Are there any consumer-grade GPS with sub-meter accuracy?

In short, can I get 1/2 m accuracy for under $1000.

I want to measure, as precisely as possible, the location of several
points around a lake (bridges, buoys, etc). I know about "averaging",
but how long do I need to dwell "on-station" to get sub-meter accuracy
with WAAS? I don't know of any consumer-grade GPS that displays
position less than one meter. To get the accuracy I desire, I am
willing to post-process.

You could try the Earthmate Bluelogger - http://www.earthmate.com/eabllogpsw=
ig.html

The price is good, but the workflow is not the greatest. To get
submeter you will need to use the carrier phase option, which will
require static occupation of each point for long periods (30mins
should give good results, you may be able to get away with shorter if
the conditions are good).

US Positioning sold a garmin based solution, but there website looks
to be out of action at the moment.

Otherwise you might have to roll your own solution. Most of the
components you need are around, but it will take a lot of tinkering to
get something going. CMC Superstar used to be very popular with
universities for its carrier phase, not sure if they still sell this.
Some Garmin receivers output carrier (older models are more likely to
provide this), but the data suffers from half cycle slips. u-blox
LEA-4T sounds like a good choice. A good starting point for all these
options is Denis Milbert's software page:

http://home.comcast.net/~dmilbert/softs/ant2rin.htm

On Thu, 17 Apr 2008 02:53:58 -0700 (PDT), raewyn.gilford@gmail.com
wrote:

http://www.earthmate.com/eabllogpswig.html

http://www.gpspassion.com/fr/articles.asp?id=185&page=2

Post-processing is the wat to go, averaging won't do you much good.

The Garmin reported number is not the maximum error (impossible to report)
to expect but a statistical parameter that measures the error and as that is
the case, the actual error (this is a better term than "accuracy" for a
single measurement) will sometimes be larger than the Garmin reported error.
Unfortunately Garmin does do a poor job of stating exactly what accuracy
statistic their units report. Your disappointment between the averaging
acuracy and the reported averating accuracy of the Garmin unit on the web
page could be very well due to the above.

The Garmin reported number appears to depend on only the satellite
geometry (and maybe only of the best N), plus if you average a waypoint,
then the number of samples. It doesn't do any useful statistical maths,
and I've had occasional (persistent) errors of 10m or more vs a reported
2m.

You can also observe the lunacy by averaging a waypoint while walking
along, and watch the 'accuracy' number improve, whereas a unit actually
keeping track of the answers and their divergence would note that it
didn't have a clue where the fixed waypoint actually was.

--
GSV Three Minds in a Can
11,517 Km walked. 2,259 Km PROWs surveyed. 40.9% complete.