Last month we introduced a little of the practical aspects of radiowave propagation and outlined the criteria for making comparisons among software packages designed for HF propagation prediction. This month we will begin a series of software package reviews by taking a look at a software program (actually a suite of software programs) from the Voice of America (VOA) of the US Department of Commerce. The name of the program is VOACAP and there are two versions. The earlier version, which is now frozen in development and will not be further revised, is a DOS-based program. Recently, VOA released the first version of a Windows-based program also called VOACAP. Development of this software is an ongoing project and new versions of the program are released from time to time. All future development of this software will be limited to the Windows version.

There are many good reasons to review VOACAP first. One good reason is the price of the program - which is free (unless of course you are a US taxpayer). It can be downloaded by anonymous ftp from the VOA ftp site (ftp.voa.gov in the folder /pub/software/VOACAP). The latest Windows version is voawin.exe and the final DOS version is voados.exe. A second reason is that VOACAP is one of several software packages that are based on a program called IONCAP developed in the late 1970's by the Institute for Telecommunication Sciences (ITS) of the Department of Commerce. Actually VOACAP and others are "shell" programs that simply act as a user interface for IONCAP and run it for you. IONCAP is probably the best studied propagation program and is arguably the best basis for judging other programs. Thus, we will begin our review of VOACAP with an introduction to IONCAP as seen below.

Since the late thirties, many different organizations have been involved in the study of HF spectrum radiowave communications. A worldwide effort to measure ionospheric parameters, including noise, was established and detailed records have been obtained for variations in system performance over various paths. All of this research has shown that HF system performance is related, in a very complex manner, to solar activity, time of the day, day of the year, and the details of the radiowave path. In 1978, ITS released a FORTRAN program called the Ionospheric Communications Analysis and Prediction Program (IONCAP). Prior to the release of IONCAP, much of the path analysis that was done had to be handled manually - a very time consuming process. IONCAP was written in a modular format which allowed essentially separate development of models for the key parts of the program.

We won't go too far into the details of IONCAP, but it is instructive to consider some of them. For example, there are separate subroutines in the IONCAP program for antenna analysis. For any path, the gain of the antenna in the direction of the path and at the elevation of the specific signal needs to be considered. In the earliest versions of IONCAP only simple antenna geometries were included but, since it is fairly easy to extend a modular program, VOACAP and other software offer more complex antenna geometries, or the opportunity for you to quantify your own particular antenna system. We'll talk a little more about this later.

An important subroutine in the IONCAP program is the ionospheric parameters subroutine. Explicit electron density profiles were included in the form of look up tables, rather than mathematical approximations for these important parameters. But keep in mind that these kinds of approximations are not all bad, which we will discuss in more detail in a few months when we look at some simpler (and thus faster and less computationally intensive) programs. The concept of look up tables for very complex systems can increase computation speed and allows the use of alternate ionospheric models which we will describe a little later in the context of VOACAP.

While IONCAP is the standard for judgment, it isn't perfect and has some limitations which may, under certain circumstances, be important. For example, IONCAP breaks the year up into twelve months but no further. Thus conditions near the end of the month may look like an average of the results for the present month and the next. IONCAP is designed around the 12-month running average of the sunspot number, not the day-to-day measured solar flux. During sunspot lows, which is the present situation, this doesn't matter much but near sunspot peaks the differences can be large. IONCAP also does not include geomagnetic effects related to the A- or k-indices. Some of the shell programs, however, do include corrections for this when high latitude paths are involved.

An important and necessary limitation of IONCAP involves paths longer than 10,000 km - which are often important for DX considerations. The IONCAP single hop model (distances less than 3,000 km) is very complete and considers all possible ray paths for the circuit. But extension to paths that require three or more hops seldom indicates that there is a path available, which we certainly know to be incorrect. IONCAP uses a correction for these multi-hop paths longer than 10,000 km. Empirically it has been shown that these circuits are dominated by "control areas" which are the regions within 2,000 km of either end of the path. If a propagation path does exist, it is because the control area at the transmitter end allows the skywave to be launched (first hop) and the control area at the receiver allows what is left of the signal to be returned (last hop). In between, the path can be characterized simply by a loss per distance function and the noise and signal statistics are the same for all paths. This approximation is quite valid and well tested, but implies that one should recognize that the physics is more complete in the model for shorter paths.

When personal computers became powerful enough, IONCAP was incorporated into a form that would run on a PC. The first versions used text input files that mimicked the old FORTRAN punch cards with rigorous requirements on the form and position of the data on each line of input. Output consisted of long printouts of data, also in text format. It didn't take too long for programmers to create shell programs like VOACAP to make data entry easier and to put the tabular output into a graphical form.

Let's take a look at VOACAP. Because the DOS version is frozen in development, I'll only describe the Windows version, but they really aren't that different. After you download the file named voawin.exe, simply run the program in some temporary directory on your hard drive. Note that a total of more than 26 MB of hard drive space is necessary to get the program running. After executing voawin.exe, this directory will have 11.5 MB of files in it but you can delete voawin.exe if hard drive space is really critical. Then run install.exe from this temporary directory. The install program will create and fill up a directory called "itshfbc" (about 14.5 MB). The files in the temporary directory can be deleted at this point. You're done!

The install program will have created a program manager group containing seven programs. There is the usual readme file and a viewer for the collected release notes for VOACAP for Windows called News. There is a program for creating custom antenna profiles called HFANT. VOACAP is the basic point-to-point analysis program, VOAAREA is a version of VOACAP for analyzing point-to-area coverage, and S_I VOACAP is a version for calculating signal-to-interference profiles.

Figure 1

After you've launched VOACAP, you'll get the data input screen shown in Figure 1. All of the input screens have modest balloon style help for the input window corresponding to the mouse pointer. Since most of the input criteria are common to IONCAP-based programs we'll discuss this screen in some detail. Much of this discussion will also apply to other programs later. IONCAP has a number of different methods but only three of them are commonly used. Method 20 is the complete system performance, method 21 is the same but forces the long path, and method 30 is unique to VOACAP and smoothes the transition between the two IONCAP models in the distance range of from 7,000-10,000 km. You have two choices for ionospheric coefficients - CCIR (default) or URSI. We may discuss these in a later article.

The menu choice Groups allows you to choose a number of different months and the associated sunspot number (SSN). This is not the same as the solar flux index given by WWV at 18 minutes past the hour. VOACAP is designed around the 12-month running average of the sunspot number. The average solar flux (SFI) and average sunspot number can be related by

SFI = 63.7 + 0.728 SSN + 8.9x10-4 SSN2

During periods of high solar activity, the running averages can be very different from the daily values while during low activity they don't differ much. When in doubt, run the program for different values and see what happens.

Transmitter and Receiver can be selected from files called transmit.def and receive.def in the userdb directory. These are text files and can be modified by the user as long as the format is strictly maintained. I have created a set of these files from the CT contest program DX countries database file and will try to find an ftp or web site for those of you who want copies of them. Path can be used to select either short path or long path. Freq(MHz) can be used to select the default short wave frequencies or either of two user-definable frequency lists. These frequencies do not affect the calculations - they only are used for marking the scales on the output graphs.

We only need to concern ourselves with three values under the System prompt. The noise parameter you choose depends on the receiver location. The help menu will give you the information you need to select a value. Required reliability was described last month. 90% is the default and is most useful for contests and DXpeditions but 50% gives you an idea of what may be possible. I usually run the program for both. Required S/N ratio in IONCAP is normalized to a 1 Hz bandwidth so the value inserted must include the bandwidth of the receiver which, of course, depends on mode. For SNR use

SNR = 10 + 10 log (bandwidth in Hz)

A 500 Hz bandwidth, CLOVER for example, gives a required SNR of 37 dB.

Most users can simply use the default parameters for Fprob. TX Antenna allows up to 4 frequency ranges to be chosen with a different antenna for each. There are 70 different antenna files that can be utilized or the user can create a custom antenna file using the HFANT program described below. Only one RX Antenna can be specified. This may seem a little odd at first but consider that VOACAP was written for broadcast stations not hams. Of course, broadcast stations build complex transmit antenna arrays for specific frequencies and test their systems assuming some basic receiver antenna, such as an isotropic element. From the ham point of view, the choice of antenna and choice of which station is the receiving station is much more complicated. The path between the two stations is not reciprocal, in the sense that you may have different power output at each end. Also, you may not know what kind of antenna the DX is using or even where it's aimed. I usually assume the basic rule that you have to hear 'em to work 'em and pick KA6A as the RX station. Then I will use realistic receive antennas and guess conservatively on the transmitter antenna and power. Thus, for my purposes, VOACAP would be more convenient to use if it allowed for more RX antennas as well.

Figure 2

Once the data have been provided, it's easy to run the program. Simply choose Graph from the Run menu. When execution is completed, you will be given a window full of choices, shown in Figure 2, from which to select the graph output. There is a lot to learn from each of the graphs but, for illustration purposes, look at SNRxx as shown in Figure 3. SNRxx is the signal-to-noise ratio at the probability you specified in the data input window. The solid line is the MUF and the color-coded graph shows you the SNR at any given time and frequency. The horizontal gridlines are the frequencies you specified earlier.

Figure 3

That's a capsule view of VOACAP. Let's take a look at the other programs included with the package. The operating window for HFANT is shown in Figure 4. The simplest way to use this program is to open a similar antenna file, modify it however you like, and then use the Save As function to leave the original file unchanged. I created a new directory call Amateur to keep track of my versions. Look at the HFANT plotting capability. It's great.

Figure 4

VOAAREA is a program that at first glance makes more sense to broadcasters than hams but can be a very useful educational tool for hams. The data input screen is similar to the other VOACAP data input screens. It allows you to specify the transmitter location and the receiver plot center as well as the dimensions of the plot area. An example output window is shown in Figure 5. This is for a transmitter located in Libya and a plot area comprising the continental US. You can see from this graph why we in the Society of Midwest Contesters call the upper Midwest the Black Hole. The SNRs in both 2-land and southern California are 10 dB greater than in 9-land!

Figure 5

The last program in the VOACAP suite is S_I VOACAP, a program that calculates the signal-to-interference ratio. You specify a receiver and two transmitters and the program runs VOACAP for both and automatically calculates interesting interference parameters. This is a brand new part of VOACAP and offers enough potential that I think we'll cover it by itself in a later column.

The VOACAP suite is powerful and certainly makes a useful tool for analyzing radiowave propagation. But there are some weaknesses. The program runs in a reasonable amount of time on a 486-DX2-66MHz machine. A slower machine may be frustratingly slow. This program was developed by a US government agency and, like the government, it has gotten huge. The compressed file is almost 6 MB, so downloading it via modem is going to take a long time. Even after the program is completely loaded and temporary files are deleted, it takes almost 15 MB of hard drive space! Now, there are a large number of data files included that could be eliminated, but that takes time and some experience with the program. I've already mentioned that geomagnetic activity is not considered. In fact, the readme file explicitly warns against using VOACAP for short term prediction. It can indeed be used for short term predictions, of course, as long as you understand the limitations. I have a mixed opinion of the graphics. The color scale 3-dimensional graphs work fine on the screen but do not lend themselves to black and white printing. A black and white switch is included in the program but the results are unimpressive. The program does not have single band output graphs - a feature I think is important.

My overall recommendation? C'mon...the program is free! Seriously, if you have the disk space and a reasonably powerful machine, this one is worth getting and using. Is it the best program available? Other programs do some things better, even much better in some cases, but all things considered (including price) this might indeed be the best program available.

Next time we'll look at another IONCAP shell program called CAPMAN. If you have any specific questions related to radiowave propagation, send them to me and I'll try to answer them either as a short answer at the end of one of these articles or as the basis for a later article. And, as usual, if I don't know the answer, I'll try to make up something plausible. Hi!