 UTC 2003


Piotr,

I don't know what your DATAPLOT reference is talking about. White phase
noise has slope -1 on the log-log plot and random-walk frequency noise
has slope +0.5. This is confirmed by many journal articles and by my own
measurements. Forget about data plotting packages and go directly to the
physics involved. Your quotes about slopes are plain wrong. Believe it
really and truly.

Some articles talk about frequency spectra rather than time series.
While one can be derived from the other, watch carefully what the
reference assumes.

Our fascist news system won't let me include all the text of your
message, so others may be totally confused by my answer. Thus be it so.

Dave



More information about the questions mailing list