NEC-LIST: NEC Problem

I'm having a serious problem with my use of NEC2 to calculate the response of a 1-m diam. 1-long cylinder
connected by a .01m radius axial wire 1-m long to a perfect ground plane. The cylinder is created by a line
of pathches rotated and reflected. The result varies greatly with the number of segments into which the wire
 is divided. One and 11 segments produce similar results, both seriously wrong. The best result, as compared
to a similar calculation with a 1-m cubic box, is with 2 segments. I may be making a stupid error, but can't find it.
 Is it possible that I've encountered a NEC bug? This problem is similar to one I reported for the cubic box
when the segments to which the wire was attached were too large, but I've created small segments on the bottom of the cylinder.
Will someone please help me?

The cylinder calculation always produces a Windows error message, "convert.exe has encountered a problem
 and needs to close", but proceeds anyway. My computer's hard drive already is partitioned as NTFS,
so I don't know why convert.exe is invoked.

-- 
The NEC-List mailing list
NEC-List_at_robomod.net
http://www.robomod.net/mailman/listinfo/nec-list