Topband: Best wire antenna for roof top location

Guy Olinger K2AV k2av.guy at gmail.com
Mon Aug 10 21:48:56 EDT 2015


There is a rather stern W7EL caution in the EZNEC doc about using two
ground media. He particularly nixes using the inner media as sea water
to mimic a high conductive surface.

------------(EZNEC doc)-----------

"Using Two Ground Media

"The ground may be broken into two "media", each having its own
conductivity and dielectric constant (relative permittivity). The
second medium can be at a different height, but must be at the same
level or below the first medium. The media can be arranged in parallel
slices or concentric rings. One use of two media would be to model an
antenna on a lake surrounded by land. A very important thing to
understand is that the second medium is used only for far field
pattern calculations, and is ignored for all other purposes. Be
careful when using two media, and keep the following in mind:

"Even if you place the antenna over the second medium, EZNEC will
always use the ground constants and height (z = 0) of the first medium
for calculation of the impedances and currents.

"The second medium is used only for far field calculations. Near Field
and Ground Wave calculations assume that the first medium is of
infinite extent, and ignores the second medium.

"The effect of the second medium is taken into account only in a very
simplified way. The vertical pattern is generated by tracing "rays"
direct from the antenna and reflected from the ground. When a second
medium is used, the ground reflection "ray" is determined by whichever
medium it strikes the top of. The "ray" does not penetrate either
medium, and diffraction or similar effects aren't considered. Because
of this, a highly conductive inner medium and normally conductive
outer medium is not a good model of a ground radial system, and
shouldn't be used for this purpose.

"Buried wires in EZNEC Pro/4 will always be treated as though they're
immersed wholly in the first medium."

-----------(end EZNEC doc)---------

The effect of this is to render any use of "ground" to represent the
roof of a building as potentially grossly inaccurate, particularly as
applies to building loss. The simplest way to summarize is that NEC
x.x cannot model this situation without creating a presumptive model
of the building using resistively loaded wires. Can be done, but a
real pain in the *ss. Even then the question remains as to WHAT values
to assign to anything.

This is one place where anecdota from trusted sources and other
considerations besides pattern are far more valuable decision tools.

Noise and lots of lossy conductors beneath would get me away from the
roof for low bands.

Personally, in that situation, I would try an FCP on the roof as
counterpoise and then pull a wire down and out at a 45 degree angle.
End feed the wire at the roof against the FCP. The FCP would couple
the roof far less than a single wire, and have far less pickup from
from horizontal conductors below. I would make the wire 5/16 wave on
160 which would get the current max on the sloping wire away from the
building. Use a trifilar winding on a T400A-2 #2 powdered iron core,
and then an autotuner. Given my band use on 160 (low end CW only), I
would use a tank circuit tuner tuned at 1.83 MHz.

Not having coax running up to the sloping wire, not fed as a doublet,
would allow a far less visible installation and a durable pull on the
wire.

When I was a freshman in college we strung #12 copperweld-like wire
*between* two dormitories, and fed an 80m halfwave off center, one of
those Heath double bifilar coil baluns hanging off the wire connected
to RG59 down to a window on second floor, and used it on 80, 40 and
20. The wire had very secure mount at either end and supported the
weight without problem. Was a killer antenna.

The dormitories were right at the edge of a ridge line looking out
over the Kentucky Blue Grass region. That was a four story building
down to the ground, which went 100 feet to the northwest and then
quickly dropped 150-200 feet to the Blue Grass plain below. The high
view was from southwest to northwest to northeast. Could see 25 miles
out to the horizon for 180 degrees. You could see the incoming weather
from the northwest hours ahead of time.

Will give you that was one wire off a building that depended on
location, location, location.

73, Guy K2AV

On Mon, Aug 10, 2015 at 2:00 PM, Jim Brown <jim at audiosystemsgroup.com> wrote:
> Hi Bud,
>
> Thanks for comments.
>
> I used sea water in the model to correspond to your copper plate example. I
> also tried other less conductive mediums for the roof as a sanity check.
> What changes is the strength of the high angle stuff.
>
> No surprise that a dipole 5 ft above a conductive surface would be a train
> wreck both as a radiator and as a match.
>
> I just looked at an 80M dipole 20 and 30 ft above that sea water roof.
> Feedpoint Z about 15 ohms at 20 ft, about 30 ohms at 30 ft. Narudi's shack
> is on the roof, so a feedline will be short enough that a 3:1 mismatch on
> 80M should not be an issue with RG8. Obviously, a different conclusion if
> his shack was on the first floor, with 400 ft of coax running to it. If you
> change the roof medium to less conductive, feedpoint Z is about 28 ohms for
> an 80M dipole 20 ft above the roof.
>
> I understand the goal of your 5 ft high antenna in the example for a
> domestic contest. Narudi is in Jakarta, and his objective is CQWW.
>
> I still think the major issue will be RX noise.
>
> 73, Jim K9YC
>
>
> On Mon,8/10/2015 5:08 AM, W2RU - Bud Hippisley wrote:
>>
>> Jim — I’m not sure you’re “missing” anything from a theoretical
>> standpoint.  Some comments from a practical standpoint, however:
>>
>>    1.  In my dormitory roof example, our objective was to work W/VE — most
>> of whom were very close to us, hence very high angle.  We might well have
>> had a great DXing pattern; I just wasn’t “into” DXing back then.
>>
>>    2.  In a  two-medium model such as the one you’re describing below, the
>> relatively small inner medium (the roof of Narudi’s building directly
>> beneath the dipole) sets the feedpoint impedance of the dipole, which will
>> be very low because the dipole height above Narudi’s roof is very low when
>> measured as a fraction of a wavelength on 80 meters.  This leads to what I
>> call “super-gain” models which can have the far field “boost” you describe
>> but which are extremely difficult to realize in practice, due to the
>> difficulty of matching such low impedance feedpoints without substantial
>> feedline or antenna matching unit losses.  For brevity in my initial posting
>> I didn’t mention that the low-Z feedpoint on 80 and 40 made it impossible to
>> properly match our dormitory dipole with just the pi-network output of the
>> 813 rig.
>>
>>    3.  You say you used sea water as your inner medium.  I’m not sure I’d
>> equate a rooftop with sea water — even “my” rooftop with a solid sheet of
>> copper flashing under the tar and gravel.  I think having a dipole really
>> close to a rooftop tends to compress the entire dipole pattern because of
>> losses in that “ground” system directly beneath the dipole.
>>
>> So, overall, I think the _pattern_ you got from your model is probably not
>> too horribly off, but the overall efficiency and gain of a real-world
>> implementation of that antenna system is not the greatest.
>>
>> Bud, W2RU
>>
>>
>>
>>
>>>
>>> On Aug 10, 2015, at 1:56 48AM, Jim Brown <jim at audiosystemsgroup.com>
>>> wrote:
>>>
>>> On Sat,8/8/2015 10:36 AM, W2RU - Bud Hippisley wrote:
>>>>
>>>> but, with the usual wire sag, the feedpoint was about 5 feet above the
>>>> gravel.  We weren’t worried, because the roof was at least 70 feet above the
>>>> surrounding terrain.
>>>
>>> Hmmm!  Let's remind ourselves of Nuradi's situation. The roof is 110m
>>> high, 45m x 33 m. Corner to corner is less than a wavelength on 80M, more
>>> than a wavelength on 40M, but the distance to a corner from a wire strung
>>> between the two corners is less than a quarter wave on 80M, less than a half
>>> wave on 40. Assuming an ideal conductor on the roof, it's going to act as a
>>> reflector going upward, but the low angle pattern will be determined in the
>>> far field.
>>>
>>> I've not worked before with two ground media, so I pulled out W7EL's
>>> instructions for doing so. I built a very simple model attempting to roughly
>>> simulate Nuradi's situation. I'm running NEC2 with EZNEC Pro5. The first
>>> ground medium is sea water, with a radius 120 ft (it's a rectangular
>>> building so that's an approximation. The second medium is Very Poor: cities,
>>> industrial, and it's at -360 ft. Yes, I'd like to elevate the first medium
>>> and have the second medium at 0 ft, but EZNEC won't let me do that. I
>>> simulated 40M and 80 dipoles in the range of 20-30 ft. What I got was a two
>>> lobe vertical pattern -- a VERY strong, very narrow low angle lobe, and a
>>> broad upward lobe whose strength depends on the height of the dipole above
>>> the roof. Very low (5 ft) makes the bottom lobe VERY narrow and VERY low
>>> (about 2 degrees) and makes the high lobe a lot weaker. Yes, it's a poor
>>> antenna -- IF its low to the roof. But if it's up 20-30 ft, a horizontally
>>> polarized wire looks like a nice DX antenna.
>>>
>>> What am I missing?
>>>
>>> And, like I said before and several others added -- all that stuff on the
>>> roof is likely to be mondo noisy.
>>>
>>> 73, Jim K9YC
>>
>> _________________
>> Topband Reflector Archives - http://www.contesting.com/_topband
>
>
>
> _________________
> Topband Reflector Archives - http://www.contesting.com/_topband


More information about the Topband mailing list