[TowerTalk] RF Ground is not a Myth

[Jim Brown]

On Fri,1/23/2015 3:20 PM, Steve Maki wrote:
> Can't one specify a perfect ground (even though impossible in the real 
> world) when modeling antennas? Is that not a useful exercise as an 
> educational tool?

"Specify?" I think you mean "assume." But few antennas depend upon a 
connection to earth for their operation (some RX antennas do).

One can ASSUME a homogenous earth of known properties for the purpose of 
modeling an antenna, but that's not what this discussion is about, nor 
does in involve a CONNECTION to the earth, nor is the fairy tale that is 
the subject of this email useful. Rather, modeling programs consider the 
interaction of electric, magnetic, and electromagnetic fields (and 
wavefronts) with the earth. The coupling of the fields results in loss 
(for real earth) and the interaction of the wavefronts produces the 
vertical pattern.

The only good reason for assuming homogenous (uniform) earth is that it 
makes the math simple enough that we can write the equations and solve 
them on computers that we can afford to have on our desks, and within a 
reasonable length of time.

There's another major point about this. Circuit common is NOT a single 
point. When we draw schematics, we show only half of the circuit -- the 
signal flow through components. But current flows in loops, the return 
path -- what EMC guru Henry Ott calls "the hidden schematic lurking the 
ground symbol." Those return paths can be well controlled -- a 
transmission line -- or they can be random and uncontrolled.

When we build our circuits on multi-layer boards with a continuous layer 
forming a ground plane under a layer with traces on it, the trace and 
ground layer form an unbalanced transmission line, and the current on 
the trace returns in a narrow area directly under the trace, minimizing 
crosstalk and making the circuit more stable. If there is no ground 
layer, or if the ground layer is broken under a trace, the return 
current finds whatever random path it finds, depending on the whim of 
the equipment designer, and the circuit is subject to crosstalk, noise 
pickup, noise radiation, and circuit instability.

If you look inside a good power amp or antenna tuner, you will see coax 
running between the antenna switches and relays. The signal returns on 
the coax shield. Poorly designed units omit the coax, using single 
wires, and the current returns on the chassis. The otherwise very nice 
Ten Tec antenna tuners have that design defect -- try measuring a 6M 
antenna through that tuner. It's a mess. By contrast, Ten Tec HF power 
amps use coax for antenna switches, and 6M goes through just fine (with 
the amp inactive, of course).

73, Jim K9YC