HF Propagation tutorial

by Bob Brown, NM7M, Ph.D. from U.C.Berkeley

Propagation modes and DXing (VI)

Having spent some time with the ionosphere, now we have to be more practical, speaking of propagation modes and the things that can go wrong when DXing. But modes are the first order of business. In that regard, everyone knows about HF hops from the various regions - in the range of 1,500-1,750 km from the E-region and about 3,000-3,500 km from the F-region. Of course, it depends on frequency and the radiation angle at which signals are launched.

The electron distribution, having greater density at the higher altitudes, always refracts signals downward. That may seem a bit strange but that is the case; rays which are ascending are bent back toward the earth and the same is true of rays which are going down. The rate of bending is greater at the higher altitudes, when rays are close to the greatest concentration of electrons, but it is always AWAY from the region of higher ionization. And as I indicated earlier, how far rays proceed in the ionosphere depends on the effective vertical frequency (EVF) when they were launched, just like the baseball. Remember?

Let's take the case of some rays where the EVF is very close to the critical frequency at the peak of the F-layer. In the figure at right, Ray A is one where the EVF is less that foF2 and it is bent back toward ground while Ray B is one where the EVF is greater than foF2 and it penetrates the F-peak and goes on to Infinity.

But notice that both rays A and B are bent or refracted AWAY from the region where the ionization is the greatest, the F-layer peak. That's a general feature of refraction in the upper range of the HF spectrum. Now one other thing; it seems rays can be reversed in electromagnetic theory so Ray B could be the path for galactic radio noise which penetrates the F-region below. OK?

Now we come to Ray C, one where the EVF is very, very close to the critical frequency of the F-layer. That type of ray, moving almost parallel to the earth's surface is called a Pedersen Ray. Those rays can give very long hops but they are essentially unstable in the sense that any little increase or decrease in the electron density and they diverge, going back to ground like Ray A or off through the F-peak to Infinity like Ray B.

Just in case you missed the idea, Pedersen Rays at the peak of the F-region involve the upper portion of the HF spectrum as the oblique path must reach those altitudes; that is not possible for the bottom of the HF spectrum (3 MHz) as even vertical rays can't penetrate that far up in the ionosphere as foF2 is just too high.

But that is not to say that Pedersen Rays are impossible at the bottom of the HF spectrum; it's just that type of refraction takes place down around the E-region where the electron density levels off for a short range of altitude. So let's look at some ray paths there, for 80 and 160 meter signals with EVF close to the value of foE, especially at night as shown at left.

Ray path A corresponds to a E-hop where EVF < foE and covers only a short distance to a receiver. But Ray B is one where the signal has an EVF that's very, very close to foE. But it penetrates the E-layer and ascends into the F-region; however, its EVF is still too low to reach the higher portions of the F-region and so it is refracted back down.

If the down-going angle of the ray has not been affected, it will continue for a distance along the level of the E-region and then be returned to ground. In a sense, the path resembles that followed by a Pedersen Ray but there is that short excursion into the F-region making it an E-F path.

Whether at the level of the E-region or the F-peak, paths which have Pedersen-like refraction cover greater distances than the simple E-or F-hops. As such, they would contribute to paths with few hops and stronger signals; however, as noted earlier, they may be unstable and only have brief existences. With the varied paths that amateurs use, such situations are not readily identified; however, for fixed paths in commercial use, it is a different story. In that regard, it is pointed out in Davies' book that HF Pedersen rays tend occur around local noon on fixed paths across the North Atlantic, when the density gradients along the path are at a minimum.

So the above examples cover the simple, single hops that can occur, from short E-hops to long E-F hops, then F-hops and even long Pedersen hops. After that, we get into multiple hops; those are more complicated, of course, but there is some simplicity in the second and third hops in that reflections involve equal angles of incidence and reflection from a surface. But even then, there is the odd chance of complexity if the surface is not flat or not smooth. The former would, in effect, change the next launching angle of a ray, adding or subtracting the tilt of the surface to its original angle relative to the horizontal direction.

As for rough surfaces, they can give a diffuse reflection and that serves to reduce the power carried forward in the original direction. At surface reflections, there can be some signal loss, depending on the signal polarization, surface material and the frequency.

As you know, we distinguish between horizontally and vertically polarized waves, meaning the electric field of the wave is either parallel to the earth's surface or perpendicular to it, as for radiation from a horizontal dipole or a vertical antenna.

While there may be signal loss (in dB) on reflection, the process is discussed first in terms of reflection coefficients, meaning the amplitude of the reflected wave compared to the incident wave. The graphic at right illustrates the case for good ground material and 14 MHz signals; clearly, the small reflection coefficient for vertical polarization around 25 degrees means there would be a large signal loss for waves incident at that radiation angle. But horizontal polarization is much better in that regard and is the reason why most DXers prefer horizontally polarized antennas.

Of course, once signals leave an antenna, their progress is part of the discussion of propagation. Everyone knows that salt water is the best reflecting surface for RF and fortunately 78% of the earth is covered by oceans. That really helps DXing. But a significant fraction of ground (and amateur population) lies in the northern hemisphere and the rest of the earth involves ice and snow in the polar caps so the distribution of surface material shown below is of some interest to the propagation of signals.

We'll do more with reflection loss later on but for the moment, it is important to know it is there and extracts signal strength with every bounce. But there is one more point to bear in mind; the angle of reflection can be as important as the polarization, the surface or frequency. Thus, losses off of water at low angles are about 1 dB, about 3 dB off of the various forms of ground and in excess of 6 dB off of snow/ice. The situation gets progressively worse at higher radiation angles so low radiation angles should be the order of the day. But you knew that, just because the hops are longer at low angles.

 The reflection power of the earth varies from excellent to extremely poor depending on whether you are over salt water (dielectric constant k=81, conductivity G=5 S/m), pastoral with medium hills and forestation (k=13, G= 0.006 S/m) or cities with building and heavy industries (k=3, G= 0.001 S/m). A poor ground affects also performances of vertically polarized antennas due to the Pseudo-Brewster Angle (PBA), a situation similar to the one we experiment when the sun is low and its light reflects from the water's surface as glare, obscuring the underwater view. The reflection coefficient can exceed 90% at 15° of elevation (21 MHz).

Finally, it should be noted that we've pretty well assumed the ionosphere to be concentric with the spherical earth. That is a simplification, of course, and we have to expect tilts in the ionosphere and those will have effects on waves returned from the higher altitudes. For one thing, a tilt ALONG the path will change the angle of return to the ground; for another, a tilt ACROSS the direction of a path will affect the polarization in the sense that what was a horizontally polarized wave may now have a vertical component to it. So the next ground reflection becomes a bit more complicated, the signal loss now depends on how the two polarizations are reflected. And then there are phase changes on reflection. But nobody said radio was simple, did they?

Let's go on with multiple hops, putting in more of the details. One matter of interest is the radiation angle throughout a path. Thus, one might pick one angle, say at the peak of the antenna radiation pattern, and try to follow it along a path. But while the Laws of Optics apply, with angles equal for incidence and reflection from a surface, the angle may change due to a tilt of the ionosphere on one hop or change of inclination or slope of ground at a reflection point.

So there could be some variability in the radiation angle. And, of course, the height of the ionosphere is not constant along a path, changing if the path goes from being in sunlight to being in darkness. All those aspects of the path serve to change the distance per hop or, for that matter, how close the path for a given radiation angle comes to the target QTH.

Leaving aside the variations which result from surface reflections and the like, one can illustrate path structures by making various combinations of hops. Without citing any particular type of the ionospheric circumstances, some common paths are shown below.

 At left a propagation path close to the gray line involving together an E- and an F-hop. At right an F-to-F propagation involving a plasma cloud known as an E-sporadic. Both phenomena contribute to long-path propagation.

Of course various other combinations are possible. The modes shown above are specified as as E-F and F-Es-F. For longer paths, the number of E- and F-hops may be larger, depending on how the path is located relative to the terminator. As for desirability, the rule is that E-hops on a path are where most losses occur, with ionospheric absorption on the sunlit legs and ground losses, while F-hops in darkness have less loss, with fewer ground reflections for a given distance from point A to B.

The presence of a sporadic E reflection, without any intermediate ground reflection between reflections from the F-layer, brings up another type of path that contributes to long-path propagation.

Here, the idea is the same as with the Es reflection except that the ground reflection is missing because of ionospheric tilts, shown at left in cyan, between the two portions of the F-region.

This figure is "Flat Earth Physics" but in reality, the ray reflected off the first part of the F-region did bend downward but it didn't go down far and the curved earth fell away from it so it missed the earth and went on to the F-region again. OK?

While the tilts shown below are exaggerated, such circumstances are found regularly on paths going across the geomagnetic equator in the afternoon/evening hours and give rise to long, chordal hops with correspondingly stronger signals. But it should be noted that "tilts" really are another way of representing the changes in the electron density distribution along a path. Thus, an upward tilt, one that gives a longer hop, really is the same as the case where the electron density DECREASES along a path direction and results in less downward refraction. That is called a negative gradient and, of course, a positive gradient is just the opposite.

Finally, there is another interesting variation on path structure that results from a negative gradient along a path, ducting as displayed above right. In that case, the situation is like the E-F hop discussed previously but the excursions into the F-region are repeated several times.

Again, the representation shown at right is "Flat Earth Physics" and involves a negative gradient, just like the chordal hop mentioned earlier. But those long hops are more characteristic of the upper end of the HF spectrum, 14 MHz and above, and require almost the full height of the ionosphere for their completion. That is the case as even a reduction in electron density along a path does not reduce refraction at the higher frequencies to a great extent.

The ducting shown above right is for the low end of the HF spectrum and involves smaller vertical excursions of ray paths than the case for chordal hops. That is the case as refraction varies with the inverse-square of the frequency; thus, for the same gradient or reduction in electron density along the path, the change in the downward refraction is much greater at the low end of the HF spectrum and less of the ionosphere is required for the same type of effects.

Now, having gone through a wide range of mode structures that are possible, one can use those ideas in dealing with propagation. But, face it, the RF from one's antenna pattern goes off into all the possible modes, be they E-, E-F or F-hops and, depending on the operating frequency, some of the exotic modes, like chordal hops or chordal ducting are possible too. But the mode that gets through for your DX contact is something of a "survivor", giving signals where the others have died out due to absorption or have the wrong radiation angles for the path or receiving antenna.

But at this point, about all we're prepared to think about are the more common modes and those would be in relatively calm, stable conditions. In short, we'd be looking at the indicators, SSN and the like, perhaps a map with great-circle paths on it and pointed our beams in the right directions. But the "when, why and how" have yet to be discussed, to say nothing of circumstances that are out of the ordinary.

Myself, I consider "when, why and how" to be the "propagation imperatives", the ideas that every DXer should have in mind before turning on the rig in pursuit of a "New One". In short, those ideas should be "Second Nature", the sort of thing you'd have in mind if shipwrecked on a desert island with nothing but the makings of a ham station at your disposal. You should be able to think of the DX QTH, have a feeling for what could be done on a given date and think of when to get on the band of your choice. Sometimes the answers are not to one's liking but an answer should be forthcoming without too much head-scratching.

So let's see what we can do to get that right, at least for normal conditions, and then deal with disturbances and see what they'd mean for us. That won't be too burdensome as once the broad outlines are established, you'll have a propagation program to fill in the quantitative details, case by case.

Now we've discussed some of the general ideas behind propagation in the HF part of the spectrum and you should have a good grasp of what it all depends on - enough ionization overhead to refract signals downward, keeping them in the F-region, and signals getting through the ionization down in the D-region with enough strength to overcome the local noise.

Case study

With that in mind, let's explore propagation with a practical case, say making a contact between a central location in the USA and Togo, West Africa in the upcoming CQ WW CW contest in late November 1998. That'd be a good test to see just how far we can go in predicting propagation using the simple ideas developed so far. That done, we can look at how computer programs do it and see what other details they offer.

So let's use Omaha, NE as our QTH in the USA; that's at 41° N, 96° W. Togo is a bit harder so we have to go to the ARRL Operating Manual or DXAtlas to find that it's located in the Horn of Africa, at 6° N, 1° E, close to the Greenwich Meridian. Looking at those coordinates, one thing is immediately clear - it's quite a ways from Omaha to Togo, more than 90° difference in longitude and more than 35° difference in latitude.

 To reach Togo (5V) by shortwaves from Omaha (NE., USA) you have two solutions : use the short path (10,100 km bearing 81°, 5 hops) or the long path (29,900 km, bearing 261°, about 10 hops). The direct and short path to this central african country seems "workable" and accessible as the 3/4th of the path runs over the ocean that offers a high reflectivity to shortwaves. The long path is also interesting as it runs on 5/6th of the path over the dark side of the earth. But first of all is there an opening to Togo on 28 MHz by 1600 UTC on this November 21, 1998 ? The solar and geomagnetic conditions as well as the MUF and the F2 critical frequency maps will help us to answer to this question. In the negative we will have to work this entity at another time and another band, maybe much lower. The answer is given you below. map created with HFProp.

Considering that the distance around the earth is about 40,000 km, one can conclude immediately that the distance to Togo from Omaha is better than 10,000 km, a quarter the way around the world. That's confirmed by going to the azimuthal equidistant map for Central USA in the ARRL Operating Manual or any logging program showing the world map (e.g. DX4Win); Togo is half way to the antipodal circle, making it quite a haul. But it's not all that hard if you're on the right band at the right time.

Now we're talking about late November so we can take the effective sunspot number SSN as around 80, judging by recent reports from NOAA. The chances of making a contact on the higher bands are pretty good when you consider that Togo is at a low latitude, where the the electron distribution of the F-region is quite robust. So we only have to worry about launching the high band RF from Omaha.

As a first approximation, let's think of trying for a contact on 28 MHz. For that, ionization and the MUF are the important things and tell us that the contact should be tried during the time the path is well illuminated. So with a longitude difference of about 97 degrees, we'd like to have the sun at least midway between the two QTHs, say at about 47° W of longitude. With the sun advancing westward at 15° of longitude per hour, that means the time should be about 3 hours after 1200 UTC or 1500 UTC.

But remember Togo is at a low latitude so the critical frequency of the F-region there is less of a problem than at Omaha. That being the case, it would be better to choose a later hour, one when the sun is closer to the longitude of Omaha, raising the critical frequency near there. But the time should not be so late as to have the sun set anywhere on the path. That means we have to look into the sunrise/sunset tables in the ARRL Operating Manual or any astronomical calendar, paper or program, and see when the sun would set at Togo.

 At left, the MUF iso-contour map calculated with HFProp shows some "islands" up to 30 MHz between Equator and Kenya, at first sight good to reach Togo; the short path looks open. But at right, the F2-critical map shows many areas where the critical frequency is much lower (6-8 MHz). Other maps, especially the MUF, the propagation map, SNR and reliability charts will confirm that in fact all bands from 20 to 10m are open to Togo in this evening (1800-0100 UTC) with signal between S6 and S7. Unfortunately in the afternoon signals are not higher than S3.

In that regard, the Operating Manual gives SR/SS data for November 21 and we can use that as an approximation, taking the ground sunset at Togo as 1736 UTC. That would suggest, as a first correction, that the 28 MHz band be tried between 1530 UTC and 1730 UTC. The same would apply for 21 MHz too, knowing that less ionization is needed for propagation on that band, so an operating window might be better if widened to start earlier and end later, say from 1500 UTC to 1800 UTC.

As an aside, I should say that last idea has some generality to it, at least for the bands where MUF are important. So from a given QTH, the lowest bands open the earliest, the highest bands the latest, and band closing is in reverse order. Of course, that is just the availability of the path; the signal/noise situation still has to be looked at for the best times of operation.

As for the transition bands, 10 MHz to 18 MHz, absorption plays a role there and good sense indicates the effect can be minimized by avoiding times when the path is well illuminated, with the sun around its midpoint. In addition, we know that ionization lingers after sunset, thanks to the role of the geomagnetic field and the slow recombination rate of electrons and positive ions up there in the F-region. As a result, propagation on those bands would be supported around sunset and on into the evening hours.

In addition, the rising sun on the path near Omaha would open up propagation, at least until absorption became too great. That being the case, we can expect the bands to open shortly after the sunrise at Omaha, roughly 1350 UTC according to the Operating Manual. And with sunset around 1730 UTC at Togo, another two or three hours could be added to the operating time.

Things are shaping up, at least for the bands where F-region ionization and D-region absorption are important. That would give a starting point as sunrise at Omaha, about 1400 UTC, and a closing time of about 2030 UTC for the transition bands. The higher bands would start later, of course, and end sooner, the general principle mentioned earlier.

The lower bands, 160 meters - 40 meters, where D-region absorption dominates, would be open from sunset at Omaha til sunrise at Togo. Going to the Operating Manual, we find low-band operations could start at Omaha around 2300 UTC and end around 0545 UTC.

 At left the MUF calculated by DX Toolbox over Togo does not extend over 24 MHz around 1800 UTC and goes down to 10 MHz during the night. In fact from Omaha, at 1600 UTC the only opening on 10m is toward South America. To reach Togo in good conditions we must go down close to 20m and work late in the evening (2100-2300 UTC and even early in the morning on the low bands as shown at right). The signal are predicted weak, thus to work preferably in CW or in SSB with 1 kW in a directional array. You lost your time if you try to work 5V at another time. Other charts will confirm these conditions.

But there is the question of noise, man-made or atmospheric in origin, to compete with signals. Here, experience shows that man- made noise is less as the hour goes past the end of the working day. And atmospheric noise, say at Togo, would be the lowest at times close to dawn. So low-band operation probably would be more productive in the later hours of the operating window. But in view of the high level of ionospheric absorption and distance involved, it could be much more difficult to make a contact on the lower bands than the higher ones. In addition, antennas and power play a greater role in that part of the spectrum. Those resources are developed over time by DXers and related to their operating experience in that part of the amateur spectrum. Put another way, DXing on the lower bands, 80 and 160 meters, is tough and not always rewarding for casual operators.

Now, to add a realistic twist to this discussion, let me say that I worked 5V7A on 20 in CW at 2312 UTC on November 29, 1997. If you look into it, you will see that was over five hours AFTER ground level sunset at Togo! (See? Ionization does linger on in the dark, especially at low latitudes!) I would hope you could do the same this year. At least, the above example shows how you can "sharpshoot" for a New One, even with only primitive tools at one's disposal. Give it a try. OK?

 At left for a S/N ratio reliability (SNR) of 38 dB in CW and a required reliability of 90% at the specified date (Novembre 1997) and with a SSN of 40, VOACAP predicts a S/N ratio of 40 dB in Togo at the time of Bob's QSO with 5V7A at 2312 UTC. At right taking into account the date and URSI/88 Coefficients, ICEPAC predicts for the opposite circuit a signal power in Omaha of -117 dBW, or close to S7. A QSO can be sched at that time with good signals on both sides. Bob selected the best time; according forecasts, signals were the strongest on 20m between 2100-2300 UTC as predicted DX Toolbox as well. Note that both circuits (K-5V or 5V-K) are quasi reciprocal with very light differences in the signal strength, MUF and FOT.

Now I didn't work out all the aspects of contest propagation for the 5V7A group; you'll see what their own propagation guru came up with but I'm sure it was based on the principles I outlined above. I have done that sort of thing before, for the recent 8Q7AA and 3B7RF DXpeditions. In that sort of circumstance, the idea is to forecast so they can "Work the World". So every time interval has to be looked and in every direction to find the best way for them to operate in the contest.

The first one for the 8Q7AA group went very well, operations going essentially as predicted. But the second one for 3B7RF got into a bit of trouble; that was interesting in itself as it will lead us into the matter of ionospheric disturbances of geophysical origin. Leaving that to later, let's go beyond slow, mechanical methods, how "The Ancients" handled the propagation problem, and look at how it's done by computers.

As you know, they do everything practically at the speed of light. But how well do they do it? That's a good question. As a matter of fact, given what you know now, you might wonder if they just do the old-fashioned calculations faster and not add much to the problem. So we'll go with that for a while, looking at how computers handle these questions and then look at a few new ideas.

Reference Notes

If DX contesting is the sort of thing that interests you, let me say that the 5V7A crew were kind enough to provide me with their '96 and '97 contest logs for analysis. I was more interested in them for the aspects of 160 meter propagation but you might look at my article in the March/April '98 issue of The DX Magazine. It also shows how demographics overpowers propagation.

Next chapter

Page 1 - 2 - 3 - 4 - 5 - 6 - 7 - 8 -

 Back to: HOME Copyright & FAQ