[TowerTalk] Tuning raised radial verticals

[Dan Maguire]

Gary, NA6O wrote:
>>> Why the manuals say to tune the verticals 150 or 200 kHz low makes no sense to me.

I spent some time with that question.  From the DXE Transmit Four Square System (DXE-TFS4-80B) manual:

"Set the resonance to the low edge of your desired operating range. This is because array performance peaks approximately four percent higher in frequency than the low resonant point of each individual element. Example: To obtain minimum dump power for 3.65 MHz, each vertical is tuned for resonance at 3.51 MHz."

The Comtek (COM-ACB-80-A) manual has similar wording except it recommends element self-resonance of 3.55 kHz for an array design frequency of 3.65 MHz.

So I started with a model having these characteristics:

a) Conventional 4-sq (ground mounted qtr-wave verticals) with an assumed ground loss in the radial field of 5 ohms per element.
b) Diameter of the vertical elements set to simulate the DXE-7580FS-VA-1 Qtr-Wave Vertical (stepped diameters ranging from 2.125" to 0.375" over 16 taper sections).
c) Array design frequency at 3.650 MHz.
d) No modifications to the hybrid coupler.

Then I ran comparisons for the gain at 20° elevation, the F/B at 20° elevation, and the dumped power, with the elements set for resonance at either 3.51 MHz or 3.65 MHz (140 kHz difference).

https://s26.postimg.org/hmszkjhs9/NA6_O-1.gif

https://s26.postimg.org/6yp8lp7t5/NA6_O-2.gif

https://s26.postimg.org/4jxd19rk9/NA6_O-3.gif

If one considers either maximum gain or minimum dumped power (per DXE) as the metric for "array performance at 3.65 MHz" it looks like the elements should be tuned about 90 kHz lower, not 140 kHz.  Of course, that's just for this particular modeled scenario.  And although that F/B peak of ~52 dB looks impressive it's important to keep in mind that's for exactly 180° azimuth.  The Front/Rear response (gain at 0° azimuth compared to the range 90-270°, not just 180°) is almost identical for both cases.

The above study was done using the "Vertical 4-sq Hybrid sweep.weq" AutoEZ format model, available for download at the end of the first section on this page:

http://ac6la.com/aecollection8.html

======

Gary also wrote, regarding the N2WQ "single radial, 'L' shaped wire elements" array:
>>> Single radials are not good for pattern, as previously explained. ... I cannot see what adding 1/8-wave radials would accomplish, intuitively nor in simulation.

I ran some models for that as well.  N2WQ had proposed adding three 1/8-wave radials per element, in addition to the single 1/4-wave radial.  A top-down view of the model looks like this, with the central tower and its grounding system temporarily removed to avoid clutter and confusion.  (Those wires were put back in for the actual modeling runs.)

https://s26.postimg.org/7cv5z1hah/N2_WQ-5.png

With the extra 1/8-wave radials added, the vertical portion of each element was changed from 68.5 to 68.7 ft to maintain element resonance at 3.413 MHz, same as the previous single radial model.  Then I ran sweeps to compare the gain at 20° elevation, the F/B at 20° elevation, and the dumped power.  For all sweeps and with both model configurations, the elements are being fed via a commercial hybrid coupler (DXE or Comtek) with no adjustments or modifications.

https://s26.postimg.org/q6gyw1fih/N2_WQ-6.png

https://s26.postimg.org/5wjngwedl/N2_WQ-7.png

https://s26.postimg.org/f31y46jm1/N2_WQ-8.png

I'd have to agree with Gary, it's hard to see any advantage with the extra 1/8-wave radials.  The gain is worse, the F/B is better, and the dumped power is about the same.

If anyone would like to experiment with Rudy's setup I have AutoEZ format models available for both the single 90° radial and the "1x90° + 3x45°" configurations that I'll be happy to share.  Drop me a line off-list.

Dan, AC6LA
http://ac6la.com