Frank's wrote:
. . .
These results appear to agree with your findings.
The structure could be fine tuned by including the insulation
and element tapering. Also I have not fully balanced the
segmentation. The only question I have is
how does this compare with a conventional 6 element
yagi at the same nominal height of 85 ft.

Out of curiosity, I brought up an EZNEC model of two stacked 5 element
beams from the ARRL Antenna Book, 20th Ed., model ARRL_5L15 95'.EZ. I
deleted half the elements to leave a single 5 element array, and lowered
it to 65 feet. I also changed the current source to a voltage source for
simplicity, and removed the wire loss (which EZNEC translates to NEC as
a bunch of loads) -- the wire loss makes a difference of only 0.05 dB.
Here's an NEC model of the 5 element array up 65':

CM 5L15 95'
CE
GW 1,11,-3.464585,-3.528538,19.812,-3.464585,3.528612,19.812,.008906
GW 2,11,-1.766258,-3.345701,19.812,-1.766258,3.3457,19.812,.0088644
GW 3,11,-.4925144,-3.283913,19.812,-.4925144,3.283913,19.812,.0089841
GW 4,11,1.120895,-2.98782,19.812,1.120895,2.98782,19.812,.0087506
GW 5,11,3.464584,-2.793987,19.812,3.464584,2.793987,19.812,.009137
GE 1
FR 0,1,0,0,21.2
GN 2,0,0,0,13.,.005
EX 0,2,6,0,1.414214,0.
RP 0,1,361,1000,80.,0.,0.,1.,0.
EN

Gain as-is is 14.11 dBi; with loss, 14.07 dBi. Zin = 21.51 - j22.26
ohms. Takeoff angle is 10 degrees.

Seems to me this would be a lot easier to build and support than the
"Gaussian" model. And I'll bet you could make up the half dB gain
difference quite easily by adding a sixth element.

People with the 20th Edition of the Antenna Book can open the model,
make the same modifications I did, and run it with the EZNEC ARRL
program furnished with the Antenna Book, or any EZNEC program type
except the demo.

Anyone who's impressed with the gain figure of either antenna should
model a dipole at the same height for comparison.

Roy Lewallen, W7EL