You guys are getting dangerously close to getting me going on this
GPS thread. Danger is that I find the topic interesting and so will
probably ramble. With that warning, here are some comments on some
of the statements that have been made:

Accuracy: Gps works by measuring the distance to 3 or more satellites
at the same time. Each satellite trasmits two codes which are
precisely time-tagged, and the receiver computes the distance to the
sattelite by shifting an identical code (which it generates) in time
until it matches the codes from the satellites (this allows it to
measure the time it took for the signal to reach the receiver). The
more accurate the timekeeping in the receiver, the better it can do
this-- and the internal "clocks" are one of the main reasons why GPS
receivers vary so much in price. Surveying grade receivers cost
about the same as any other electronic surveying tool-- $10,000 or
more (sometimes much more). Then there are "mapping grade"
receivers that are in the few thousand dollar range,
followed by the "hiking" grade receivers.

Clock error is one reason you need fixes on three satellites to get a
position. Another quality that affects cost is the programming built
in to the receiver to correct for the errors that keep the three
distances from the satellites from ever intersecting at an exact
point. Receivers do some fancy calculations to actually get to a
position readout.

All civilian receivers read the "course acquisition" code, which is
subject to "selective availability". SA is an error purposely introduced to
the time code at the satellite, and can range up to about 30 meters,
in reference to the effect it has on your reading on the ground.
Military receivers read the "precise" code, which is not subject to
SA and so is more accurate. Military receivers also pick up the code
on two bandwidths at the same time, which minimizes errors caused by
ionospheric delay.

For a civilian receiver you have to expect about 20-30 meters error
from ionospheric delay in the daytime (3-6 meters at night), 1-3
meters error from tropospheric delay, 1 meter or so error from the
satellite clock, 4 meters or so from the ephemeris the satellite
broadcasts, and 1-3 meters from multipath, which is what you get if
you try to take a reading when standing along side a building and the
signal bounces off the building before getting to your receiver. SA
adds another 30 meters. Then you have dilution of precision from the
geometry of the satellites relative to you when you read them-- all
this adds up to an expected accuracy of ~100 meters horizontally and
~140 meters vertically 95% of the time. (this is a 2drms value for
the statistics nuts)Military receivers are expected to get about 22
meter accuracy horizontally and 29 meters vertically.

It matters not how many channels your receiver tracks, or how much
you paid for it. It will never be significantly better than this. SA
is due to be phased out over about a 10 year period. The 10 years is
supposed to give the military time to develop a good program to
degrade the signals for the "enemy" during national emergencies, yet
keep it useful for the commercial market at other times. On this
note, the pentagon admits that the only time they ever used GPS in
war-- desert storm-- they had to turn off SA because their weren't
enough military receivers to go around and they had to buy civilian
ones.

So-- why are there surveying and mapping grade receivers? Even when
SA is turned off you still have significant errors from the
atmospheric interferences. The way to get around all the errors is
to simply put one receiver on a known point, and have it track the
same satellites that your "rover" receiver is tracking at the same
time. Then, with computer processing, you can correct errors in the
rover's position by comparing its output with the "drift" of the base
receiver. (There is another way-- SA cycles over about a 24 hour
period, so some receivers, such as the Trimble Scoutmaster, can be
left on a point for 24 hours logging continous readings-- then it can
average out the SA error).

Using two receivers is what has made GPS take off. You don't have to
own the base receiver. The Coast Guard has two or three working
continuously off Washington, you can log on with radios or get the
data off the net to correct readings after the fact. There are
several commercial outfits that have base stations at radio stations
in Seattle, Portland and other cities. You get a subscription that comes with
a pager that you hook up to your receiver, and the corrections can be fed to it in
real time. Base receivers are being set up at airports, when the
system is in place it will be viable to use GPS for air navigation.
Several cities have set up base stations for their own GIS mapping.
So on and so forth. Radio linked base stations is what makes them
work in automobile navigation systems, and now some areas
are putting headphones on them with computers that talk
to you as you walk down the street-- "Turn left here"- to guide blind
people. Us private surveyors usually set up our own base
station for each job-- then with survey grade receivers we can get
centimeter level accuracy. It is when you are using a base station
that you start worrying about the actual quality of the receiver's
inner workings. It makes no sense to pay for a subscription for 1
meter accuracy and then plug the pager into a 15 meter unit.

Extra Channels: today many hiking grade receivers tout having 12
channels. That doesn't make them any more accurate for static
positioning. It does make them work better for tracking courses while
you are moving, for instance when you use them on a boat. It also
makes them a little better in the woods because when a two or three
channel receiver cycles back and forth between the satellites it is
more likely to lose lock. So the 12 channel receivers may work
faster in the woods. There are several moderately expensive
receivers that are designed for mapping under tree canopies, which
work quite well in the forest.

For my money, the most valuable feature in a hiking grade receiver is
that it NOT give you a reading at all until it gets a good fix. I
have used some $189 Co-op purchased receivers for mapping plots in
the woods, and they got fixes pretty fast all right, but I threw out
about half the readings when I plotted them. Clarice and I used a
Trimble Ensign to map rhododendron populations in China a few years
ago, and compared it with a Sony Pyxis at the time-- I noticed the
Sony gave readings faster but then, after a while, the readings would
jump quite a ways. The Trimble kept quiet until it got a good PDOP,
and gave the same reading the Sony did after the Sony jumped-- and
here I'm talking quarter mile jumps, not 100 meter blips.

Incidentally-- when using one in the mountains ONLY accept a 3
dimensional reading-- and then only for the horizontal position. The
reason is that elevation is a necessary part of the equation, and if
your receiver is telling you that it just has enough satellites to
give you a 2-D fix then it is fudging in the last elevation it read
to solve the problem. This can throw you off, again by about a
quarter mile. Remember-- the biggest recreational market for these is
with boaters, who don't have to worry about changing elevations, or
with trees.

By the way, neither Clarice nor I have ever used one for birding.
Too much like work. It will come about, however, eventually.
(Really, the reason Clarice asked about section lines on DeLorme
atlas's in the first place was because so many people use sections
for locations-- ultimately she is mapping rhododendron populations
and needs long. and latitude anyway. A good easily used grid that non
techie types understand is really what we need).

So, if any of you wandering about the woods finds a R. macrophyllum
patch, take a few GPS readings and send them to her. Thanks.
Jerry Broadus
P.O. Box 249
Puyallup, WA. 98371