Much confusion exists about the radiation patterns of antennas. In particular, the horizontal dipole is misunderstood by most who use it. It is generally assumed that a horizontal dipole has a radiation pattern which aims most of its signal toward the horizon, perpendicular to the line of the antenna. As we will see here, that is only true in one particular circumstance, and, even then, not quite an accurate picture.
Following is a series of dipole radiation pattern plots, produced by a popular antenna modeling program. The first plot below shows a dipole at an elevation of two wavelengths above a standard earth ground model.

Note the multiple lobes at varying elevation angles above the horizontal. These lobes represent a division of the radiated energy, so that much energy is sent off at angles which may not be helpful to the particular communication in progress. In addition, the nulls represent angles at which no energy is transmitted. Depending on the height of the ionosphere, and the location of the distant target station, the null may exist at the precise angle necessary to effect the desired contact.

Page 2 of 2
The next plot shows the dipole at 1.75 times the wavelength above the earth.

There are fewer lobes and fewer nulls, but they are still present. In addition, two of the upper lobes have moved up and merged to produce a single, wide lobe aimed straight up.

Next is the dipole at 1.5 wavelengths elevation.

Page 3 of 3
Again, fewer nulls and lobes, but still not necessarily concentrating the energy in the desired direction. Next is the dipole at 1.25 wavelengths high.

Page 4 of 4
Next is the dipole at .75 wavelength.

And, the dipole at .5 wavelength over the ground below.

Page 5 of 5
Finally we see a pattern that looks somewhat like what we expected. But, notice that, rather than the energy being directed at the horizon, the peak of the pattern is elevated at about a 30 degree angle (actually 28 degrees). However, the lobe is broad enough so that there is considerable signal strength from about ten degrees to about 55 degrees elevation - the 6 dB down points on the lobe. Remember that 6 dB represents one "S" unit on the signal level meter of most HF receivers. Thus, for a quite broad range of elevation angles, and for a correspondingly broad expanse of geographical area covered by the signal reflected off the ionosphere, there is substantial signal strength.

As we continue to lower the dipole, we continue to see the now familiar raising and combining of the lobes, until, at .25 wavelength shown below,

the lobes have combined to produce a single broad peak, aimed straight up. Clearly, this is not going to make for great DX communications, with most of the energy concentrated above the 30 degree elevation angle. But, notice that there is still considerable energy directed even as low as 15 degrees, the 6 db down point.

This is the radiation pattern which is normally used for NVIS operations. NVIS stands for Near Vertical Incidence Skywave. Below a certain critical frequency (which depends on the density of the ionization in the ionosphere), a signal directed straight upward will be reflected back down into an area near the transmitter. This can allow communications within a 100-200 mile radius of the transmitter, the area normally skipped over by signals transmitted from the dipole at lower radiation angles (generated by higher positions above ground). NVIS antennas are normally elevated at about .1 to .25 wavelength above the ground.

The next plot shows the dipole at .1 wavelength.

Page 6 of 6
At .1 wavelength, we see the single lobe even more tightly pointed up, with the 6 dB points now at about 25 degrees. This appears to be a very good NVIS antenna. But, as we will shortly see, there is another factor involved, the antenna's impedance.

Before we leave the NVIS discussion, lets go back and look at the .75 wavelength elevation pattern shown below:

Page 7 of 7
Notice that this pattern has both the low-angle lobes and a broad vertical lobe. The vertical lobe is about 2 dB below the low-angle lobes, an almost insignificant difference in signal strength.
It may be that a dipole at .75 wavelength height over ground is the best of all possible setups, providing both DX communications and NVIS communications at the same time.

For each calculation, the antenna's length was adjusted for best SWR, using 50 Ohms as the reference impedance. The calculated dipole was modeled using #12 wire for the elements.

Page 8 of 8
The radiation pattern for a dipole set up according to the 16.7% / 47.7% chart looks like this:

This is not the ideal DX antenna, but it does provide good signal strength at reasonably low angles (6 dB down at 20 degrees elevation angle). It is essentially the same pattern as the .25 wavelength height dipole, but with the benefit of being matched to 50 Ohms. In addition, it provides a very broad NVIS signal lobe. This looks like a good compromise between both modes, and is fairly easy to set up on frequencies above 10 MHz. At lower frequencies, the height above ground becomes a problem to achieve with a reasonable support structure.

73, John, WB4YJT