Re: NEC-LIST: 4NEC2 Optimizer

Hi Ed

The reason that I made the director element lengths a single optimisation
variable and spacing between director elements a single optimisation
variable during the optimisation process was more for a manufacturing point
of view. I also only allowed all the 'variables' to vary with a 1mm
resolution.

The boomlengths for the yagis were also set to a Maximum of 0.8 m for the 5
element yagi's and a maximum of 3.2m for the 17 element yagis (maximums for
boomlengths increased proportionally to number of elements).
Another bonus was that time to run the optimisations, for the designs I
previously specified, was only in the region of between 10 to 15 minutes on
a 1Ghz PC, as I found that 80 chromosomes and 80 generations (saving and
inserting the best 10 chromosomes from each generation into the next
generation) were more than sufficient for this relatively small size
problem.

I am going to attempt to find our '97 April AP magazine and look forward to
reading your article.

Kind Regards,

Brandon Orchard
Poynting Software (Pty)Ltd

----- Original Message -----
From: "Altshuler Edward E Civ AFRL/SNHA" <Edward.Altshuler_at_hanscom.af.mil>
To: "'Brandon Orchard'" <brandon.orchard_at_poynting.co.za>; "neclist"
<nec-list_at_gweep.ca>; <mikes_at_signull.com>
Sent: Wednesday, December 11, 2002 4:44 PM
Subject: RE: NEC-LIST: 4NEC2 Optimizer

> Hi Brandon,
>
> We used a Genetic Algorithm a number of years ago to optimize two types of
> Yagi antennas. For the first application we optimized for gain only. We
> used boomlengths from 3.6 to 6.1? at 432 MHz. We allowed both the lengths
> and the spacing between elements to float and obtained simulated and
> measured gains that were consistently higher than those that could be
> obtained using a conventional design approach. This was reported in the
> Winter 1998 issue of Communications Quarterly. The second application was
> for a broadband Yagi with low sidelobes. This Yagi was simulated for 219
to
> 251 MHz for sidelobes below 25 dB. It was scaled and measured at 1310 to
> 1510 MHz. These results are included in a review article in the April
1997
> issue of the Antennas and Propagation Magazine.
>
> I was surprised that you kept your spacing and element length constant. I
> suspect that you may have obtained a higher performance Yagi if you had
> included these parameters in your optimization.
>
> Best regards,
>
> Ed Altshuler
>
> -----Original Message-----
> From: Brandon Orchard [mailto:brandon.orchard_at_poynting.co.za]
> Sent: Wednesday, December 11, 2002 4:35 AM
> To: neclist; mikes_at_signull.com
> Subject: RE:NEC-LIST: 4NEC2 Optimizer
>
>
> Dear Mike
>
> I have optimised a couple of yagi's using SuperNEC's Genetic Algorithm
> optimiser(some of my MSc work) and I have listed the results below. Most
of
> the results were nice and broadband with very good Gains and VSWR's
.Please
> contact me if you require any of the antenna parameters for any of the
> designs listed.
> All the yagi designs below use elements with a 0.008m radius, a folded
> dipole feed, directors of all equal length, spacing between directors all
> equal, and the design frequency was set to 300Mhz (you can always scale
the
> designs for the right operation frequency) .
>
> -5 element yagi(vswr for 75ohms)
> Gain BW 22.3% VSWR<2 BW 8.7% gain @ 300Mhz 10.19dBi
> Gain BW 21.7% VSWR<2 BW 5% gain @ 300MHz 10.4dBi
>
> -6 element yagi (VSWR for 75ohms)
> Gain BW 14% VSWR<2 BW 5.7% gain @300MHz 11.62dBi
>
> -12 element yagi (VSWR for 200 Ohms)
> Gain BW 29.3% VSWR<2 BW 20% gain @300MHz
> 12.11dBi
>
> -15 element yagi (VSWR for 200 Ohms)
> Gain BW 11.3% VSWR<2 BW 11.3% gain @ 300MHz 15.38dBi
> Gain BW 14% VSWR<2 BW 12.7% gain @ 300MHz 14.61dBi
>
> -17 element yagi (VSWR for 200Ohms)
> Gain BW 15.7% VSWR<2 BW 11% gain @ 300MHz 14.59dBi
> Gain BW 17.3% VSWR<2 BW 12.7% gain @300 MHZ 13.97dBi
>
> where %gain BW = ((Fh-Fl)*100)/Fc
> where Fc=300MHz(design frequency) Fh=highest frequency at which gain is
3dB
> lower than gain at Fc and
> Fl=lowest frequncy at which gain is 3dB lower than gain at Fc
>
> Regards,
> Brandon
> --
> The NEC-List mailing list <nec-list_at_gweep.ca>
> http://www.gweep.ca/mailman/listinfo.cgi/nec-list

-- 
The NEC-List mailing list <nec-list_at_gweep.ca>
http://www.gweep.ca/mailman/listinfo.cgi/nec-list