Topband: coax cable counterpoise

[Terry Conboy]

On 2011-09-24 6:32 PM, Bob Kupps wrote:
> Hi when I lived in the Mojave desert I fed my 70' crank up tower as a folded monopole on 160m and used 8 radials made from electric 1/4 waves of old coax cable laying on top of the ground. The center conductors of the radials were fed with the shield of the feed line and the radial shields all connected back at the center to the DC tower ground. The far ends of the radials were left open and therefore reflected a direct short back at the center ends. This seemed very effective and I had good results even running 5 watts in contests.
>
> Anyway I now live in northern Thailand and am planning a full size 4 square array built over a flooded rice paddy which is about the opposite to my previous QTH in ground conductivity. I want to farm the land as much as possible however so am considering some type of elevated counterpoise instead of ground radials, possibly using only 1-3 tuned coax radials as a counterpoise for each vertical.
>
> Any thoughts on this approach?
>
> 73 Bob HS0ZIA
Bob,

Sounds like you should have a killer array on 160m with ground 
conductivity that is probably the next best thing to sea water!

The input inpedance of each of the of the open-circuited 1/4 wavelength 
coax radials, assuming RG-8A coax, isn't quite zero, but 1.23 ohms at 
1.83 MHz, according to TLW.  So you are adding 1.23/8 = 0.15 ohms of 
loss to the ground system, which you probably would never notice.  These 
impedances become inductive above resonance, and capacitive below 
resonance, so you're causing a slight reduction in the system bandwidth, 
since the reactiance moves the same way as the antenna vs. frequency.  
Again, these are all in parallel, so probably no big deal.  (Of course, 
smaller coax would have higher loss: RG-8X about 2 ohms each, and RG-58 
about 3 ohms.)

Otherwise, the coax radials just look like large diameter wire radials 
that are shorter than a free space 1/4 wavelength on the OUTSIDE of the 
coaxes, where it counts.  With on-ground (or on-sand, in the desert), 
any radial is essentially non-resonant and only serves to lower the loss 
resistance for currents flowing in the ground back to the base of the 
antenna.  So you probably have a bit higher ground system loss, since 
the shorter than 1/4 wl radials are somewhat less effective than longer 
conductors in most cases.

At your Mojave QTH, you would probably have seen hardly any difference 
if you had just shorted the center conductor and shield together on each 
of the coaxes at the base of the tower.

However, for elevated radials, they need to be pretty close to 1/4 
wavelength conductors (on the outside, in the case of coax radials) to 
be resonant.  So an elevated ground system made with [0.25 wl * VF] 
length coaxes  with have ground system impedance that is capacitive and 
this will appear directly at the feedpoint.  One way to compensate for 
this is to make the coaxes longer than [0.25wl * VF], so their input 
impedance will be inductive enough to cancel the effect of the outside 
of the coaxes being shorter than 0.25 wl.

To check this out, I made a model in EZNEC with a 132.7 foot vertical, 
and three 132.7 foot copper wire radials (elevated 10 feet) with 
resonance at 1.83 MHz.  Gain at 16 degree takeoff over "Very Good" 
ground is 2.74 dBi and feed Z is 39.4 ohms.   Assuming RG-8A coax 
radials, I added the input impedances of the open-circuit coax lines in 
series with the radials and adjusted the radial lengths for resonance.  
This required the coax radials to be 125.1 feet each (only 7.6 feet 
shorter than normal conductors).  Gain with the coax radials is 2.63 dBi 
and the feed Z is 40.4 ohms.   SWR bandwidth is very close to the same 
for both.  Modeling might get you in the ballpark, but will undoubtedly 
require some field adjustment due to variations in coax velocity factor, 
the dielectric constant of the outside jacket (which I've never seen 
specified), ground conductivity/permittivity, and the approximations 
inherent in modeling conductors near ground in NEC2.

Another alternative is tuning your vertical radiators longer than normal 
to supply enough inductive reactance to tune the system to resonance.  
For a single radiator, this probably isn't a big deal, but when dealing 
with a phased array, where the mutual coupling between the elements 
needs to be taken into account when designing the feed/phasing system, 
this will only make your life more complex.

Have fun!
73, Terry N6RY