[time-nuts] Thunderbolt precise positions needed

Tom Van Baak tvb at LeapSecond.com
Fri Aug 14 01:46:36 UTC 2009


> I have developed some code that processes a 24-48 hour survey
> to calculate a precise position.  With a good (geodetic/survey)
> quality antenna I get my location to within a few inches.  With a
> conical timing antenna,  I get around 8 inches of error.  With a
> cheap patch antenna,  around a foot.

Nice results. Did you start to see any diurnal patterns in this
data?

Did you check if the 0x8E-A9 TSIP command lets you increase
the duration of the survey from it's default of 2,000 fixes? I recall
doing this once through TBoltmon for 86,400 or more fixes. This
presumably allows you to run your survey long enough to get the
precise position you were looking for and avoids the limitations
of the 0x32 command.

For my TBolt monitoring program I display the double-precision
values in the 8F-AC packet (to a few more decimal places than
TBoltmon shows). There's a double-precision 0x84 command too.

In determining an accurate position statistically as you are, would
having the Az/El or DOP values be of benefit? Or are using a form
of double-precision position averaging based on the 8E-A9 info?

/tvb





More information about the Time-nuts_lists.febo.com mailing list