Wire junctions? / free NEC input/output 3D file viewer available

1. How does NEC handle wire junctions? The simple reaction-integral mm
codes I've written handled them by making the junctions themselves nodes,
and I never did figure out how to join more than two wires without solving
for many nodes at each junction. It appears that NEC somehow allows
non-zero current at the ends of joined wires, without assigning a variable
to calculate the current at the junction. An understanding of how NEC
handles the following simple dipole modeled with three single-segment
wires would probably answer my question:

    ----------+-----(v)-----+----------
         1 2 3

If I were modeling this configuration, I would add nodes at the junctions
for a total of 5 nodes (and 5 unknowns). Then the fields from each
segment are easily calculated because each mode has zero-current on the
ends. The fact that adjacent nodes overlap make for a smooth total
current distribution. (These adjacent-node mutual impedance calculations
for overlapping segments require much finer integration than the other
mutual impedance calculations.) I think that NEC, however, manages to solve
the problem with only three unknowns, without calculating mutual
impedance of overlapping modes. If this is so, how can it calculate the
fields from, say, segment #1 above without knowing what the current is at
the junction (+) ?

Any help would be appreciated.

2. I have put the NEC input and output file viewers that I wrote at:

http://www.metronet.com/~dave.des

The "nv" program views simple geometry input files, and the "nov" views
output files with the segments color-coded for current magnitude. I've
found this helpful for locating hot spots. PC executables,
QuickBasic/QBasic source, and examples are included. Please send me any
comments or suggestions.

                        TNX
                                Dave
Received on Fri Dec 01 1995 - 00:19:00 EST