Contacter l'auteur / Contact the author

Recherche dans ce site / Search in this site







Software review

DX ToolBox propagation analysis and prediction program (III)

Propagation Map

The “Propagation map” window displays an estimation of propagation conditions at earth scale rather than between two stations. A gray scale is incrusted on the world map centered on your home location, which brigthness and extent depend on all involved parameters.

The required information are the solar flux, the working frequency expressed in MHz, the transmitter power level expressed in Watts, the date and time in UTC. Some of these data can be extracted from the “Current Conditions” window in pressing “Now” or entered them manually to simulate another time and/or another solar flux.

Once this information entered, press the “Update Map” button to get an estimation of the propagation. This forecast displays also in the upper bar the estimated signal strength in dB and the number of hops between your “home” location and any location pointed on the map.

The system makes the assumptions that the receive bandwidth is about 2.5 kHz, and the minimum sensitivity of the receiver is -123 dBm (~2 μV), typical for most modern receivers.

The “Propagation Map” and charts calculated with DX Toobox. At left, for a circuit between LX and CE on August 2016 using a vertical antenna in SSB with 100 W. At right, another prediction calculated for the 20-meter band for a circuit between ON and VU for August 2004.

This map uses an increment of 1 hour only; the forecast of 1030 UTC for example is thus the same as the one of 1000 UTC or 1059 UTC; it a pity because you have no possibility to "intercept" a sudden opening in real-time like that occurs quite regularly on the low bands. The same step is used to display the propagation chart. Of course no one program is today able to take into account real-time conditions. But it should be interesting that a forecast is be able to predict opening in a time-lapse shorter than one hour.

This map is also an educational mean to understand what is the skip distance (or silent zone displayed in dark gray) around your emitter according the frequency used or to know how many hops there are between your station and a remote one.

This window cannot be adjusted, only moved or closed, and fills the 2/3 of an SVGA screen. It is thus superimposed to the Grayline map or any other chart and you cannot always click on some area hidden behind other windows. This is annoying when you want to display several windows (even with two it is a problem) or run simultaneously another application as it will fill at least half of you screen as well. A variable-size window should be welcome and very appreciated.

The accuracy of the prediction is however more than disappointing as the system uses very few inputs and simple ionospheric algorithms. The propagation map works better on the highest frequencies with a high solar activity. It also displays from time to time some minors bugs like a thick black line across the middle of the screen or displays a "No reception possible" message in an area close to a silent zone (dark gray) but where the propagation is still open (signal stronger that -12 dB below which it is considered as lost by the program). These small "bugs" that do not prevent you to use the program should be corrected in a next release.

Propagation Chart

The “Propagation Chart” window displayed below pops up in fact each time that you request a Propagation Path Estimation in clicking a location on the Gray line or the Propagation map.

The Propagation Chart displays the signal strength in dB between the location A (usually your home) and any location B when you click somewhere on the Grayline or the Propagation map (in this case in the area YI-HZ). The wired black rectangle at right highlights the current time (UTC)to the user (location A).

Preset to use the current conditions data, the “Location A” is the location that you set in the Preferences (your home usually), while the “Location B” in any station pointed on the map or coordinates entered manually.

This is in this window that you can simulate other times and solar flux entering their value manually.

Following whas has been told about the accuracy of the Propagation map, using the same algorithms, this chart displays obviously the same limitations as it displays the same forecast, but no more at the global earth scale (area coverage) but between two stations (point to point); the propagation to some DX looks for example open with a strong signal over 15 dB and a MUF well above your working frequency but in the field, at the radio, the ears tensed or the mike warming, you cannot either work any station nor hear the beacon located in that country; another factor not considered affected the propagation.

Hopefully, at other times the forecast matches the real conditions but I cannot say exactly yet in which conditions values reverse the prediction from reliable to not reliable, all the less that the reliability is a term unknown from DXTB.

Of course the frequency used and the solar flux level are the main concerned factors but it is hard to say what is the triggering level : probably when the sun enters in a quiet cycle (SSN < 100), at low frequencies (14 MHz and below) where forecast becomes less accurate, and also at distance longer than 6000 km, triggers to confirm in the field or if we know what algorithms are used.

The bar-graph shows also sometimes one or more missing bars : this is not a bug but a drop in the propagation (see below). This chart using a 1-hour increment, the decrease and increase of the signal strength fall between two steps and thus they cannot be displayed to smooth the "curve".

Neither the propagation map or the chart takes into account the ground properties (conductivity and dielectric constant) or the change in propagation along the gray line. Yet, these factors affect the radiation pattern and the maximum range of your antenna. A simple example : if you can easily reach VK for K as most of the circuit is over the sea, it is much harder to reach JA from ON through Russia. Idem with the gray line that helps much for DXing.

In my last mail to the publisher (August 2004) I suggested him to use a smaller incremental step, trying to use the ground properties in forecasts, to add more series in this chart, like for example the signal strength as seen from the receiver at night, the signal-to-noise ratio and reliability, the best usable frequency or other custom series. At last the format could be customized too, allowing the user to select the colors and the chart type (bar, line, area, etc). An alternative is using a color mapping as he does for the MUF/LUF estimation (see below). Of course some users will prefer bars while other will prefer the rainbow, but the choice should exist.

Locations Window

The "Locations" window.

This window is only available through the Propagation or MUF/LUF Edtimation Chart. Logically it should be available from the main menu ("Windows") as well, what I suggested to the publisher.

By default DXTB provides no location contrary to other applications that list several thousands cities worldwide.

If you click on the “Edit Locations” button, a new window pops-up in which you can add, edit or delete locations. The program is limited to a maximum of 100 locations without accentuation. Don’t forget the minus (-) sign when encoding southern latitude and western longitude. Type also the coordinates as decimal integer with the decimal point (.) even if the "Preferences" display a coma (,). If you enter a coma the program reacts stupidly, display a black vertical line in the middle of maps and shuts down (bug). If you type well a dot, you’ll need to close and re-open this window for the changes to take effect.

The three columns can be sorted but it always didn't work in the release 2.2.0 under Windows XP or ME. New cities added were not saved each time neither, and I had to retype them several times for the change to take effect.

To add conviviality, I have suggested to the publisher to correct these bugs, to insert a "refresh button", to permit the user to encode angular values on request as he does for the QTH locator (grid), and to provide in a next release names and coordinates of main international cities.

MUF / LUF Estimation

The “MUF / LUF Estimation” window allows you to estimate the signal strength for a specified path for a range of frequencies over which propagation is expected or impossible. Recall that the MUF is a statistical value representing the median value of the predicted propagation estimation[1]. But this prediction alone does not tell all the story.

The “MUF / LUF Estimation” window for a path from ON to TA with 100 W PEP. The propagation looks open with signals at night 40 dB higher than at daytime where low ionopheric layers absorb signals. Such signals are common at short distances.

MUF and LUF strongly depend on the solar flux, the transmitter power, the time of the day and of the year as well of the location of the transmitter. 

You cannot simply trust in this forecast to plan your QSO because any geomagnetic storm will also affect the propagation of your signal, and thus the MUF, as well as the S/N ratio associated to the bandwidth of your operating mode while the D-layer status will affect the LUF. It is thus a chart rather complex to interpret without more information, all the more that it displays simultaneously three variables as a time function : LUF, MUF and the associated signal strength estimation in the scale at left expressed in dB or its equivalent in the rainbow meter.

If you understand well what represent LUF and MUF, this chart is then very useful and it is also user-friendly. The color mapping based on the rainbow colors ranges from purple for the weakest signal to red for strongest signal, the green being in the average. This color mapping is really efficient and speaks by itself, as good as the Propagation Chart bar-graph expressed in dB.

More interesting, you can bring this window up by holding down the “Shift” key while clicking on a location on either the Grayline or the Propagation Map window.

I think that the capabilities of this color mapping associated to the signal strength could be pushed much further to provide more dynamic forecasts displaying for example the SNR, S/I ratio, reliability, etc, if these parameters could be taken into account by more accurate algorithms.

Grid and Grid Map

At last the Grid calculator lets you determine the grid square (QTH locator) from the longitude and latitude. You enter them as decimal integer values or using the degrees, minutes, and seconds. The given value is correct.

The “Grid Map” windows shows a map of the world showing grid square and time zones at large scale. You cannot browse it except to enlarge it in full screen and moving inside with the lifts. If you move very slowly the mouse over your QTH locator you can find it but its resolution is perfectible as the way to move on the map.

Accuracy of forecasts

In buying DXTB, the amateur expects to get accurate forecasts. But he also knows that buying this program at the price of a good book about propagation these predictions could be as accurate as astrologers' ones. On the other side, a more complete model like VOACAP is free. So what to think about the accuracy of DXTB against the one of its competitors ? What reliability or degree of confidence can we grand to such a program ? 

Except the common inputs, SSN, geomagnetic indices, the frequency, the date, time of the day, and the power, you cannot ask the program to take into account an additional variable. DXTB doesn't take into account either parameters of a complete circuit (ground, transmitter, antenna, tolerance, S/N ratio, reliability, etc) nor the operating mode or the hop structures. As all these inputs are bypassed, it only works with the ouput power, a predefined receiver sensitivity and raw assumptions about the propagation mode or the signal strength injected in simple ionospheric models like Fricker's MUF and F2-peak models. But without considering many data, the results can only be biased, and therefore DXTB cannot thus gives you an accurate prediction for a specific path but only "the big picture" with some percentage of confidence.

DXTB predictions compared to the ones of the "golden standard", VOACAP. Above propagation conditions to Europe as seen from Australia (Canberra) on 20 meters for September 2004 at 0700 UTC (SSN = 27, SFI = 85). DXTB predicts no reception possible in Europe and Canberra is in the silent zone. It is unable to predict the MUF/LUF and forecasts a S/N below its threshold of -12 dB (both windows displayed at right are gray without the least bar). At right, VOACAP set for the same date and SSN, for a Yagi transmitter side with 8 dBi gain and 100 W output, a S/N reliability of 73 dB (SNR) and a required reliability (SNRxx) of 90%, thus good signals and conservative values. It predicts some opportunities between 06-09 UTC around 14 MHz with a signal strength at receive between -145 and -150 dBW or S2 and a S/N of 25 dB, thus weak. In the field VK/ZL stations arrived on my 2l dipole between S1 and S4 with a weak to strong audio, values that matched VOACAP predictions but not at all the ones of DXTB. Below, the same imprecisions occur when one tries to predict propagation conditions on evening.

Above a forecast calculated for Belgium on August 5, 2004 at 2200 UTC on 20 meters and 100 W output. The audio reports match partly the prediction from DXTB : signals from FY stations matched. LA or UA stations, so-called unreadable, were as strong as FY. K and XE were still weak but could be worked. The MUF over New York is predicted close to 9 MHz where VOACAP predicts it higher, close to 11.5 MHz at that time (not shown). At right calculated for the same time VOACAP predicts for UA (Moscow) a signal strength of -140 dBW or S3 and a S/N ratio of 27 dB; UA will be weak but could be worked contrarily to what states DXTB. Differences between DXTB and VOACAP come from the fact that the first does not consider either the path, ground properties, antenna gain or the reliability, but uses only median values, good for an overview but far to be enough to give an accurate forecast for a point-to-point circuit.

DXTB is for example unable to predict a drop in the propagation (strong QSB for tens of minutes) or a gradual increase of the signal due to a change in the LUF at daytime. It predicts well a global change in the propagation at earth scale (closing down or opening with an increment of 1 hour only) but it doesn't see these small variations that a program like VOACAP for example can predict[2].

As DXTB does not let you select the parameter to display (there are none except the propagation estimation per band, MUF/LUF associated to a power strength meter, and a point-to-point signal strength meter), you have to trust in the dB scale, the color mapping or the signal strength meter displayed in charts. Impossible to know the parameter reliability or whether the value displayed is a median (like the MUF I hope), a low or a high probability (e.g. a high or a lower decile, etc).

This application provides no access to curves like FOT, BUF and HPF (we can estimate them of course), there are no cross-sections of the ionosphere to get an accurate view of the height of the ionospheric layer(s) used in the circuit, no iso-contour or global maps showing critical frequencies (MUF, LUF, F2, etc), not even a S/N ratio or a power signal decile chart.

In fact you have no mean to know whether the forecast is right or false or to discover slight variations in propagation. In other words your unique way to know the accuracy of the prediction is to switch on your radio to check if the signal of the DX station that you hear or are going to work fits well in the estimation calculated by the program.

This inaccuracy appears quite rapidly when you request a prediction. In the Propagation map, DXTB refuses for example to consider the possibility to work at 100 W PEP a station located at short distance, say 1000 or 2000 km away (1600-3200 miles), at night on the 20 or 17-m band in summertime. Even without using a computer but only understanding the propagation fundamentals, there are of course several objective reasons to deny this QSO (season, D and E layers vanished, skip distance, low solar flux, in other words a band almost dead or experiencing strong QRN and QSB). But in the field you will observe quite often that the band is not as closed as estimated and even on clusters amateurs spotted some contacts. The same inaccuracy appears with DX stations (located over 6000 km away) that are estimated too weak to be heard ("Reception not possible") and who arrive quite strong and that can be work using a beam.

But usually silent zones represent well what they are : an area in which there is no propagation at the specified time, frequency and power. Hereunder is an example showing the propagation chart from ON to TZ (Mali). The "missing signals" between 0300-0500 UTC are real, correlated with the silent zone associated to the darkness on this part of the earth at that time.

Propagation chart calculated for August 21, 2004 at 2200 UTC (left) and 0400 UTC (right). I ran the simulation earlier to understand the reason of the void (no signal or lower than -12 dB) in the propagation chart between 0300-0500 UTC. This is not a bug but a real effect related to the silent zone associated to the darkness over TZ in these early morning hours.

Well, and at the end, if a forecast states that the band is closed, is it really closed or not ? In fact this problem concerns as much the operator's working conditions (his/her antenna system and the mode used) as the propagation : even if the propagation is so-called closed, a beam could still reach a DX station where the omnidirectional vertical gave up for a long time. If you add some power, that will be still easier, and still more if you work in CW. However, even if DXTB takes into account the power, nobody tells it either that you use a high gain beam placed 12m high or an isotropic antenna at low height for example, or that you work in CW. These factors affect thus the prediction of a serious bias.

This dependency on the working conditions is obvious checking beacons that transmit without interruption in CW. In the Propagation map, DXTB displays more than once "Not reception possible" at daytime with a signal strength below -12 dB (thus "below" the first S-unit) although the beacon is very strong at 100 W (ZL6B reached S-7 and ZS6DN was S-5 during hours); at other occasions it forecasts a strong signal (+18 dB) at daytime although the beacon located a short distance was unreadable (e.g. 4X6TU or RR90). In most of these cases, but not always, at the specified time the MUF, LUF and signal strength are simply not displayed because the program considers the signal too weak (what usually means that the darkness is not far but the contrary is also true, or that the MUF is lower than the working frequency or, at daytime, that the D-layer is still too active, but many other interpretations are possible). 

MUF/LUF estimation, propagation and signal strength forecasted for ZS6DN beacon at 1127 UTC (left): no signal to expect in ON. At the earth scale, in the propagation chart displayed in the below right corner it was obviously confirmed : there was no propagation (dark gray) to South Africa. However, at the radio the 100 W beacon arrived S-5 on a simple dipole 40m long ! By 1800 UTC (right) the forecast predicted a much stronger signal exceeding 10 dB and the propagation chart below right confirmed a strong signal (+12 dB) up to South Africa. That matched: this time it arrived S-3 with a strong audio.

Most of the time, as you have no other chart at your disposal to cross-check these values (especially the median S/N, SNR, receive antenna properties, signal power lower decile, etc) you stand in ignorance of the real reason that led to these lost of propagation or sudden openings to such or seach country. This is not really the objective of a propagation program that should highlight the true reason of these poor conditions (showing a signal strength in the purple, a very low power at receive, low reliability, or conversely a strong signal, high SNR, low losses, etc).

The problem is emphasized by the fact that you must also interpret graphs and figures, although a map or a chart  is by definition self-explanatory. But even if a green color means "good signal" or if a signal strenght of 10 dB looks strong, 10 times stronger than a 0 dB signal, it does not mean much as long as you don't hear it in the reality. Add also the fact that this interpretation acquires also with experience in working on the air, still the best way to try correlating what you hear with what state, with more or less accuracy, maps and figures. In all cases to explain a changing propagation, display always both Propagation and Grayline maps together, because if the information is not self-explanatory on the first map or in its chart, the second one will give you some clues. Another trick is to run the Propagation map some hours earlier or later to see how it changes, knowing however that such a program is not really suited to work with short-terms but rather with longer lapses, of one or more days.

Trying to set a prediction for a long path to VK/ZL, at the limits of the map (52°S, 176°W) DXTB traces a surprising path.

Knowing all this, we can say that where other propagation programs lack of parametric maps to cross-check predictions, DXTB lacks of additional charts to validate forecasts. But we have to put all these problems into context because, as we told, other more sophisticated programs display also some degrees of imprecision. Reasons are known and multiples : all these programs use algorithms more or less accurate, with or without interpolation, or use functions extracted from the same ionospheric model and set their probability or their degree of reliability to median values.

At last, we explained earlier that DXTB shows also some errors at the limits of maps. Stressing the program, I tried to display the long path to New-Zealand and Australia, and the concerned propagation chart. Impossible, the function is not supported. Annoying when you want to work VK/ZL using the long path from Europe (say between 0400-0800 UTC) to benefit of the darkness and a stronger signal...

If you select an island close to ZL and the left frame, at some coordinate points located near 52°S and 176°W, very close to the frame, the path is erratic, it begins to shows steps instead of a smooth curve. The function has been improved and does no more shows straight lines and steep angles instead of a curve like do other applications (remember that we work with geodesics). But the reason of this error is obvious : many programs doesn't take very well into account conditions at the limits, especially the high angular values at high latitudes due to the cartesian projection. I suggested to the publisher to still improve the display accuracy and  to add a long path calculation.

My final impression

A lot of features are available in “DX ToolBox” at a few keystrokes. Not only it is cheap and good-looking but it provides a very intuitive graphic interface with maps, charts, images and reports that will satisfy all casual amateurs and even amateur astronomers who are looking for near-real-time solar and geomagnetic bulletins.

If you want to understand the behaviour of the ionosphere and figures listed in bulletins displayed on-screen, its “literature” is made for you, I mean all its warning messages and other status available in its "Reports" window. In addition tens of images, including your own links, will illustrate its most verbose text. And all this is accessible simply in clicking in one submenu. Time consumed : a few seconds once all images and reports and downloaded. Even forecasts display quickly. That cannot be simpler and faster !

This simplicity has unfortunately a drawback. The engine hidden behind this program, algorithms or functions, are not as numerous, flexible and powerful as the ones included in a coverage analysis program like the VOACAP model for example. Its accuracy is thus limited. But as we told, as well as in other pages dealing with the ionospheric modeling, many other programs in this category show restrictions too, and either VOACAP or DXTB doesn't go against the rule.

So, without to be either powerful nor a complete tool, DXTB covers most requirements of a casual amateur and will probably please to the advanced DX addict amateur looking for a simple propagation program. However; much cheaper than more sophisticated programs using the VOACAP engine, do not expect the same accuracy and the same functionalities as its competitors.

But positive side, if you don’t understand anything at all in the propagation, I make the bet that thanks to its online reports and dynamic windows this program will help you to understand better this complex subject, and that you will do a big …hop ahead, Hi !

Download, purchase and support

“DX ToolBox” is now at version 4.6.3. It is available for all Windows and Mac platforms (incl. OS/X, iPhone, and iPad). It requires an active Internet connection for get online updates.

“DX ToolBox” is free to try out and can be downloaded as an installer or a compressed Zip file from the publisher website, a file that "explodes" in about 13.1 MB of data distributed in five files. Some of its competitors are over three times bigger without offering more accuracy.

In this “free to try” version it displays at regular interval a warning message. If you decide to continue using it, you can buy a copy for just $24.99, which gives you a registration code to remove the reminder messages (code to enter in "Edit" menu, "Enter Registration Code").

“DX ToolBox” runs on a single computer. If you wish to run it on multiple computers simultaneously, you must obtain a license for each system, or the appropriate site license. Please contact Black Cat Systems for pricing and details about site licensing.

Last but not least, when you purchase you'll be entitled to use all new releases and updates to “DX ToolBox” released over the next year, free of charge.

Close this review saying that the response of Chris Smolinski to users enquiries is fast at first contact and comes back with tips and a fix less than 24 hours later. This is very appreciated when you are face to problems. Congratulations!

I hope that you will appreciate this product.

Last note from the publisher, Chris Smolinski

July 26, 2004 :  "Thanks for the suggestions, I'll consider them. I also am in the process of writing a brief tutorial on propagation, what the various indices mean, etc". This tutorial is already partly available on the publisher website (see below).

For more information

If you are seriously interested in propagation, I warmly suggest you to read some books and studies that you can find in the ARRL or RSGB bookshop. Completed with simulations, you will become without any doubt a guru in this matter, Hi !

DX ToolBox propagation software, Black Cat Systems

HF Shortwave Radio Propagation Information, Black Cat Systems

HF Propagation tutorial, NM7M (on this web)

Propagation Studies, RSGB

ARRL Bookshop

RSGB shop

Back to Menu

Page 1 - 2 - 3 -

[1] MUF is between FOT, the Frequency of Optimum Transmission and HPF, the Highest Possible Frequency. For example if FOT = 10.7 MHz, HPF = 17.4 MHz then MUF = 13.6 MHz. Statistically speaking HPF and FOT values lie about 15% above and below the MUF values. MUF represents thus the monthly value during which the frequency exceeds 50% of the time, or 15 days per month. The upper and lower decile values of critical frequency have to do with the 90% and 10% limits.

[2] Here is a chart showing QSB around 0900, 1300, 1600 and 2000 UTC on 7 MHz predicted by the VOACAP Power Signal Lower Decile function, a statistical function like the MUF but instead of giving a median value of 50% for the path gives a frequency which is exceeded 90% of the time, or 27 days a month. Power variations along the day are clearly visible.

Back to:


Copyright & FAQ