How close should 3 GPS units read when next to each other?

[Guest dmdeitz]

e them alone for an hour. How close should they read on their coords?

[Guest bwolv]

Well, ideally they should agree. but it is not a perfect world (or system)

Very good, but not perfect. I think the current accuracy values

for non-WAAS signals is better than 10 meters 95 percent of the time.

Usually they will be much better than that. But if one were

to read 10 meters north, and another to read 10 meters south of

the "actual" location (unlikely since many of the factors causing error will

apply to both gps's) this would cause a difference of about 10/1000

of a minute of latitude, and somewhat more for longitude. (since longitude

units are smaller than like latitude units except at the equator.)

Your GPS's should easily be within 10/1000 of a minute of latitude. They'll

probably be within 1,2, or 3/1000 or so. (each 1/1000 (.001) minute of

latitude is roughly 6.1 feet over the ground. I would expect they would

also be

within 5/1000 of longitude as well. And surly be with 15/1000 of Longitude...

 

WHat did you observe? Do they utilize WAAS?

Do they "drift" in some

unison? Does one always read higher or lower than another?

[Guest k2dave]

if they are too close they will interfere w/ each other. I don't know how close is too close or what the interferce will be.

[Guest dmdeitz]

I realized that the three of them (1 Magellan 315 and 2 etrex) that were touching were interfering.... next to each other for 1 hour on the deck, with the 2 story house overhanging their view of the south, I got:

 

for lat: ...893, .917 and .902, range= .024

lon: ...840, .832 and .836, range= .008

ele: 385, 401, 410 range= 25

 

And the etrex's even said weak signal as I leaned over them to read the numbers- but directly overhead was cloud free sky!

 

then I moved more into the yard away from the house and had 2 on the gorund while holding the 3rd at chest height:

lat: .900, .903, .904 range= .004

lon: .826, .824, .823 range= .003

ele: 404, 384, 401 range= 20'

 

so yeah, having them right next to each other has a big influence.

 

I guress i need to also see if the e etex' typically come in near each other in readings vs the 315 or if the 315 is in the middle of the range..

 

But being new to this, even the 2nd set of numebers seems frustrating for finding a cache.. but do I have the wrong expectatio that you walk to the coords on the cache info page and look down and there it is? Not right, or even in the spirit of the game, right? needing to look is part of the game? although for me, like the guy in USA today, coming home without a find is a bummer to me and my kids....

[Guest bwolv]

Well I'm guessing that moving away from the house made the difference.

The first set of values seems too high. So there was clearly something

degrading the accuracy. And one even indicated poor coverage. The yard

values seemed very consistent with my similar experiments.

As to your next question, Yes it will involve some searching. But that's

part of the fun. Your experiments would yield an area about 24 foot by 24

feet to search. In reality, they will probably be larger. Thus there will be

some analysis like "where would I put it if I were hiding it?" You really wouldn't

want it to be so easy as at your feet each time would you? Well maybe. There

will be some times when you don't find it however.

 

If your like me (and it sounds like you are) you'll want to know which

of your units is more "accurate". You can find a survey mark with this link.

 

http://www.ngs.noaa.gov/cgi-bin/ds_radius.prl

 

But use caution: Make sure you've loaded the survey mark waypoint data

in the same datum as it is posted in. EX: If the survey mark is listed in NAD27,

Then you must set your GPS to datum NAD27 before entering the waypoint.

Then you can return it to WGS84 (before looking for a geocache)

You'll also need to enter Deg Min Secs, versus Deg Min Decimal Min. Look

for a first order survey mark. Even a third order should be adequate. I was

looking for survey marks nearly 10 years ago with my first GPS. It's almost

as much fun as geocaching!

[Guest brokenwing]

Don't expect the etrex units to get more accurate if you leave them for an hour like you mentioned. To the best of my knowlege, they don't average like the Mag 315 will. When you leave it like this, the 315 will continue to refine the reading and should be pretty accurate. The etrex units are continuing to get new readings, some more accurate than others.

 

As far as expectations, without a good WAAS lock, about 25-50 feet is all that you can reasonably expect when geocaching. You have to remember that you are also dependent on the accuracy of the posted coordinates in addition to the accuracy of your own unit.

 

Another point, GPS sattelites are all over the sky. some lower on the horizon, so the idea that the sky is clear directly above you, does not necessarily relate to good conditions.

 

thanks,

brokenwing

[Guest Kerry]

Since SA was set to zero averageing really doesn't achieve a lot (as compared to pre SA times) and the difference in averages is a lot less dependent on the time period.

 

Also with handhelds face the equator as there are no Sats approx 45 degrees either side of the poles up to approx 45 degrees elevation.

 

Cheers, Kerry.

[Guest BigDoggie]

-line cache logs for more than a few caches, you'll understand that everything is not always perfect. One thing that you will see is finders posting revised coordinates, saying that the original poster was off by several, or evcen quite a few, thousandths. That might make the 20 foot circle into a 200 foot circle.

 

All this is to say that your GPS *should* get you very near the cache, but your mileage may vary.

 

A pretty common recommendation is that you carry a decent compass as well as the GPS. When you get close, use the GPS to get bearing and distance, use the compass to know which way to walk, and count paces.

[Guest brokenwing]

quote:

---

Originally posted by Kerry:

Since SA was set to zero averageing really doesn't achieve a lot (as compared to pre SA times) and the difference in averages is a lot less dependent on the time period.

---

 

Agreed, averaging is less critical post SA, but I think you may have misunderstood my post. I'm not talking about manually averaging readings over time. The Magellan 315 and 330 will go into averaging mode any time it is stationary. When doing this, it is constantly averaging ALL coords received. If one were to leave it turned on and stationary for a while, the listed coordinates do get more accurate because it is averaging hundreds or even thousands of readings. On my unit, it seems to stablize after about 5 minutes, but will occasionally change slightly after that until it settles in on a single set of coordinates.

 

If averaging didn't do anything as you suggest, why do you suppose both Magellan and Garmin offer this feature?

 

Thanks,

brokenwing

[Guest Iron Chef]

quote:

---

Originally posted by brokenwing:

As far as expectations, without a good WAAS lock, about 25-50 feet is all that you can reasonably expect when geocaching.

---

 

Even with a WAAS lock it is possible to get an EPE of 25-33 feet if the signals from the other sats arn't very good.

 

------------------

-Iron Chef

_ _ _ _ _ _ _ _ _ _ _ _ _

agefive.com/geocache/ ~ Fe-26

Lets Drive Fast and Eat Cheese!

[Guest Kerry]

Brokenwing, that 5 minutes as you have found is probably more than enough time these days after which averaging looses any further benefit due to precision.

 

Since the "accuracy of the system" and the "precision of the GPS" are reasonably close without SA averaging is really lost into a black hole after a very short time compared to pre SA days where certainly the longer one averaged the closer one became.

 

The original spec for Repeatable accuracy, which is what users are dealing with here was 141 metres @ 95%, Averaging then certainly had benefit. The new "official" specs should be available within 6 weeks so it will be interesting to see what they are?

 

Why it's still there in "some" units is probably a result of zero SA only being 12 months old and also it's one of those features that users think they want/require etc. Having said that it can/does have a limited purpose but really some days it's about as far as one can jump sideways and thousands of measuerments won't improve the accuracy below the precision.

 

Now if display output was available to a higher precision then averaging would have a bit more purpose (if 6' is important?) but those extra digit's on some/most displays will probably require a whole new layout to incorporate. This has to happen in the not too distant future to keep up with the system capabilities (for those users who really require it).

 

Cheers, Kerry.

[Guest GeoForse]

There have been a number of comments that the averaging is not as necessary in post-SA times. It was always my impression the averaging with SA wasn't that effective since the introduced errors were unpredictable, but generally of a long period. To eliminate them, you would have to average for hours. Also, I am amused when posters state that their readings were "at zero" when a cache was found, exceeding the precision of the entered coordinates.

 

I do have a question. What, exactly, does EPE mean? How is it derived in the various units. To what extent is it related to pdop? Is it related only to geometry, or does signal strength enter. Also, what is the effect of acquiring more than 3 or 4 sattellites? Does the unit make a judgement on which satellites to use?

 

Just curious...

[Guest GeoForse]

There have been a number of comments that the averaging is not as necessary in post-SA times. It was always my impression the averaging with SA wasn't that effective since the introduced errors were unpredictable, but generally of a long period. To eliminate them, you would have to average for hours. Also, I am amused when posters state that their readings were "at zero" when a cache was found, exceeding the precision of the entered coordinates.

 

I do have a question. What, exactly, does EPE mean? How is it derived in the various units. To what extent is it related to pdop? Is it related only to geometry, or does signal strength enter. Also, what is the effect of acquiring more than 3 or 4 sattellites? Does the unit make a judgement on which satellites to use?

 

Just curious...

[Guest Kerry]

Simply but interesting questions but some of the answers are not available.

 

EPE - Estimated Position Error and different manufacturers have some other different versions. What does it mean really? My impression, it's only a "relative" accuracy measure for that particular unit.

 

How is it derived - generally propriety information which none of the manufacturers will divulge. Some state their implementation gives gives their units the edge.

 

Most certainly related to geometry and "other" things.

 

More than 3 or 4 Sats - redundancy and a few other issues with lock times/obstructions etc. Some units will use all Sats while others can be selective.

 

Also have been interested in the averaging comments (or maybe that should be averaging impressions?).

 

My comments on averaging http://www.cqnet.com.au/~user/aitken/gps/gps_avg.htm

 

Cheers, Kerry.

 

 

[This message has been edited by Kerry (edited 02 August 2001).]

[Guest GeoForse]

Kerry...

 

Very interesting report on averaging. I wish I was smart enough to understanding it all. I think the important thing to keep in mind was your quote in an earlier post..

 

"Since the "accuracy of the system" and the "precision of the GPS" are reasonably close without SA averaging is really lost into a black hole after a very short time compared to pre SA days where certainly the longer one averaged the closer one became."

 

So now the question is, do YOU average? Under any situations?

[Guest GeoForse]

Kerry...

 

Very interesting report on averaging. I wish I was smart enough to understanding it all. I think the important thing to keep in mind was your quote in an earlier post..

 

"Since the "accuracy of the system" and the "precision of the GPS" are reasonably close without SA averaging is really lost into a black hole after a very short time compared to pre SA days where certainly the longer one averaged the closer one became."

 

So now the question is, do YOU average? Under any situations?

[Guest Kerry]

Geoforce, High order post processing relies on changing geometry, which occurs of a period of time. So to a point post processing could be classed as a ?type? of averaging, which if that was the case then one could say I do that.

 

But without making it sound complicated averaging and post processing are not entirely the same thing but there is ?some? analogy between the two. Also the precision capability is radically different.

 

Averaging in fact doesn't really fit as a good term with specifying GPS accuracy which is why the most often seen terms will be accuracy X@95% or CEP,RMS,2DRMS etc.

 

Cheers, Kerry.

[Guest CharlieP]

Since SA was eliminated over a year ago, I have found that the accuracy of my GPS depends mostly on the number of satellites it can lock on and the geometry of those sats. The GPS reflects these factors in the reported EPE (Estimated Position Error), but I have found that the EPE from my GPS requires some interpretation, probably because the algorithm was developed for use with SA on. With 7 or more sats and an EPE of 8 to 15 meters, my GPS will usually return me to within 10 feet or so of a position, but with 4 or 5 sats and an EPE of 40 meters, I may be 100 feet away. So it is important to take notice of the EPE figure. Since terrain and trees will interfere with reception, the number of sats locked and accuracy tends to be minimized on most cache hunts here in North Georgia, which are typically in hilly wooded areas. So the major factor in determining GPS accuracy here seems to be the ability of the antenna and receiver to pick up sats through trees.

 

CharlieP

[Guest Kerry]

Charlie, the EPE figure is certainly another interesting figure that as you say doesn't make all that much sense now that SA is set to zero. It generally appears that in some cases it might have been tuned to SA on.

 

One can see what manufacturers are trying to do with EPE but why don't they implement some of the DOP's (the better ones do) which have specifications, mean something and would be common between all receivers?

 

Cheers, Kerry.