[TowerTalk] modeling LPDAs

[Leonard Kay]

Hi all,

Anyone have any experience to share modeling LPDAs?

I'm trying to model my T10 using NEC4WIN, and I'm having 
what looks like ill-formed matrix problems. Granted, there are 
a LOT of segments - LPDAs have a lot of 'wires' since 
the boom is hot - I have 39 wires @ 4-10 segments each.

When I graph any quantity (gain, f/b, swr) over a frequency 
range, I get mostly smooth curves, but every so often 
I get a very sharp spike off the chart. (This is in free space)
I've tried twiddling everything, from wire size to 
# of segments, etc. Also the feed impedances don't seem right.
The azimuth pattern _does_ look good, and apart from the
sharp spikes the average numbers for gain, f/b, swr look good.

Have I hit on a limitation of NEC4WIN? Past a point, the
results don't change as I increase the # of segments, so I
don't think it's a resolution problem....  

Tnx & 73
Len K1NU


--
FAQ on WWW:               http://www.contesting.com/towertalkfaq.html
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm