Just for kicks I modeled a pair of half wavelength antennas, one is end fed with no radials and the other is end fed with four radials at 45 degrees. The left pictures are the no radial variety the pictures on the right have the four 45 degree radials.
First the data...
It looks to me like the 45 degree radials raises feedpoint impedance over no radials significantly...
And now the current distribution...
Hmm, next to no current is flowing in the radials. 3 dB gain over a dipole seems like kind of a stretch to me based on this information...
This shows that the additional gain is insignificant, 0.12 dB additional gain really doesn't amount to anything. The take off angle, if there is a change, isn't enough for NEC to pick it up. As a matter of fact, the patterns look identical to me so I would be surprised if there was anywhere near one degree difference in "take off angle"...
Now for a few notes...
1) These models didn't include any feedline or mast, and simply compare antenna to antenna in near ideal, if not lab like conditions. The a99 would be affected by the feedline and mast more simply because it doesn't have a low impedance option such as radials for said current to flow. Because of this, the electrical length and running of the feedline and the electrical length of the mast can change the radiation pattern, current distribution and phase of said currents, and feedpoint impedance properties of said antenna as there is no low impedance option for the RF to flow that a set of radials would provide.
2) This potential in variance in potential feedpoint impedances may explain why many people claim that the A99 is a lossy antenna, such losses would help make tuning an antenna with such a problem easier by lowering the feedpoint impedance over the entire useable range of the antenna and beyond.
3) One thing to watch out for with the Big Hair design is the overall length of the vertical element. As you pass a half wavelength mark the radials are no longer in phase with most of the vertical element. This causes an initial drop in gain then a switch to a much higher angle dominant node. Because of this if there is anything in the area of the antenna that effectively lengthens the electrical length then the benefits of that design quickly evaporate.
The DB
