Topband: Ground Radial Discussion

[k8bhz at alphacomm.net]

I've taken the liberty of opening a new thread, which I probably should 
have done before.

To Dan, AC6LA,

First let me also thank you for your civil replies. As you stated, 
"Refreshing in this internet age".

I think I have a simple comparison to show the problem, using the 1/4 WL 
Vertical (36 ohms R radiation) and 1 kW rf.

First, consider antenna operating over the perfect ground plane. All the 
power is radiated. P = I sq X R radiation. So, I = 5.27 Amps.

Then, let's replace the perfect ground with an 8' ground rod, which in 
my soil is about 100 ohms.

Now the antenna feed impedance is R radiation + R ground, or 36 + 100 = 
136 ohms. Then apply our 1 kW. I is now down to 2.71 Amps. Radiated RF 
power is now 264.7 W & 735.3 W goes into heating the ground. This 
illustrates two things. One, why we don't use a ground rod, but more 
importantly, when R ground increases, using constant power, the current 
and efficiency drop.

Let's just see how much power is required if we were to achieve the same 
ground current as the perfect ground model. As the I squared term would 
be identical, the power required would simply be the ratio of 
impedances, or P = 1 kW (136/36) = 3.78 kW!

Now, realizing that the electrical 1/4 WL radials present the lowest R 
ground to the antenna, you can see that greatest efficiency is to be had 
there. Likewise, with 1/2 electrical WL radials, R ground is the 
highest, so efficiency is the worst.

In Rudy's discussion of a constant current into the radials (and that 
current MUST be the same into the vertical), the RADIATED power is 
constant. However the ground losses vary considerably. As the radial 
resistance increases when exceeding 1/4 electrical WL, Power input must 
be increased to achieve the same current. Importantly, as the radials 
are now not being fed at a current maximum, but elsewhere on the 
sinusoidal pattern, larger values will be observed elsewhere along the 
radials, increasing their losses.

Hope that helps...

Brian  K8BHZ