I built two radial systems for this same 23' vertical and moved the antenna back and forth. One radial system is made up of 64 wires extending from an aluminum disk as shown in part 1 of this article. The other consists of four radials made of welded-wire fencing laid flat in an X pattern. Each is 25' long.

With a perfectly reflecting ground the resistance should be 36 ohms or one half of a what a dipole is. Using an MFJ-259 analyzer both radial systems yielded the same measurement, 36 ohms, at resonance, indicating a perfectly reflecting ground, at least within the limits of the analyzer's accuracy. Based on thexe measurements I conclude that it doesn't matter whether you use fencing or 64 wires. However, fencing is much easier to lay down.

At one time I had only 18 wire radials and the resistance measured 47 ohms, indicating a considerable ground loss. It was easy to match and I did work a lot of DX, even with the losses. How much was I losing and how much was getting out?

I calculated power lost as I-squared-R, where I is current and R is resistance. If we operate only at resonance we have only resistance to deal with and don't have to be concerned with reactance. Power is lost as heat, and the loss is heating up the soil slightly at the base of the antenna. With 100 watts, solving for I the nominal current at 36 ohms would be 1.67 Amps (Square root of 100/36).

At 47 ohms the same power is delivered from the transmitter assuming a matching network or tuner is virtually lossless and is presenting a 50 ohm load to the transmitter both times. Again solving for I, the current is now 1.46 Amps (Square root of 100/47). Before, there was only one resistor and that is the radiation resistance of 36 ohms. Now, there are two resistors, the 36 ohms of radiation resistance which doesn't change, and now the 11 ohms of ground loss. The difference is the portion of power wasted, probably as heat generated in the ground loss resistance. That loss can be calculated as I-squared-R in the 11 ohm resistor, or (1.46 * 1.46 * 11). The result is 23.4 Watts, or about 23 per cent of the 100 watts.

As the analyzer shows, it is possible to obtain a reading of 36 ohms. One doesn't always need the broadcast standard of 120 radials to achieve a perfect ground. A lot depends on soil conditions. The better soil conduction the fewer radials you need. If you don't your soil condition you can let the analyzer show when you have enough. Over poor soil one probably would need 100 and over salt water one needs very few, if any. At my QTH 64 radials appeared as a perfectly reflecting ground.

Top Hat
The vertical dipped at 10 MHz which is great for 30m but I was trying to build a 40 meter antenna. Making the antenna taller was not an option so I needed some sort of loading. I had read a top hat adds loading without adding loss like a coil.

I added six top hat radials by salvaging three elements from an old 2 meter beam. The diameter is 38 inches. The resonant frequency of the original 23' vertical measured 10.0 mHz. This addition lowered the resonant frequency, but only to a disappointng 9.0 mHz.

Next I tried adding a ring of aluminum wire using small hose clamps to hold the wire to the tip of the top hat elements. Now the resonant frequency measured 7.8 mHz and this is where I left it. Resistance at 7.8 mHz measures 36 ohms on the MFJ analyzer.

If I want the antenna resonant at 7.0 mHz my next move is going to have to be the introduction of some other kind of loading such as a coil which I would put about two thirds of the way up. Or I could figure out a larger top hat.

Meanwhile, I am temporarily coupling the transmission line to the antenna with a SGC-230 Smartuner. The Smartuner forgives all shortcomings. It adds inductance to tune out the capacitive reactance present due to a higher resonant frequency. It also transforms the 36 ohms resistance to 50 ohms for the transmitter.

Snapshots of the fencing radial system, the split bolt connector I used to connect to it, and the Smartuner.

Footnote: In 2006, I replaced this vertical with a commercial version from DX Engineering (Model DXE-40VA-1). The top hat is bigger and the whole antenna is physically stronger. It works great. One big advantage to me is, I can use higher power if I need to. On 40m full legal power can produce a lot more dx contacts. Now, in 2007, I am thinking about putting up a second one and phasing it. The more I read about phased arrays the more I realized how tricky the phasing is. I have not decided one way or the other at this point.