Perhaps this might be of interest to some of you.
When I moved to my current QTH 6 years ago, one of my priorities
was to find a place up on a hill or ridge, in hopes of better radio
performance. I was not a contester at the time, and simply thought
that higher was better.
Then along came some interesting information. First, there was W3LPL
stating convincingly that there was no "magic" in station performance,
just physics, mathematics, and engineeering. Then there was N4KG's
article on elevation angles and antenna height. Finally, there was
AA4NU expounding on the vitures of computer antenna modeling, not to
mention getting me interested in contesting.
It seemed that these folks were saying things which made a lot of sense.
There was always one problem though -- I'd moved to a fine ridge top
location with quite irregular and varying terrain. The NEC-based models
could not predict accurate antenna patterns (especially elevation angles)
because of my irregular terrain.
Then along came N6BV with YTAD. This was obviously a step in the right
direction. Finally there was a model which was really taking into
account the effects of ground on antenna patterns. I heard Dean talk
about the new version at Dayton this year, and was also convinced by
his contest results.
There was still something I wasn't comfortable with, and that was the
generation of the terrain profiles required by YTAD (YT). There were two
problems I encountered. First, creating the profile plots manually from
1:24,000 USGS topo maps was a rather tedious task, and somewhat prone to
error. Secondly, a small change in azimuth (5 degrees or so) in highly
irregular terrain could produce a very different profile, and hence very
different results. I mentioned this to Dean. He recommended generating
several profiles across a range of azimuths, then observing how the results
change over that range. This made sense, but I started dreading the many
hours it would take to generate all of those profiles for my highly irregular
terrain. Since the power of YT is based on calculations made against
the elevation profiles, it made sense to try to generate them as accurately
(and as easily) as possible.
The problem still remained of how to generate all of the profiles which
were needed for a wide range of azimuth headings. I started thinking of
a better way to create the profiles. It turns out that the US Geological
Survey has created digital terrain data called Digital Elevation Model
or DEM data. The DEM data is available in a variety of scales, from
1:250,000 "1-degree DEM" down to 1:24,000 "7.5 minute DEM." The 1-degree
DEM's are the lowest resolution (least accuracy) and the 7.5 minute DEM's
are the highest resolution available from USGS. The 1 degree, 1:250k scale
data is available for download from the Internet via:
http://edcwww.cr.usgs.gov/doc/edchome/ndcdb/ndcdb.html
Technical details of the data are available through the USGS WWW pages.
The problem remained that the 1-degreee DEM was not sufficient resolution
to generate the kind of terrain profile needed by YT. Through my work,
I was able to acquire the higher resolution 1:24k 7.5 minute DEM data for
much of Tennessee. Using Geographic Information System (GIS) software tools,
The DEM was converted into a "lattice," or a raster grid with floating-point
values. The lattice was used to generate a "surface," which is a smooth
representation of the elevation. The surface can be viewed in 3-D, or have
raster or vector data 'draped' over it, and other such techniques. For the
purposes of YT, the surface can be 'sliced' at any azimuth, from any location
to create any number of elevation profiles. The profiles can be viewed or
saved to an ASCII file, which can then be brought directly into YT.
The results look very promising so far. Here is part of a profile for
K4RO at zero degrees azimuth (due north) :
$RECNO DISTANCE ELEVATION (all units are in feet)
1 0.000 742.224
2 44.245 750.692
3 88.491 758.812
4 132.736 766.702
5 176.982 774.076
6 221.227 781.147
7 265.472 784.858
8 309.718 788.569
9 353.963 789.040
10 398.209 788.977
11 442.454 783.336
12 486.700 774.471
13 530.945 765.911
14 575.190 757.791
15 619.436 749.950
16 663.681 743.337
17 707.927 736.724
18 752.172 725.847
19 796.417 714.794
20 840.662 702.655
21 884.908 690.127
22 929.153 681.757
... ... ...
(values continue on for a few miles...)
The results listed above agree very closely with those obtained using
7.5 minute topo maps, looking at contour lines with a magnifying glass
for hours. Using the DEM data is much less tedious, and probably more
accurate. The DEM values also agree with elevation values measured at my
QTH using a quality Global Positioning System (GPS) satellite receiver.
I believe using the DEM data generates the best elevation profiles available
for civillians today. A few cursory runs of YT yield results that seem
more like what I experiece on the air. Since the power of YT is based
on elevation profiles, it makes sense to have as accurate a profile as
possible. My goal of course is to be able to use this terrain data along
with YT to maximize the site I have already. If I ever plan to move again,
you can bet I'll be doing some major terrain analysis first.
What's next? First, preparing more DEM data into surfaces for more areas
of TN. It's a lot of work preparing the DEM data, but it's quite useful when
completed. One can generate shaded 3-D relief views, and do many interesting
things with the surfaces. Much of the work is currently performed 'manually'
-- meaning lots of GIS commands and steps. I hope to develop more automated
methods of generating ranges of "needed" profiles (e.g. EU, JA, SA) through
the use of scripts.
The ulitmate tool would be a version of YT which can calculate reflections
and diffractions in 3-D, integrating the results of ray tracing against a
*surface* instead of just a 2-D profile. Does anyone know of any progress
in this area? I'd be interested in any comments.
-Kirk K4RO k4ro@music-city.tdec.state.tn.us
--
FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com
Search: http://www.contesting.com/km9p/search
|