[antennaware] Radial Study

[Gary Breed]

Fellow antenna modelers:

I recently did a small modeling study to help shed some light on the elevated radial discussion. I have chosen to post it here instead of other reflectors, as I think this group is more reasoned and less blindly opinionated.

The study is simple in concept -- it compares three elevated radial configurations in transition from a START point of a vertical dipole to an END point of a quarter-wave vertical with four elevated radials.  By examining the differences in the intermediate configurations, I hoped to gain some insight into the magnitude of the ground losses involved and perhaps discover what radial types to use or avoid.

This study is limited in its scope -- it DOES NOT compare elevated to buried radials and it only compares resonant systems (non-reactive feedpoint impedance).

The fixed parameters are:

Frequency: 1825 kHz
Ground: EZNEC "average" -- conductivity = 5 mS/m; dielectric constant = 13; "high accuracy" ground model
Wire size: #12 copper

START -- vertical dipole with bottom eight feet above ground

END -- 1/4 wave vertical with four radials electrical 1/4-wave (128 ft. long), eight feet above ground

TYPE (1) -- lower half of the vertical dipole is split into four radials, sloped as needed to keep the ends eight feet above ground.

TYPE (2) -- bottom of vertical dipole split into four short radials (like a "capacity hat" on the bottom), with lengths adjusted to keep structure resonant. Bottom kept at eight feet above ground.

TYPE (3) -- like the END configuration with horizontal radials, at heights that put the top of the vertical the same as (1) and (2).

Two intermediate positions were modeled between start and end points --

Position (A) START -- top of vertical dipole is 270.75 ft.
Position (B) NEXT -- top of three types at 248.95 ft (feedpoint 118.85 ft.)
Position (C) NEXT -- top of three types at 204.95 ft. (feedpoint is 72 ft.)
Position (D) END -- top of antenna is 142.75 ft. (feedpoint is 8 ft.)

Raw maximum gain and takeoff angle data:

Position (A) -- START -- 1.21 dBi at 16 deg.
                [Type (3) at this height -- 1.0 dBi at 13 deg.]

Position (B) -- Type (1) (sloped radial) -- 1.66 dBi at 16 deg.
                Type (2) (bottom "hat") -- 1.09 dBi at 17 deg.
                Type (3) (horiz) -- 1.21 dBi at 14 deg.

Position (C) -- Type (1) -- 1.32 dBi at 18 deg.
                Type (2) -- 0.79 dBi at 19 deg.
                Type (3) -- 1.33 dBi at 16 deg.

Position (d) -- END -- 0.96 dBi at 22 deg.

CONCLUSIONS 
The Type (2) "bottom capacity hat" system was clearly the worst of the three.  This suggests that the use of short horizontal radials is not a good choice, since this type in position (C) is actually worse than the lower END configuration with its 1/4-wave radials.

The "best" of this group is the sloped-radial configuration with a well-elevated feedpoint. It is significatly better than a full-size vertical dipole. This, of course, is the traditional elevated ground plane with radials sloped to achieve a 50-ohm feedpoint. (This data may be more useful when building a ground plane for the higher bands.)

The behavior of the horizontal radial Type (3) is consistent with the findings of Christman, KB8I. Its maximum radiation peaks at some moderate height.

In general, a higher feedpoint results in improved gain and a lower takeoff angle (no surprise). To get this improvement, you can either slope the radials downward from the feedpoint or leave then horizontal, with little difference between them.

FOLLOW-UP FREE-SPACE COMPARISON
After seeing the big dip in the Type (2) performance, I took the configurations of Postion (C) and compared them in free space:

DIPOLE -- 2.14 dBi
Type (1) -- 1.73 dBi
Type (2) -- 2.01 dBi
Type (3) -- 1.53 dBi

The fact that the worst of the configurations over ground is better than the others in free space highlights its unusually high losses.

QUESTIONS RAISED FOR FUTURE STUDY
If you had a fixed available height to work with, would it be better to add more top-loading to the radiator in order to raise the radials?

Would more intermediate heights add any new data?

Does this comparison hold with 8, 16 or 32 radials?

Does anyone see something in the data that I've missed?

---------------------------------------------------------
Personally, this study has given me greater confidence that resonant radials are needed.  It's also interesting that sloped radials are essentially equal to straight ones at low to moderate height, and are the best choice for heights approaching a vertical dipole.

I'm ready for your critique and comments!

73,
Gary Breed
K9AY
gary@noblepub.com


--
FAQ on WWW:               http://www.contesting.com/antennaware-faq.html
Submissions:              antennaware@contesting.com
Administrative requests:  antennaware-REQUEST@contesting.com
Problems:                 owner-antennaware@contesting.com