[BC] GPS

[Cowboy]

On Wednesday 04 February 2009 11:17 am, Mike McCarthy wrote:
>  I don't see how going out 30+ miles makes it any more accurate for a proof. 

 Within the confines of what FCC will accept, you may be right.
 For me, I prefer to start with as minimal an error as possible, as
 other errors WILL creep in anyway, no matter how much we might try
 to minimize them.

>  Since most receivers resolve to 0.5 deg, that would 
>  result in a displayed accuracy of something less than 450 arc ft. or +/- 
>  225 arc ft. or +/- 3 seconds geographic at the E/W cardinal ordinates.

 Which is also why a consumer grade unit in and of itself is not
 up to the task, any make, any model.
 ( except very close-in, where one can walk back and forth about
 20 feet, and find the center of the trip point with reasonable accuracy )
 BUT, when combined with real mapping software, where the GPS unit
 is not being called upon to resolve anything, they are more than
 up to the task of simply telling you where you are, to the greatest
 precision of which they are capable.
 At max zoom, this will put you within a few feet at any distance, assuming
 your radial was plotted with sufficient accuracy to begin with.

>  At a distance of 10 miles, you're well beyond anything the FCC has ever 
>  remotely required for accuracy and precision, thus GREATLY diminishing any 
>  returns for the time/fuel invested to drive 4+ times the mileage of a 10 
>  mile radius.

 Assuming you can get ALL of your points within that ten mile maximum, and
 that your target result is "good enough" and not necessarily "right."
 What you propose will "work" but I still maintain that it is an expensive
 way to get mediocre results at best.

>  If anything, the greater distance makes it harder to locate  
>  the radial center since you need to cover more than 1/2 mile and then back 
>  track a greater distance to find the true center of the arc degree. Which 
>  could be as many as 10 seconds on a typical receiver.

 Again, why consumer grade GPS are not "of choice" for locating anything
 any more accurately than "close enough" to see the street on which the
 target address exists.
 Even if the GPS unit is capable of map display, routing, and such, I won't use
 'em. ( those features ) Those units, though costing much more, are no
 better for telling you where you are than the cheapest units available.
 As in a previous, all I want from any GPS, is a reliable, repeatable way for
 it to tell me where I am. Nothing more. NMEA output to mapping software
 beats any consumer grade GPS "feature" EVERY time.

>   the basic question is really WHY? The typical GPS can't 
>  resolve anything tighter than 30 ft. anyway..which is critical at closer in 
>  points.

 I once heard an engineer argue that we are required to read our instruments
 to no greater than 2%. Calibration to no greater than 5%.
 SO, assuming the previous reading was low, and yours is high, anything
 within 14% was "good enough."
 I disagree.
 Errors compound, so eliminating as much error as possible at all phases
 results in data of much greater accuracy in all cases, regardless of
 whether or not FCC requires it.
 In the worst case, it's why I do a better job, and get better results than most.

 The GPS is probably not capable of displaying any more accurately than that,
 so don't ask it to display anything ! Just tell me where I am, and let me
 use something that *is* capable resolve it for me.
 Most all of the GPS available at all, are capable of resolving the satelite
 signals to within a few inches. That's all I expect it to do.
 That's all I ask it to do.

>  Also, unless your GPS can resolve to 1/100 of second, you'll never have a 
>  tighter box than a 10 ft. square (at this latitude).  Which is the error 
>  margin at 1 mile radius typically on DGPS enabled RX and negates the 
>  "precision" of using a really distant end point.

 If you're trying to use a consumer grade GPS alone, then I'd agree.
 ( any GPS without monument recalibration reference that day )

>  With that said, a 30 mile end point is appropriate for longer distances, 
>  such as when running conductivity measurements. It allows a tighter cone to 
>  locate a point and be "on" the true radial at greater distances.  But it 
>  still does nothing for proof accuracy/precision.

 If one is trying to use the GPS alone, then you do have that cone to deal with.
 In the example you cite, both the cone created when you plot, and the cone
 created when you attempt to resolve using only the GPS, which should
 overlap by some unknown amount, probably, introducing a potential
 error of several hundred feet at best, large fractions of a mile at worst.

 The way I use them, there is no cone. There is a true on-bearing line, and
 the GPS unit need only resolve it's present location.
 When doing conductivity measurements, I plot to about 300 miles typically.
 If one is trying to resolve a bearing by driving back and forth, looking
 for when the GPS display trips to the next degree, one could be wasting
 an inordinate amount of time, and still be off by miles.
 In real world practice, the cheapest yellow hand held Garmin tied to the
 serial port ( or simulated serial port ) of a laptop running decent
 mapping software resolves to about 4 feet. ( WAAS enabled )
 Accuracy something greater than the pencil line on a topo-quad.
 At worst, I have a 16 foot error. At best, it's repeatable as I say, "close
 enough" that anyone can get within a few feet of the point I've described,
 and if my description is good, a few inches of where I originally read
 a point, and exactly repeat my process with little to no effort.  

 Greater error will be introduced if you are left handed, and I am right, by
 which way you turn to orient the FIM than by my mapping methods.

 I'll agree that if you're trying to find Ohio, that it may seem like over-kill
 to locate the third parking space east of 793 West State Street,
 Columbus, Ohio, that anything within 100 miles or so would be "good enough"
 but if I do, then there is NO chance you'll end up in Weirton, WV by mistake.

 Yes, driving around to locate bearings is time consuming and expensive.
 What you might do by driving a ten mile circle, taking a day or so, I can
 do at my desk in about ten minutes.
 Yes, it took me a few days to write the software, but now, even at ten times
 whatever rate you get paid, I'm both more accurate, and less expensive, AND
 I don't need to even have a clue what your rate might be.
 Send me the reference coordinates and bearings, and I send you back
 Garmin format files good for loading directly into a Garmin GPS and/or
 Garmin mapping software, and Delorme formatted text files, good for
 loading and displaying the radial in any Delorme mapping package.
 At $25/radial, you're not going to beat it.

 Of course, if you're re-running an existing proof, then none of this means
 much, as you must be where the point was previously read, located
 correctly "on bearing" or not.

-- 
Cowboy