To: <towertalk@contesting.com>
>Date: Thu, 31 Dec 1998 10:02:34 -0500
>From: "w8ji.tom" <w8ji.tom@MCIONE.com>
>
>Hi Eric,
>
>I initially wrote:
>
>> >A J-pole is not as ground independent as one might think or
>> >read. It is an end-fed 1/2 wl antenna , fed by a 1/4 wl
>> >stub. Because the end impedance of a 1/2 wl is not infinite,
>> >but rather some finite impedance between a few hundred and a
>> >few thousand ohms (depending on conductor diameter), the
>> >terminal current of the balanced line driving the actual
>> >radiator are not balanced. This causes the balanced feedline
>> >section of the J-pole to have common mode current adding
>> >destructive radiation to the pattern that raises the wave
>> >angle and lowers gain.
>
>> We did not observe any telltale indications of an excessive
>> feedline radiation problem with the one we used on 160 a few
>> years back. And we were looking for this effect during the
>> tuneup phase of the project.
>
>Respectfully Eric, it depends on how you "looked". The antenna
>doesn't work "bad", it just doesn't work the way most people or
>models predict.
We were looking explicitly for RF current flowing on the outside
of the coax feedline shield. We were admittedly not using a
hypersensitive method to quantify it. But if there had been
significant RF current on the shield, we would have easily been
able to detect it.
>
>> We did isolate the coax at the feedpoint (turned out to be only a
>> few inches out from the short for 50 ohm feed) with a bunch of
>> ferrite beads. But we never had any bead temperatures above
>> ambient even when running significant power into the system.
>
>First the terms. The top 1/2 sticking straight up is the
>radiator and intended to radiate. The parallel section below it
>is the (matching) stub and is assumed to not radiate in the
>regular J.
Yup! No warranties about the validity of the assumption.
>
>Was the stub part of the pole isolated or grounded to another
>structure?
It was isolated except for the connection via the coax shield.
>If the stub section is isolated from other conductors, including
>the mast and feedline, common mode excitation is minimized by
>the stub section and antenna forming an unbalanced to balanced
>connection. In that case the antenna will work about like a 1/2
>wl vertical, especially if the radiator is thin compared to
>effective diameter of the two wire stub used to feed the
>radiator.
That is basically what the modeling predicted. The model
included the 1/4 wave stub as wires. The radiator and the stub
were both constructed from #14 hard drawn 7 strand copper wire.
Snip... (discussion about conductor diameters not relevant to the
antenna in question)
>Grounding the stub at the bottom or making the stub thin
>compared to the radiator can make the stub radiate as much or
>more than the actual 1/2 wl element, and current is 180 degrees
>out-of-phase with the intended radiator. That places a pattern
>null at zero degrees elevation.
Neither of these cases was met. I should probably describe the
actual antenna so we don't have to waste time talking about
nonexistent conditions.
The antenna was entirely supported by a nylon rope through a
pulley at the 450 foot level on a 650 foot tower. It was not
exactly vertical. It actualy hung off to one side of the tower
in a kind of two stage catenary curve. One curve for the single
radiator wire and a different one for the heavier stub. So the
radiator was probably 10 to 15 degrees off vertical and the stub
was roughly 45 degrees off (linear approximations of the
curves). This configuration was selected by modeling the way
the antenna hangs and then adjusting pulley position and lower
anchor point location to put the radiator where it got the maximum
benefit from the tower as a reflector. This location took about
10 dB off the back side and put 3 or so on the front according to
the model. A quick check of relative signal strengths with a
portable receiver and an attenuator confirmed that there was
indeed approximately 13 or 14 dB F/B. I have no idea what the
rest of the pattern looked like. But I bet the model wasn't
terribly far off.
>If a model does not include real-world effects like grounding or
>less-than-perfect feedline choking, it was not an accurate model. This very
>thing happen in an article about end-fed zepps, where someone modeled the
>end-fed zepp as a perfect antenna with a perfect ground independent power
>source exactly where the antenna and feeder were optimized. In the real
>world, especially if someone attempts to feed an end-fed zepp (which is
>really a bent J-pole) with a voltage-type balanced tuner or transformer,
>the feeder (or matching stub in the J-pole) will radiate quite
>significantly.
I am fairly conversant with wire antenna modeling techniques. I
did a pretty complete model of not just the j-pole but the
tower and all of its guys. We had pretty good info on the soil
characteristics for that particular area as well. The only
coductor that I can tink of that wasn't modeled was the power
connection from the rig back to the generator (about 100 feet of
power pigtail mostly laying on the ground).
>You can use that feeder radiation to advantage in the J-pole by
>intentionally feeding the bottom of the stub with a balanced
>voltage source, intentionally making each voltage at the
>terminal offset from center with reference to a ground plane. By
>using the correct offset, you can force the antenna to be
>excited with in-phase currents and behave like a 1/4 wl in-phase
>with a 1/2 wl and get a few dB of stacking gain.
Here is how the structure was fed from coax. The stub was
constructed of homemade 6 inch open line from the same #14 wire
used for the radiator. The stub was hoisted into place and tuned
for resonance at the desired frequency.
The shorted end of the stub was connected to the shield of the
coax. From there the coax was run up to the feedpoint location
and connected across the stub legs. The ferrite decoupling beads
were placed on the coax shield between the point where it
attached to the shorting bar and the transmitter. The beads were
located right next to the shorting bar. We used 30 or so 73 mix
beads. It was probably not an optimum configuration but it was
what we had available and none of the beads got hot (or even
warm) when running high power key down for minutes.
Once we had a feedpoint, we attached the radiator, hauled the
entire structure into place and went through several iterations
of locating the feedpoint for minimum SWR and then adjusting the
radiator length for minimum reactance at the desired operating
frequency. When we were done, we had a 50+j0 feed point at 1.830
MHz.
>
>> Also, we had no "RF in the shack" types of problems and the
>> coax feedline dress or length had zero effect on the observed
>> feed impedance. And that was true both with and without the
>> beads on the coax.
>
>Feedline radiation is another topic. RF in the shack is an
>esoteric term. You might have had "RF in the Shack" and not
>even known it.
Trust me. If there had been significant RF current on the
feedline shield, we wouldn't have been able to touch any of the
equipment cabinets when running full power. Although all the
radio equipment was bonded together, there was no external ground
reference other than back to the generator through a rather long
power cable. The generator itself was not grounded and the
trailer was a fiberglass shell. The RG213 was a 50 foot length
hanging in the air from the feedpoint to the place where it
atached to the amplifier output connector. No grounded entry
panel here.
>Or the feedline might have coupled off all the RF through mutual
>coupling to surroundings before it got there!
I don't think so. Not much other than earth for it to couple
into and it wasn't very close to earth for most of its run.
>Once again, try to find Danny (K6MHE's) CQ article on
>J-poles. These electrical problems are the reason why commercial
>systems never adopted the j-pole, despite its mechanical
>advantages.
I will.
Of course I have no idea whatever what the absolute efficiency
for this thing was. But I don't think it was particularly lossy.
During testing a week before the contest, we had no trouble
working anyone we heard on SSB (we were running 5 to 10 watts
peak from a battery power supply for testing). That included
stations in the carribean and labrador. Not a measurement but a
pretty good indication. Another indication is that the Q seemed
to be pretty high (2:1 SWR BW for a full size radiator was 60
KHz). Again, not a measurement but a confirming indication.
We also may very well have been suffering from radiation from the
stub. We have no way to know that one way or the other. But we
are pretty certain that there was very little RF on the coax
feedline shield. Whatever stub radiation there may have been was
not apparently a problem for the operation of the system. We had
tons of fun. There was not one operator there who would not have
immediately put that same antenna up at his home QTH if it was
feasable.
73, Eric N7CL
--
FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|