The consistent error (difference) Karl reports might be due to the
difference in datum used in the gps receiver and the datum of the
benchmark.  For example, there is a difference in my area of about 20
meters between the North American Datum of 1927 (NAD27) and the later
datum NAD83.  GPS rcvrs can  usually be set to output coordinates in a
wide selection of datums.

I have found very good agreement in my measurements comparing GPS and
several benchmarks in the area.  The difference among benchmarks is
typically less than 1 cm using "homebrew" differential gps and
inexpensive single frequency gps receivers.

If you are interested in GPS surveying (or freq standards & timing) you
might like my GPS webpage at www.rt66.com/~shera.

Brooks


>Subject: Re: New SDR release, GPS without SA and new A/D card
>From:    Karl Cunningham 
>Date:    Sat, 16 Jun 2001 10:18:04 -0700

>A few thoughts on the GPS accuracy --

>The following URL
>http://www.ngs.noaa.gov/cgi-bin/ds_radius.prl
>has a search engine to find survey monuments within a specified radius
of a
>particular lat & long.  I found one near me that had been surveyed to
>within fractions of an inch and took my GPS there.  It was consistently
>about 15 feet off in the same direction.

>How about the following poor-man's DGPS.
>1.  Lock the satellites in use by the GPS so they don't change.
>2.  Make an offset measurement at such a benchmark.
>3.  Travel to another close-by location.
>4.  Could the offset be applied to improve location accuracy?

>I suppose the real test of this is to get an offset from one surveyed
>benchmark and then testing the offset at another surveyed benchmark.  I
>realize there are errors due to reflections in GPS signals off
buildings,
>mountains, etc., but I would  bet some improvement could be made over
the
>uncorrected position.

>Karl







__________________________________________________________

Public Seismic Network Mailing List (PSN-L)