Hi, Richard,
A bit of errata to correct my previous post: The difference in
current will not be as large as I stated in my last post, because for
the pair of 1/16 radials I did not divide the source current between
the two 1/16 radials, and therefore the difference is not nearly as
large. This came clear when I got the time to do a series of models
on the loaded 1/16 to quantify sensitivity to coil resistance.
The FCP is specified in the NEC 4 model literally as bare #12 copper
wires, which they are in reality. My "analysis" is simply retelling
the NEC4 calculated current on the wires and field strengths
underneath. The specification manner for the wires in the model is
very straight-forward and not subject to any of the "gotcha's" that
can sometimes bedevil modeling. What one can say it behaves "like" is
open to individual elaboration. I would point out that the three
parallel wires all all in the same field, and that by itself
discourages me from a description other than what NEC4 says the
currents would be and field densities at the ground underneath.
Further indication that quite different devices are in effect: With
all loading out of both models, at the feedpoint, the 1/16 pair under
the 125' vertical wire has -j626 ohms to tune out, but the FCP has
only -j176 ohms. When the 626 is taken care of with 55 mH of coil, the
1.5:1 SWR bandwidth on the pair is 26 kHz. When the 176 is taken care
of by pruning, the 1.5:1 SWR bandwidth is 51 kHz. L's with shorter
vertical runs will be proportionately tighter bandwidth for both,
making the SWR bandwidth on the loaded 1/16 an issue.
Also for the loaded two 1/16 setup, particularly with shorter vertical
runs in the radiator with higher feed current, the NEC4 model shows
that method is hugely sensitive to the effective series resistance in
the coil, and would need to be made of 1/4 inch copper tubing or the
like, regardless of the power level, to keep it in the running.
Coming up with the proper lumped resistance for the model to represent
a given coil in a given situation IS problematic, but it is easy to
prove sensitivity or not to coil resistance by gradually varying the
resistance up from zero and rerunning the model. Even 5 ohms with a
full size 125 foot vertical run costs nearly a dB.
In your analysis you said "a" radial, but the FCP clearly has fields
underneath proportional to the PAIR of loaded 33 foot radials. So it
most looks like a PAIR of opposed 33 foot radials with reduced
current.
The loss in the FCP in NEC4 is not subject to assumptions about wire
losses in coils or a representation of such as a transmission line
device because the wires are literally modeled as bare #12 copper in
their respective positions. Wire losses, current phases and
magnitudes are simply computed by NEC4. Very straightforward.
Overall, I think there is a reason why we aren't doing a lot of loaded
dual 1/16 wave radials out there. If it worked it would be too much
of an advantage, and ON4UN would have published it already. His
break-even point would appear to be at the 4 loaded 1/8 wave radials,
and I agree with him.
73, Guy
On Wed, Nov 16, 2011 at 11:06 PM, Rick Karlquist <richard@karlquist.com> wrote:
> Guy Olinger K2AV wrote:
>
>> Counting FCP segments 1 through 5. 33 feet per segment. Directions
>> used are for illustration only.
>>
>> 1: center to 33 feet east
>> 2: 33 feet east back to center
>> 3: center to 33 feet west
>> 4: 33 feet west back to center
>> 5: center to 33 feet east and end insulator.
>
> Your analysis IMHO doesn't take into account coupling between
> the two conductors in the open wire line. I would characterize
> the above as a 33 foot radial in series with two 33 foot
> shorted stubs. A 33 foot shorted stub made of 600 ohm line
> is equivalent to about 20 microhenries of inductance.
> Two of those add up to 40 microhenries. This is close to the
> 55 microhenry loading coil you mentioned. Using a shorted
> stub of OWL to implement an inductor is an implementation
> decision. It seems less lossy because it doesn't get hot;
> the heat is spread out over a large area. But you still have
> the copper losses of a considerable length of wire which
> add up to a similar amount of loss that a big coil would have.
> You're probably right that it's cheaper than a coil, at least
> if you buy it new. Again, nothing wrong with doing this;
> I'm sure it works, but there is nothing magic going on here.
> It sounds like a nice ham-proof implementation of short elevated
> radials, which can be tricky to install the usual way.
>
> I don't see how any of this improves bandwidth except
> to the extent it adds loss to the system. There is a known
> relationship between antenna size, bandwidth and efficiency.
> Networks on the ground don't fundamentally affect this.
> Replacing an inductor with a shorted stub is always detrimental
> to bandwidth because the inductance of the shorted stub is
> proportional to frequency, instead of constant like the
> inductor.
>
> Rick N6RK
>
>
_______________________________________________
UR RST IS ... ... ..9 QSB QSB - hw? BK
|