VOACAP
                    FREQUENTLY ASKED QUESTIONS
                       (Revised 22-FEB-96)


Questions about VOACAP may be submitted to Hien_Vo@BENG.VOA.GOV.
Answers will be periodically posted in this FAQ file.


1.  What is VOACAP and why is it better than other models?

VOACAP is a modified version of IONCAP.  It is a model that has
been under development by the U.S. Government since 1942.  The
strength of the model is that it uses world maps of
ionospheric parameters to construct the ionospheric path and uses
path-specific statistics to evaluate the system performance
factors.  There is no claim that it is better than any other
model.  "Better" is usually in the eye's of the user and is based
on how well a model works for a particular application.  IONCAP
was selected by the VOA in 1985 because it provided the system
performance analysis capability we needed for design
specifications and it had a proven track record.

2.  How does VOACAP differ from IONCAP?

Minor changes were made to the logic to correct for errors which
could occur if only one mode was found, the mode was
above-the-MUF and the path had 3 or more hops.  Signal power
values in the working bands predicted by VOACAP differ from
IONCAP by more than 3 dB in only 1.5% of the cases. Very large
differences can be found at frequencies above-the-MUF and below
the LUF.

The LUF and the MUF are the limiting frequencies which define the
coverage footprint.  The changes made to VOACAP are believed to
improve its ability to predict area coverage.

No changes were made to the subroutine in IONCAP that computes
the actual value of the maximum usable frequency for the lowest
order mode.  The values of the system parameters (i.e. signal
power, reliability, etc.) may be different.  

A math error in the formula that is used to calculate Reliability
and Required Power Gain was corrected.  This change makes VOACAP
less optimistic than IONCAP by about 2 to 4 dB.

A smoothing function was added as a transition between short path
and long path predictions.  It is activated when the user
specifies Method 30 and when the path distance is greater than
7,000 km.  The parameter which is smoothed is the signal power
distribution.  All other parameters which depend up the received
signal power are also smoothed.

The most significant differences between VOACAP and IONCAP are in
the input and output processors.  This version of VOACAP includes
the ability to create area coverage maps in B&W or color.  Also
it has an atlas of city locations, country boundaries and CIRAF
zones for international broadcast applications.

In addition VOAAREA, the mapping program, will allow the user to
overlay maps and display the sum, difference, greatest value,
least value and the signal-to-interference ratio for the case of
a wanted signal and one interfering signal.

3.   Are the changes to VOACAP documented anywhere?

Yes. All of the recent changes since May 1993, have been
documented in the HFMAP Newsletter.  A summary of the changes
with cross references to the newsletter are contained in
PC-VOACAP   NEWS which is also posted on this VOA Internet
server.  The original changes made from 1984 until April 1993 are
documented in Broadcast Engineering Technical Report 01-93,
"Voice of America Coverage Analysis Program (VOACAP), a Guide to
VOACAP."   It also provides flow charts for original IONCAP and
VOACAP showing the order in which the subroutines are called and
for what purpose. This document (PB93-155174) is for sale from
the National Technical Information Service, 5285 Port Royal Road
Springfield, VA 22161 USA (Tel: +1 703-487-4650)             
(FAX: +1 703-321-8547).


4.   Is VOACAP completed now?

No.  We are working on signal-to-interference point-to-point and
mapping capability, and possibly providing NEC capability to
VOACAP. This work should be completed in 1996.

5.  How can I get updated version of VOACAP?

Updated binary versions of VOACAP will be available on
FTP.VOA.GOV

Please periodically visit VOA Gopher to check for the updated
VOACAP versions.  Make sure to read the README file for news
concerning with the updated version.

6.   Can I obtain source code for VOACAP?

Yes.  One can find the FORTRAN source code of VOACAP in the
subdirectory named VOACAP.

7.   I loaded VOACAP on my computer and nothing happened.  What
should I do?

Your computer must contain a math co-processor and 500 free
Kilobytes of RAM.  You must update the PATH statement in your
AUTOEXEC.BAT file, and have ANSI.SYS defined as a device in your
CONFIG.SYS file.  See the readme file for details.

8.   Will VOACAP run on any machine?  Or just on a PC under DOS? 

VOACAP DOS-Version runs only on DOS machines.  One can get the
FORTRAN source code from the subdirectory named VOACAP, and adapt
it to run on any machine.  Users attempting this feat are
encouraged to thoroughly benchmark results with the PC code.

9.   Who should use VOACAP?  Is it better than some of the
smaller and easier to use prediction programs?

VOACAP is intended for the user wishing to estimate the region of
coverage for a frequency, hour and monthly smoothed sunspot
number for the period of a month.  It is not accurate for a
prediction of a single day within the month.  There are other
models which are more accurate for frequency prediction on a
daily basis.  They generally take into account solar flares,
magnetic disturbances and daily sunspot
numbers.

10.   I regularly make plots of parameter SNR90 from the area
coverage program.  What is this new parameter SNRxx, and is it
better to use it?

SNR90 was specifically defined as the signal-to-noise ratio at
90% reliability.  SNRxx is the SNR at the user's specified
required reliability.  Now you can make maps at any reliability
value you wish. The user is warned that VOACAP  behaves strangely
when attempting to calculate maps of SNR90.  Be sure to use
SNRxx.

11.  Are the methods that predict HPF, FOT or LUF in VOACAP
reliable to use?

You have hit one of the problems of software development.  We
only benchmarked Methods 13, 14, 15, 20, 21, 22,and 25.  When we
first started debugging IONCAP we found that changing the Method
could result in different predictions.  So we concentrated on the
above methods and did extensive testing to assure consistency in
the predictions.

There are some real logic problems when using terms such as HPF,
MUF, FOT and LUF and the IONCAP/VOACAP prediction program. 
Because IONCAP has an above-the-MUF model, it is possible to have
sufficient system power gain to have reliabilities of 90% or
higher at frequencies above the MUF. Thus, the FOT may be above
the MUF.

Also since IONCAP is a quasi-ray trace model, there may be
frequencies that just plain won't propagate between points A and
 For example, one can have a frequency gap between the LUF of
the 1F2 and the MUF of the 2F2.  Higher frequencies work just
fine via the 1F2 mode and lower frequencies are good via the 2F2
mode.  But in that frequency gap nothing will propagate.  You can
see that on oblique ionograms or when using ray trace programs. 
If one is predicting the LUF, IONCAP may have two or more LUFs.  

Because of the confusion with these terms and the way IONCAP
makes predictions, we stayed away from any methods that predicted
HPF, FOT or LUF.  In the process of cleaning up various
subroutines and making corrections to a few logic errors in
IONCAP, the Methods we ignored began to deviate greatly.

We had wanted to eliminate those methods but others felt we
should leave them in just in case some one later wanted to fix
them.  The Users Manual has a warning statement that we never
tested the other methods for correctness.

Although HPF, FOT, MUF and LUF have intuitive meaning to many old
time HF users, there is no general definition for these terms
that has a mathematical definition.  Last time I looked, I found
at least 6 different definitions of MUF.  For example, in IONCAP
the MUF is the estimate of the monthly median junction frequency
for the ordinary ray.  It is amazingly accurate, I think, based
on some experiments we reported in the HFMAP Newsletter.  But it
certainly is not accurate to the exact hour.  There is at least
30 min to an hour of uncertainty in the predictions, especially
at the pre-dawn dip.

If you really want to find what VOACAP thinks the HPF, MUF, FOT
and LUF are, run Method 20 with lots of frequencies.  

- The HPF is approximately the frequency where MUFday is 0.10.

- The MUF for the lowest order mode is listed in the first column
on the left side of the printout.  It is defined as the frequency
having MUFDAY = 0.50.  Method 25 will let you find the MUFs for
higher order modes.

- The FOT is generally thought to be a system performance
parameter as well as dealing with the probability of ionospheric
support.  Therefore the FOT is the highest frequency having a
circuit reliability of .90. Thus the FOT is dependent on system
gain and the user's required signal-to-noise ratio.

- The LUF is usually thought to be the lowest frequency of the
working band and is given by the lowest frequency having a
circuit reliability of 0.90.  Again, like FOT, it is system
dependent.

Greg Hand at NTIA/ITS has found an error in Subroutine LUFFY
which resulted in very strange values of the LUF found in Methods
16 and 28.  He has corrected the program error but the
predictions of LUF may still be suspect for the reasons mentioned
above.

12.  What is Service Probability and how can I use it?

The Service Probability calculation in IONCAP and included in
VOACAP is a place holder for a future model.  It was a first
attempt at computing a confidence factor for the reliability
value.  In other words it is the probability that the hourly
reliability value will be achieved over the days of the month. 
For example, if the reliability for a given circuit at a
specified frequency, hour, month and sunspot number is 90 percent
and the service probability is 70 percent, then one would expect
that the required signal-to-noise ratio will be equalled or
exceeded on 27 days of the month on 70 percent of the possible
monthly attempts over numerous sunspot cycles.

Although the concept is well founded and useful in life-cycle
management of a system, it was based on the best guesses as to
what the standard error of estimate is on all of the variables
used in calculating reliability.  Since the estimates of error
are SWAGS, one really can't put much faith in the SERVICE
PROBABILITY values.  At this point in time some two decades
later, we are no closer to knowing what the errors of estimates
are on such things as MUF variation, deviation from the predicted
absorption, ad infinitum.

A very general rule of thumb for a VOACAP user is to be wary of
any predicted reliabilities that fall below 70 percent or whose
service probabilities are less than 50 percent.  In these cases,
the user can assume that the reliability is the "best guess"
given the available data as to what will happen on that circuit
hour over an undisturbed month.  Some questions are better off
not being asked.

13.  So how good is the reliability value from VOACAP?

It all depends on how carefully you modelled the system and how
closely the ionospheric and geomagnetic conditions fall to those
assumed for the month.  If the antenna patterns reasonably
represent the actual antenna and if your assumed required
signal-to-noise ratio is really acceptable, then the predicted
reliabilities can be achieved for those cases where the general
rule of thumb, described above, is applied.  Testing at VOA using
professional shortwave monitors have shown that the VOACAP
predictions are very accurate.  Basically, if you design a system
and operate it on the correct frequencies, you will achieve or
exceed a predicted reliability of  90 percent.  That is about the
highest reliability you can realistically achieve in the typical
HF radio system over a long period of time.   Your chances of
success drop off rapidly as the predicted reliability falls below
the 90 percent range.  A reliability of 50 percent might result
in a situation where no success is achieved over the month or one
in which great success is achieved or anything in between.