Don't you wish it was always like this?

[+user13371]

You don't really NEED a super accurate set of coords and a great GPS to find a cache, but it helps. Noticed this on a quick urban today:

 

http://www.geocaching.com/seek/log.aspx?LU...97-73c68303a17b

 

The GPS claimed I was within "0.05 ft" of the spot.

 

It's true the PN-40 is remarkably accurate, but finds like this are the exception, not the rule. The hider may have recorded exceptionally good coords -- maybe using another PN-40 -- or it could have just been dumb luck (or both). Other caches I found today ranged from this one to about 100ft from where the coords said they were.

Edited October 19, 2009 by lee_rimar

[GOF and Bacall]

If they were all that accurate I'd switch to the dark side and start posting soft coordinates.

[+RRLover]

You don't really NEED a super accurate set of coords and a great GPS to find a cache, but it helps. Noticed this on a quick urban today:

 

http://www.geocaching.com/seek/log.aspx?LU...97-73c68303a17b

 

The GPS claimed I was within "0.05 ft" of the spot.

 

It's true the PN-40 is remarkably accurate, but finds like this are the exception, not the rule. The hider may have recorded exceptionally good coords (maybe using another PN-40 ), or it could have just been dumb luck. The other caches I found today ranged from this one to about 100ft from where the coords said they were.

My 40 zeroed out at 0.5' on a cache @ the East portal to the Narrows Bridge, in the Memorial Park.

It truly was 6" from the cache, only time at that resolution, . . . but always w/in stated guesstimate.

I'm happy, 10X better would be better, still happy though.

 

I was just in the Alameda neighborhood this afternoon. Christmas pix for the RRRescue. Portland is a

beautiful city, you're fortunate.

 

Norm

[+Deadelm]

First off, that is great! I wish they were all like that. But I'm going with "Luck" (you can reference it as "dumb luck") I consider myself new to the sport (hobby) as evident in my profile. I run a PN-40 right along side a Triton 400 or 2000. So far the Tritons win with accuracy and the PN wins with ease of cache use. Geocaching is "as I can fit it in my schedule or route hobby", and not the reason for owning a gps. One of my first finds with the Triton 2000 was actually reading "0", I looked down and was standing on top of the container (under a rock). It was luck that the recorded location was matching with the gps at that moment. I do understand the variation of the recorded location and gps reception. Having the >1' reading while on top of it is a lot of fun and keeps me coming back for more, along with some of the great locations and ideas for hides. I think it's time to place my first.

[+user13371]

I think for the PN-40 report what it did must have used some funny math and amazing optimism.

 

The GPX contains (and the PN-40 reports) position in degrees decimal to six places. I think 0.000001 degrees of latitude is about 4.375 inches -- so to claim it's within 0.05 feet (0.6 inches) is several times sharper than the available precision.

 

But I'm not sure of those numbers -- would someone else do the figures on their own and let me know what you come up with?

Edited October 19, 2009 by lee_rimar

[+Cardinal Red]

I think for the PN-40 report what it did must have used some funny math and amazing optimism.

 

The GPX contains (and the PN-40 reports) position in degrees decimal to six places. I think 0.000001 degrees of latitude is about 4.375 inches -- so to claim it's within 0.05 feet (0.6 inches) is almost an order of magnitude sharper than the available precision.

 

But I'm not sure of those numbers -- would someone else do the figures on their own and let me know what you come up with?

 

Based on: an international nautical mile as defined to be 1852 metres (IHC, Monaco, 1929)

 

0.000001 Degrees is 4.3748031496 inches. Your math is sound if you are sure about that six decimal places.

[+user13371]

I was thinking more about this over my coffee this morning. What really matters is how wide a degree of longitude is up here ... I figure 0.000001 degree of longitude is just over 3 inches. Conveniently close to 0.25 feet.

 

Sooo... Even though the PN will only let you input and display positions in degrees to six decimal places, it must calculate them internally to at least seven. Otherwise it could never show a "distance to position" in increments smaller than .25 feet -- at least not down here in Portland.

Edited October 19, 2009 by lee_rimar

[Suscrofa]

You confuse accuracy and number of displayed digits.

Besides, accuracy in inch and feet is a non sense ! We are in the 21° century.

 

2m is at best what one can expect from a regular GPS. If you happen to be a few cm from a cache, just pure luck. Besides what about the "accuracy" of the cache location ?

[+user13371]

Suscrofa, I am not making any claims about the accuracy -- other than the tongue-in-cheek comment about <1" accuracy.

 

My main interest here is in the reported distance to the waypoint. Let's say I'm routing to a waypoint at 45.000000, - 122.000000; that point isn't going to move. If the GPS thinks I'm at 45.000000, - 122.000001, it should report "0.25 feet" from the target. But if that last digit of longitude changes, the distance has to change in increments of about 0.25 feet. The only way it can change in smaller increments is if internally the GPS is doing the caculation internally with a larger number of significant digits than it displays for lat/lon

Edited October 19, 2009 by lee_rimar

[+Rockin Roddy]

You confuse accuracy and number of displayed digits.

Besides, accuracy in inch and feet is a non sense ! We are in the 21° century.

 

2m is at best what one can expect from a regular GPS. If you happen to be a few cm from a cache, just pure luck. Besides what about the "accuracy" of the cache location ?

 

I'll be happy to acknowledge I notice the same thing...and quite often. Pure luck....maybe. But if so, I'm one lucky cacher! And yes, I too use a PN-40.

[+Red90]

That sure seems a strange thing to do, displaying precision at least two orders of magnitude higher than the possible accuracy. I wonder why the GPS chip is even designed to output values to that level?

[Suscrofa]

Suscrofa, I am not making any claims about the accuracy -- other than the tongue-in-cheek comment about <1" accuracy.

 

My main interest here is in the reported distance to the waypoint. Let's say I'm routing to a waypoint at 45.000000, - 122.000000; that point isn't going to move. If the GPS thinks I'm at 45.000000, - 122.000001, it should report "0.25 feet" from the target. But if that last digit of longitude changes, the distance has to change in increments of about 0.25 feet. The only way it can change in smaller increments is if internally the GPS is doing the caculationm internally with a larger numbner of significant digits than it displays for lat/lon.

 

Sorry, feet mean nothing to me when it comes to distance except when I use mine to walk !

[+user13371]

Nothing strange about it at all. The GPS knows (and so do we all) that the reported position is an estimate with a certain likelihood of error. In fact, if you look at the picture I posted originally, you'll see the GPS drew a circle around the position.

 

It's not saying "YOU ARE HERE" in definite terms, it's saying "I think we're somewhere in this circle, and the center of this circle is on these coordinates."

 

It just happened on this ocassion that the center of that circle was in exactly the right place.

Edited October 19, 2009 by lee_rimar

[+shutle64]

Using the coords system of DD MM.MMM Each .001 is 8ft one direction and 6ft other direction. So accuracy is never going to be better than 8 ft. I actually seen a 60csx display 8 ft accuracy on the screen before. Most companies that sell gps say 1 meter accuracy or about 9ft.

[+julianh]

That sure seems a strange thing to do, displaying precision at least two orders of magnitude higher than the possible accuracy. I wonder why the GPS chip is even designed to output values to that level?

Your GPSr contains a computer chip - don't ask me what chip, what programming language, what operating system, etc, as it will vary for each make and model, but regardless of that, there is a computer at the heart of your GPSr.

 

I will bet that the internal maths is done in floating point maths, possibly using Single Precision (32 bits - about 7 decimal places equivalent precision – unlikely,) but more likely Double Precision (64 bits - about 14 decimal places equivalent precision). (It's even possible that the maths is done in Quadruple Precision - 128 bits, or about 28 decimal places - but that would sound like overkill to me for a consumer device.)

 

I will also bet that the fundamental geometry maths is done internally using radians as the angle of measure (one radian is approximately 57.3 degrees), not degrees, degrees and decimal minutes, degrees minutes and seconds, mils, grads, UTM, or whatever. This is because all the standard computer trigonometric functions are based on algorithms which use radians as the unit of measure, and these angles can then be converted into other customary units of measure.

 

The GPSr's processor will just blindly go on doing its trigonometric maths, second after second, calculating your latitude and longitude (in radians) to the best precision that it has been built and coded to be able to determine. (There is no point in reducing the computed precision of the location internally; the rounding-off will be done on the displayed result, not the internal representation of your location.) It will then output and display this location in whatever coordinate system the user requests, and using however many places of precision the device's programmers have deemed appropriate.

 

You can see this effect in the spreadsheet of your choice. Calculate the square root of two in an empty cell. Depending on your default settings, you will probably see something like "1.414214". Now increase the precision to the maximum available, and you should see something like "1.414213562373100000000000000000". This shows me that my copy of Excel is using Double Precision maths (14 decimal places). Changing the display format does NOT change the value stored in memory, only the way it is displayed.

 

For example, assuming your device uses Double Precision maths, the internal maths can probably calculate a theoretical latitude and longitude to an accuracy of something like 10^-13 radians, which is equivalent to something like one micron (one millionth of a metre) at any instant; however, the device can also compute an estimated accuracy, which will be in the order of a few metres at best, so there is simply no point in displaying your location to a precision of greater than the nearest metre or foot. (Even this level of reported location accuracy is optimistic.)

 

For consumer GPSrs with ~3 metre maximum accuracy, the maximum meaningful displayed precision in various coordinate systems is something like:

 

1 metre

1 foot

0.1 second

0.001 minute

0.00001 degree

 

Any more displayed digits than this is simply meaningless.

 

Hope this helps.