[TowerTalk] HFTA first impressions

[Pete Smith]

At 05:22 PM 11/9/03 -0800, Jim Lux wrote:

>Just got my new ARRL Antenna book, which I bought primarily for the copy of
>HFTA.  Here are my first impressions..
>
>1) Why must they provide nagware copies of software that have time limits?
>I understand and support the concept of shareware, but, on the other hand,
>it would have been nice if the ARRL blurb said: "limited functionality
>versions of popular programs".  I wouldn't even mind getting developmental,
>beta type software on the disk, providing it was disclosed up front.


?  I hope Dean will respond to this.  Why should it be time-limited since 
there's no shareware charge involved.


>2) The instructions on getting DEM data are mindbendingly complex and
>confusing, especially with respect to properly installing it in the right
>directories, etc.  The microDEM program is quite slick (probably too
>powerful for this application, by the way), but all that capability means it
>takes a lot of configuring.  A program like RadioMobile does this MUCH
>better.


Never heard of it -- I worked with Dean on developing the use of MicroDEM, 
which was originally proposed by Kevan Nason, NS4T.  During development of 
HFTA, the USGS seam;less server came out, which offered a nicer solution to 
the problem of getting DEM data specific to a particular site.  Sgain, I've 
never heard of RadioMobile, but suggest you tell Dean about it if it's a 
better way.  MicroDEM is a very complex program developed entirely for 
another purpose, but its developer has been very helpful in developing its 
use with HFTA.


>3) The instructions given for creating the terrain profile files do not
>work.  The version of MicroDEM on the CDROM doesn't save the profiles
>created in the viewshed analysis. It doesn't allow you to click the "save
>radials" option, which is grayed out.  The HFTA pdf file tells you that you
>can enter the lat/lon for the viewpoint once you have clicked on the DEM
>data, but in reality, it's where the cursor is when you (double) click that
>determines the viewpoint.


Not if you have set up the microDEM defaults properly.  It is easy to skip 
over that point.  But if you take it step by step it works fine.


>4) It would have been much better if they had simply provided a single
>program that you enter the lat/lon of the station, how many radials, how
>far, what post spacing, etc. in a dialog, and it would spit out the required
>text files.  Sure, it's cool to use a sophisticated mapping program, but
>there's a lot more that can go wrong.


But compare this to sitting down with a topo map, a ruler, and a 
pencil.  MicroDEM was out there, and it works.  The development effort for 
a separate program such as you describe would have been prohibitive.


>5) There is no documentation about how HFTA works, in either the program
>documentation, or in the Antenna Book.  There's a general description in the
>book of the diffraction analysis, but no theory, and what's important, no
>place to go look for a better description.  The textbook cited on GTD is all
>well and good, but I wasn't looking forward to rederiving how HFTA might
>work from first principles.
>
>For any sort of modeling program, it's important that the theory be
>described somewhere, or it's impossible to validate the program.  Say you
>get some sort of output from HFTA... how would you verify that
>independently?  For a program such as this, where the development budget is
>presumably quite small (they're giving it away essentially for free, after
>all), the value of legions of unpaid testers and validators would be
>significant.  (The book mentions the work of Breakall, et al and their
>modified version of NEC-BSC that they did field tests on, but doesn't say
>whether HFTA actually uses the algorithms developed for that work.  The IEEE
>A&P paper is fairly vague on the details of the implementation)  One reason
>NEC is still popular after all these years is that there is copious
>documentation on how it works.
>
>6) there are a number of fairly obvious formatting errors in the dialogs
>(Options dialog, where the caption text on the command button doesn't fit
>and isn't readable).  Such things don't fill one with confidence about the
>testing and validation process.
>
>7) No documentation about how the "figure of merit" is calculated in HFTA

This I can speak to, because the FOM is based on a piece I published in the 
September-October 2000 NCJ, titled "Scoring Your Antenna 
System."  Essentially, the pattern of the antenna is converted to power 
ratios, and then the power at each elevation angle is multiplied times the 
percentage of the time that angle is active to a given target area.  The 
resultant weighted average is then converted back to dB.  I agree that it 
would have been helpful if the Antenna Book had cited that article and the 
correction provided by Bruce Horn, WA7BNM, which was published in NCJ 6 
months or so later.


>8) It's not clear what antenna pattern is being modeled in HFTA.  The book
>makes a mention of a cosine squared pattern for the default 4 el beam, but
>what's the pattern for the others?  Is it a cos^N type pattern? In
>particular, what's the vertical pattern assumed.


I think the documentation states that the antennas are assumed to be point 
sources.  For this reason, neither HFTA nor TA give reliable results in 
some extreme cases, such as very closely space stacks.  There is also some 
uncertainty about the results at very low take-off angles, where the 
terrain profile is highly detailed.  I defer to Dean on the cause and 
potential solutions.

73, Pete N4ZR
The World HF Contest Station Database was updated October 29.
2469 stations are listed -- 29 new and over 100 updated.
Are you current? www.pvrc.org/wcsd/wcsdsearch.htm