[TowerTalk] Where to get HFTA Software

[Jim Lux]

On 3/4/12 8:32 AM, Pete Smith N4ZR wrote:

> All this business about locating your antenna within feet or inches is
> silly.  The terrain data are markedly less precise than that, as Jim Lux
> pointed out - more like some tens of meters.  What MicroDEM does, if
> used according to the instructions, is to draw radials from the antenna
> location out some 3-4 miles, which are readable by HFTA as .pro
> profiles.  These radials are simply the electronic equivalent of drawing
> pencil lines on a topographic map and entering the elevation at, say,
> 10-meter intervals
>




Original USGS DEM files were created from topographic maps. The 3 
arcsecond 1x1 degree squares I have were created from USGS 1:250,000 
scale topo maps (which cover 1x2 degrees).  They basically turned the 
vector contour lines into elevation data, etc.  The basic accuracy of 
USGS maps ("meets US map accuracy standards") is the size of a pencil 
point: 1/50th of an inch, so the position of a feature (be it road, 
contour line, etc.) on the 1:250,000 map is accurate to about 400 feet. 
  Since 3 arc seconds is about 300-400 ft, you can see that the files 
are of the same general resolution as the map it's based on.

there are LOTS of cases where the position of a road or stream shown on 
the map is different by several hundred feet from where it really is, 
based on either lat/lon or topography from the elevation data file (e.g. 
streams on tops of ridges)

The SRTM data is nice because it's all taken by one instrument and all 
processed the same way, so at least if there are errors, they're 
consistent.  And, 10 meter data is really taken at 10 meter accuracy, 
not interpolated from 30 meter data.


This doesn't even get into the translation of lat/lon into map 
coordinates, and "where is zero" (i.e. the datum question: WGS84 or 
NAD27) There's also a whole "reference ellipsoid" issue..  the features 
on the map and/or satellite photo may not match the topographic data. It 
could be off by 10s of meters, which is a big deal if your location is 
on the side of a hill.


Actually, it's amazing that google earth works as well as it does, 
because they ingest all sorts of imagery and have to reregister it to 
whatever GE displays with, tiling and merging appropriately.  This is 
distinctly non trivial, as anyone who has tried to make a large wall map 
by combining USGS quads or FAA sectionals of different rev dates



Like all modeling activities.. you should run your model and figure out 
stuff.  Then move things "a little bit" and see if the output of the 
model changes.