Re: NEC-LIST:old question on non-physical output

This issue of misinterpreting input deck data is interesting.

I have been looking into my model problems in more detail, and I only
get more confounded.

I have a much smaller test case antenna generated with the same software.

The problem is that I get discontinuous simulation results by slight
variations in my parameters (length, feed point ect...)

The wire generating program automatically assigns the number of segments
on a wire in an attempt to keep the segment length constant for all
wires. Some wires are short so a minimum of 1 segment is assigned.

My segment length to radius ratios for all the segments are not
excessively aggressive. they should all be greater than 4.

If I change the antenna length, I can cross the boundary where my
program adds one more segment to one or more wires. The basic shape of
the antenna has changed by only a tiny amount, and the number total
number of segment may have changed from 58 to 60. This small change
produces drastic swings in the the calculated Feed Z. I also see a
swing in the gain from the expected 1.4 dB gain for a dipole to 6 or 10
dB (must be non physical)

My structure is dipole like and displays symmetry (I do not use the
symmetry card), But the currents on the wires do not have perfect symmetry.

Is it possible that some of my wires are not connected?

For some combination of parameters the simulation actually fails with a
SBF Error

I can still extract the geometry from the failed output, and the antenna
looks fine.

I tried to load 4nec2 to use the geometry and rule checker but I looks
like it will not run on windows 2000. Is this true.

Is there a program to check the input Deck for Win 2000

I would be too happy to provide example input decks that display these
problems.

This is driving me nuts... I really appreciate everybody's attention.

Davor Virkes wrote:

> Hi there,
>
> The first thing with automatically generated segments is their vast number
> of digits beyond the decimal point. My observation with NEC2D is that it
> misinterprets data expressed with more than 2 decimal places in millimeter
> scale. You could safely round up all segments' coordinates prior to
> simulation, and check currents after simulation.
> Beside everything else others suggested, I noted this particular behavior
> with reactive loads, such as bazooka, but everything converges nicely after
> proper segmentation - try hard to keep lengths of facing segments equal.
>
> Davor
>
> At 15:58 2002.05.29 -0400, you wrote:
>
>
>>I am generating some medium to large antenna files (800 segments)
>>
>>These antenna are generated automatically from a set of parameters. I do
>>my best to make sure that I do not violate any wire rules, but even when I
>>think there are no violations, I get strange results.
>>
>>For instance I may get negative port Z, or negative efficiency, or greater
>>than 100% efficiency.
>>
>>sometimes I will have a reasonable feed Z and good efficiency, but the
>>radiation pattern data is all -999.99 dB, Or the efficiency is ********
>>
>>
>>Perhaps NEC4 is more forgiving of small segments, or maybe I just need to
>>look into wire rule violations more.
>>
>>How should I interpret these strange simulation results?
>>
>>Wayne S
>>--
>>The NEC-List mailing list <nec-list_at_gweep.ca>
>>http://www.gweep.ca/mailman/listinfo.cgi/nec-list
>>
>>
>>
>>
>>
>>---
>>Incoming mail is certified Virus Free.
>>Checked by AVG anti-virus system (http://www.grisoft.com).
>>Version: 6.0.361 / Virus Database: 199 - Release Date: 2002.05.07
>>

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