Software review

VOACAP propagation analysis and prediction program (IV)

Forecasts and outputs

When all data are entered, you can at last request a forecast. That means first to calculate the propagation (via the "Run" dialog) then display the output (directly through the "Parameters" windows that will pop-up or via the "View" dialog).

If you changed your data and ask directly for a chart, the new data will not be taken into account. On the contrary, if you have already ran the calculations earlier and have not changed your data but want simply display another parameter, at this sole condition you can directly output the result.

Calculations made by VOACAP can be displayed either in tabular form (line printer style reports) or, easier to read, in graphical form showing iso-contours of parameters on a Frequency vs. Time plot.

To launch the calculations you need to click on the main "Run" dialog and select either "Circuit" or "Chart" whether you want later to print a tabular line printer text report or display a chart.

Among the other output options name the Batch, Distance and Time plots that are much less used but that show some interesting curves too.

Run "Chart"

In selecting "Chart" in the "Run" dialog, VOACAP performs propagation calculations and creates an output graphic file named "voacapg.out" that is displayed in chart form (see below "Parameters" windows).

Run "Circuit"

You can also select "Circuit" in the "Run" dialog. VOACAP performs then the same propagation calculations that previously but it displays no data in chart forms. It only creates an output file named "modelx.dat" (e.g. voacapx.dat) and then executes the propagation module ..\BIN_WIN\modelw.exe. The results is saved in the file "modelx.out" and then displayed on screen...

Note that Methods 11 or 28, as listed in the parameters, produce outputs that will be displayed graphically, rather than in a report.

"Parameters" window

When the calculations are completed (usually that last between a few seconds and 20 seconds or so), the "parameter" window will pops-up. Depending on the model this window will display a more or less extended list of functions that you can select and display in a chart form.

Not all parameters are easy to understand without good knowledges but you can learn much things on propagation if you study seriously these parameters when working on the air or just after. 

The most useful are the signal-to-noise ratio at the required reliability (SNRxx) and power signal lower decile, both expressed in dB. These parameters can bring to the fore slight variations in the signal strength in the course of the day that you can benefit to plan your activities (an increasing favorable to DXing or on the contrary a decreasing, sign of QSB and bad propagation).

Click on a parameter to display the concerned chart. Each chart provides a menu allowing you to copy the graph in the clipboard, to print it in color or B/W, to select a new parameter to display, to add custom curves (e.g. MUF, LUF, FOT, etc) or to change the scale from 2-30 MHz for example to 2-10 MHz.

Recall that if you have already ran the calculation you can simply click on the "View" dialog and select "Chart" to get the "Parameters" window. This window pops-up also in charts under the menu "Parameters".

Reliability and accuracy of VOACAP forecasts

To get a good prediction, logical, valid and thus "good" because reliable, you must not only set all input parameters correctly, but also select the output parameters that match your inputs. Indeed, some parameters depend on system parameters other are independent of these conditions.

Without set the antenna properties included in the system group for example, you can get an accurate prediction about the MUF and several methods permit to predict its frequency. The MUF and other FOT depend only on the circuit (at 3000 km from the transmitter), month, time of day, and SSN. Its frequency is not affected by the antenna design or the transmitter power for example.

Conversely output parameters like SNR, REL, DBU, RPWRG, SIGL, SNRLW are dependent on the system parameters, the signal-to-noise for example being closely related to antenna properties and the level of noise and QRM. To get a valid prediction you must set parameters of this group, as well as Tx and Rx Antenna. To achieve this you need to know the radiation pattern and the other properties of your antenna. In this purpose VOACAP provides a small application called HFANT (see below).

Among the many output parameters (SNR, LOSS, SDBW, SNRLW, SNRxx, etc), VOACAP provides REL. for reliability. Recall that his is the percent of time that SNR exceeds the “Required Reliability”, SNRxx. As we explained previously, the default required reliability is usually set to 90%, the industry standard, providing an excellent signal quality (comfortable audio, low QRN, etc). Reliability is defined as the fraction of days that successful communication may be expected at a given hour within the month in the specified conditions. Like any global forecast using median values, we must specify that reliability is the expected performance on the undisturbed days of the month, nominal conditions, and thus do not accurate forecast the effect of short-term ionospheric anomalies that are ignored in this model.

VOACAP, like ICEPAC or REC533 are really monthly median models, hence the use of statistical functions and reliability parameter. These models were never designed to predict current ionospheric conditions, near-real-time forecast or even a day-of-the-month prediction because they use anything other than smoothed sunspot numbers which accuracy increases with time considered, the longer the best.

VOACAP can handle the short-term ionospheric fluctuations but only by applying statistical factors that are based on measured SNR distributions over the month when we run the higher reliabilities. Thus, do not expected the same accuracy and the same features as a product working with real-time sounding.

VOACAP prediction compared to the one of Ham CAP, that uses also this model but with less parameters, giving thus an overview of propagation. This is a forecast calculated for a circuit between Moscow (UA3) and Belgium on August 5, 2004 at 2200 UTC on 20 meters and 100 W output (SSN = 27, SFI = 85), a minimum takeoff angle of 3° and a rural QRM in ON. VOACAP predicts a signal strength in Belgium of -140 dBW or S3 and a S/N ratio of 27 dB, thus weak. The MUF over Belgium for a 3000 km path is predicted at 11.9 MHz but DXAtlas is more acurate and predicts a MUF at 17.9 MHz. At right, Ham CAP predicts a MUF at 12.2 MHz over Belgium with a stronger SNR of 37 dB. In the field UA3 stations were indeed weak (S3-S4) but could be worked on 20 meters in SSB with few noise contrarily to what stated non-VOACAP based applications, most predicting a MUF close to 9 MHz and no reception possible across Europe. In this context, the VOACAP model shows all its utility, giving predictions much more accurate that any non-VOAVAP based application.

George Lane who devised IONCAP and wrote its manual get very concerned when amateurs use these models for very short-term predictions. We must repeat it : neither IONCAP, VOACAP or one of the additional module listed below has never been validated for that purpose. All the less when we know that these models ignore geomagnetic field indices and don't use real-time ionosonde data (but rather median values and predefined tables). Short-term forecasts have thus a limited accuracy, many disturbances are not taken into account, but these predictions are in spite of everything very useful if we know their limitations. 

For example, VOACAP doesn't forecast the auroral activity because it is not designed for automatic modeling of such an phenomenon. Unfortunately, when amateurs point their antenna at high latitudes in presence of northern lights they experience enhanced propagation to other locations. Indeed, the auroral oval acts like a reflective wall and bends sky waves along paths that are off-azimuth over very long distances. In these occasions the MUF is well above 30 MHz. Up to now VOACAP is unable to forecast these events and we must rely on other applications.

In the same way, many listeners use VOACAP to predict propagation conditions in the top band of 1.8 MHz or LW bands. Unfortunately VOACAP is ray hop model for ordinary HF sky wave propagation and has no mean to take into account propagation through the residual E-layer acting like "wave guide" when dark paths are involved in the medium wave band. According to George Lane, the only way to get a more or less realistic prediction is to run the Wang's LW model together with VOACAP and to select the prediction showing the highest signal power.

Additional modules

Several additional modules are provided with VOACAP. All are independent but are interfaced with VOACAP when you call some functions to take for example into account your antenna properties (HFANT), the receive area (VOAAREA) or interferences (S_I_VOACAP).

ICEPAC

ICEPAC stands for Ionospheric Communications Enhanced Profile Analysis & Circuit Prediction program. It is as powerful if not more as VOACAP. It is a statistical model of the large-scale features of the northern hemisphere. It uses also an improved IONCAP model but includes in addition the ICED (ionospheric conductivity and electron density) model. It provides thus some additional displays not included in VOACAP. ICEPAC recognizes the different physical processes that exist in the different regions of the north hemisphere ionosphere and contains distinct algorithms for the sub-auroral trough, auroral zone, and polar cap.

A propagation prediction calculated for a path between ON and ZD9 (Tristan da Cunha) at 9906 km away during the solar maximum using Method 20, reliable. At left calculated by VOACAP, at right by ICEPAC. These graphs show the typical difference between both programs : ICEPAC is more conservative, predicting a signal in a range between 1-4 dB or half an S-point less strong during an interval time of half an hour against VOACAP; not much more than a hiss. To compare with the prediction displayed below using Method 30.

However, ICEPAC was no as extensively validated as VOACAP and unlike this latter, it doesn't take into account a smoothing function between short and long paths (ray hop vs. forward-scatter regions) for all distances between 7 and 10,000 km. Once your reach these distances there is abrupt transition between short and long paths. Therefore, in some circumstances, if you calculate for example a prediciton between Belgium and Tristan da Cunha (ZD9) as below, the ray hop model finds an ionospheric control point with a discontinuity of nearly 10 dB at mid-path, causing artificial SNR losses exceeding 23 dB, whereas the long path model used by VOACAP finds a weak but detectable signal which is actually there, assuming a scatter propagation mechanism. This is for this reason that in ICEPAC it is well stated beside the Method 30 "For VOACAP only - S/L path smoothing (7-10000 km)".

Same prediction as above, calculated for a path between ON and ZD9 (Tristan da Cunha) at 9906 km away during the solar maximum but using Method 30 this time, not reliable at that distance. At left calculated by VOACAP, at right by ICEPAC. These graphs show the typical difference between both programs. Both give a similar signal strength at 22h on 14 MHz, about -138 dBW within 1 dB. But ICEPAC is in error when predicting a signal of -168 dBW at 14h on 28.5 MHz. According to VOACAP the signal is 23 dB stronger, -145 dBW ! A little bit more than a hiss this time, about 4 S-points and thus enough to work ZD9...

On another side, ICEPAC provides some additional iso-contours maps not available in VOACAP like the SRNxx (Method 20). Results are more accurate and the graph also more complete, but you must well known limitations of this program.

REC533

Its name comes after the ITU-R Recommendation P.533-6. This is a derived version of VOACAP offering some additional functions dedicated to S/N estimations, distance and time.

HFANT

HFANT is an add-on to VOACAP and shares its data with it as soon as your antenna properties must be taken into account to calculate a circuit. When launched in stand-alone this graphic tool is able to display the radiation pattern of antennas in both azimuthal and elevation planes.

In addition to propagation forecasts, the VOACAP package provides free HFANT that permits to simulate the radiation patterns of antennas. Above left the settings for a 3-element Yagi placed 10m high and cut for the 20m band. A right the azimuthal radiation pattern of this antenna

All sample files can be modified to suit your need as display above, showing the settings and radiation pattern calculated for a custom 3-element Yagi placed 10m high and working on the 20m band. This module will be very useful to all amateurs wishing to buy EZNEC but who cannot pay for it. This version shows an accuracy very close or even identical to this latter (or any other NEC-based modeling program) and it is free !

Here is for example a screen dump showing the largest variations between HFANT (ItsHF) and NEC models when computing the radiation pattern of a 3-element Yagi. But these differences are mainly due to the fact that the HFANT Yagi used "thin" wires while the NEC Yagi used 0.5" diameter tubing. When both models used the same specifications, graphs are similar over 99%, as diplayed in this second screen dump showing the radiation pattern of a 1/2l-long dipole cut for the 20m band and tight 15 meters high (50 ft).

So, avoid to enter any one data in VOACAP (System and Tx Antenna groups) without have first simulated your antenna in HFANT. In doing the contrary you have all chances to predict anything as you have not taken into account your real antenna properties.

VOAAREA

VOAAREA is dedicated to point-to-area coverage analysis. It was mainly developed for broadcast purposes but can be used for amateurs activities as well. The forecast is centered on the receiver location and displays in addition the size of the plot area. This is a very interesting module that displays most output parameters of the VOACAP model (e.g. SNR, SNRxx, SDBW, etc). It permits to better understand coverage problems at long distance according the radiation pattern of your antenna and to see how strong can fluctuate your signal around the receiver location (e.g. over 10 dB over a distance of about two thousand kilometers). 

At left a Point-to-Area map created with VOAAREA for a transmission made from ON on August 2004 on 20 meter taking into account the propagation conditions and a SSN of 85, an SNR of 50 dB and an SNR req.reliability of 50%. The power was set to 100 W using a G5RV dipole tight in east-west direction. The grid selected (for the map) was "1 = Lat/Lon" and set to -50/+65° of longitude and 0°/90° of latitude. To get results similar to the ones of MultiProp (right) I used the CCIR coefficients and I didn't use the "day-of-the-month" either (e.g. 8.15) to avoid VOAAREA to switch to URSI coefficients. The layer CIRAF zones was removed for more clarity and the grid size was set to a "low" resolution (61x61), eleven times lower than the maximum (361x361) available in this application. At right the same simulation calculated with MultiProp that uses MultiNEC with VOACAP. The grid size was set to the highest value, 61x61 in this case. Even at lower resolution (30x30) VOAAREA results are a bit more accurate than ones of MultiProp. Note that both map projections are different, as well as contour levels. These maps match what I experimented in the field. Recall that a signal power of -120 dBW ~ S5.

The menu differs slightly from VOACAP. To get a result you have first to setup correctly your circuit parameters like in VOACAP. In addition for more clarity remove for example CIRAF zones ("Layers" group, select "ignore" for the field "CIRAF zones"), select the grid to display ("Grid" group and select "21 =  NE hemisphere" or "1 = Lat/Lon" and enter in the "Plot" group the maximum width of the map, e.g. -70/+70, 0/+90) and at last select the grid size (default is 31x31) knowing that over 200x200 calculations exceed 1 minute on a 500 MHz computer.

When all parameters are set, on the menu, select "Run", "Calculate >" then "Save/Calculate/Screen" to start the calculation, save the output (*.vg1) and display immediately the map. If you want to display the map later, select "Calculate >", then "Save/Calculate" to start the calculation and save the result in the *.vg1 output file. Later select "Run", "Plot result", "File" and open the file previously saved. Then select "Plot to" and "Window". You will get a map similar to one display above left or at another resolution.

S_I VOACAP and S_I ICEPAC

These two models are especially designed to predict the signal-to-interference ratio. Inputs are two transmitters and one receiver station, and you have the possibility to change parameters like the receiver bandwidth, required S/I protection ratio, S/I probability. or to set SNRxx minimum value below which the S/I calculations are skipped. The program calculates then interference parameters and displays its results using the same graphs as VOACAP or ICEPAC.

Support

If you experiment any problem using the VOACAP package, do first check the readme files provided with the product. Then send all required information by email to Rob DeBolt or Paul McKenna at NTIA/ITS who maintains the three main applications VOACAP, ICEPAC and REC533.

However, as explained below, the new VOACAP online replaces advantageously the original version.

My final impression (before 2010)

VOACAP counts without any doubt among the few seldom powerful and flexible analysis and prediction program available. VOACAP is the most accurate HF prediction tool due to the extensive adjustments made during it's creation, as predictions were compared with a great body of worldwide Voice of America listener reports. Also, time passing, there were some additions to improve its accuracy - like the smoothing function of Mode 30 - that do not appear elsewhere. As a result, VOACAP is rightfully considered as the "gold standard" of propagation programs. With time it shows however some ripples.

As it, the graphic user interface of VOACAP is outmoded. It is not too bad but not really user-friendly either, essentially when you need to display an output parameter or modify a chart. Hopefully, since its release other programs have filled this gap and have provided a better user interface (e.g. MultiProp, ACE-HF Pro, PropMan-2000, Ham CAP, DXAtlas, WinCAP Wizard, etc).

Released in 1997 on a design written in Fortran 13 years earlier, VOACAP is aging. Some years ago I wrote here : "It should be welcome to design a complete new product taking advantage of new programming languages and algorithms, a down-sized version of the IRI (signal and noise models) and IGRF (geomagnetic) models and why not, cherry on the cake, real-time ionosonde data to improve its accuracy". We will see that if this time is not arrived yet, some improvements have been done !

Let's take some examples of non-fulfilment of of the original product. As we noted, VOACAP does not use geomagnetic indices. Even some major competitors consider them (Ap and Kp) as "useless". But they are wrong because either they have not seriously learnt propagation fundamentals or they never worked in conditions in which the Kp index enters into play.

All geophysicists might tell that the p-index is however useful to calculate many geomagnetic effects. If Kp index doesn't improve the accuracy of short-term forecasts, it highlights disturbed conditions like a noise makes us prick the ears and the dependency between the geomagnetic activity and the depletion of the F2 layer. Then Kp index helps estimating the geomagnetic influence on propagation at high latitude during auroral events. It also permits to predict propagation on the top band where magneto-ionic effects become important (as are the sunset/sunrise and weather conditions, two parameters ignored by most programs as well).

At last, Kp index permits to check antenna performances according to their polarization (we know that at low latitudes, vertical antennas for example don't couple properly with the horizontal component of the magnetic field, etc). Of course, everybody uses VOACAP for short-term predictions too, but we know its limitations and what mean probabilities, or we should.

VOACAP Online

We can say that since 2010 a fresh wind full of good feelings has blown on VOACAP, almost a revolution !

Indeed, since 2010 Jari Perkiömäki (OH6BG/OG6G), James Watson (HZ1JW) and Juho Juopperi (OH8GLV) have taken the responsibility to improve the VOACAP interface. They have included several very awaited modules among them the most interesting is the online interface, VOACAP.com, that allows all Internet users to get a point-to-point prediction for free without having to enter all data manually. This interface is available for the three main platforms : Windows, Linux, and Max OS X.

The User Guide is available. In 2012, Jari added the Coverage Area Map, then a Planner very appreciated by contesters. In 2015, the DX Chart was added for DXers. In addition a VOACAP blog is opened since 2012 where users can exchange ideas with developers and keep informed of the new releases and enhancements.

Also note that since that time VOACAP is included in the new DX Summit cluster (when you select a call sign, also select "Show VOACAP Predictions") to get a point-to-point prediction for your QTH.

Below are some screenshots of the new VOACAP online. The evolution is not only obvious but the ease of use is without comparison with the old interface. This time, even the youngest radio operator can use it and get a prediction in a few seconds !

The main functionalities available in VOACAP online supported by NTIA/ITS (engine) and OH6BG's team. From top to bottom, the Point-to-Point Prediction, the Propagation Planner, the the Coverage Area Map and the DX Chart.

However the IRI model as well as geomagnetic indexes (Ap and Kp) are always excluded from the interface (the VOACAP engine remaining the "classic" one) while SSN can be entered manually as in the original version (you can get the data set in subscribing to IPS maillist, see this article).

If the new online version does not reflect the propagation conditions in real-time, using accurate data and specially an accurate SSN when the sun activity is very high, VOACAP can calculate reliable preditions, even if they remain averaged and statistical data.

In all cases, with all these new options VOACAP online is became a must for all fans of propagation. Congrats to Jari and his team for all their efforts !

Other good news, some amateur products are till more accurate in some specific modules, to name DXAtlas designed by Alex Shovkoplyas, VE3NEA, that is integrated with HamCAP (that uses the VOACAP engine), IonoProbe and Band Master, and includes quasi real-time ionospheric maps.

The other "competitors", those who do not use the VOACAP engine, should follow this idea if they want to sell other thing that low-end and useless products in the future.

To be frank, today a radio amateur should only use the VOACAP online interface (and gave up ICEPAC) each time that he needs of an accurate forecast. However he will probably have to rely on other products when for example he look for the signal strength of beacons or to get online the latest solar, geomagnetic and spaceweather data or alert bulletins.

Indeed, no products could provide all these features. Currently amateurs need at least of two or three applications, including web-based, to get a complete overview of propagation conditions.

Bottom line, if you don't want to invest some money in a prediction program and have an Internet connection, go with VOACAP online (without connection use e.g. HamCAP); it is free, fully supported, and with all its modules, not only it is more complete than its embellished competitors but it is also more accurate than all no-VOACAP based programs. A good alternative that seems more accurate is Proppy.

Whatever the version you use, have good DX !

For more information

What can we expect from a HF propagation model ? (on this web)

SSN or the short history of the Smoothed Sunspot Number (on this site)

Proppy (an alternate to VOACAP using the new ITURHFProp model)

For more information about VOACAP, read L.R.Teters, J.L. Lloyd, G.W. Haydon and D.L. Lucas, "Estimating the Performance of Telecommunication Systems Using the Ionospheric Transmission Channel" (5.4 MB PDF file), Institute for Telecommunication Sciences, NTIA Report 83-127, July 1983.

VOACAP engine to download from Greg Hand's website

VOACAP Online (the online version), OH6BG, including all modules, a guide and a list of common mistakes made by users

VOACAP Mailing List (to be informed about all functionalities of the product)

NTIA/ITS (the developer's website)

User's Guide (book), by George Lane/Rockwell Collins 

ICEPAC technical and user's manuals, by NTIA/ITS

Broadcast analysis and prediction in the HF band, IEEE Xplore

Some VOACAP-based programs:

VOAProp, by G4ILO (replacing HPProp)

ACE-HF Pro, by ACE-HF

HamCAP and DXAtlas, by VE3NEA

MultiProp, by AC6LA

PropMan-2000, by Rockwell Collins

WinCAP Wizard, by Jim Tabord Software (closed down).

Back to Menu

Page 1 - 2 - 3 - 4 -